mit-6

Pd-Catalyzed Amination of Base-Sensitive Five-Membered Heteroaryl Halides with Aliphatic Amines

2026-03-14T05:51:44Z

Pd-Catalyzed Amination of Base-Sensitive Five-Membered Heteroaryl Halides with Aliphatic Amines Reichert, Elaine C; Feng, Kaibo; Sather, Aaron C; Buchwald, Stephen L We report a versatile and functional-group-tolerant method for the Pd-catalyzed C–N cross-coupling of five-membered heteroaryl halides with primary and secondary amines, an important but underexplored transformation. Coupling reactions of challenging, pharmaceutically relevant heteroarenes, such as 2-H-1,3-azoles, are reported in good-to-excellent yields. High-yielding coupling reactions of a wide set of five-membered heteroaryl halides with sterically demanding α-branched cyclic amines and acyclic secondary amines are reported for the first time. The key to the broad applicability of this method is the synergistic combination of (1) the moderate-strength base NaOTMS, which limits base-mediated decomposition of sensitive five-membered heteroarenes that ultimately leads to catalyst deactivation, and (2) the use of a GPhos-supported Pd catalyst, which effectively resists heteroarene-induced catalyst deactivation while promoting efficient coupling, even for challenging and sterically demanding amines. Cross-coupling reactions between a wide variety of five-membered heteroaryl halides and amines are demonstrated, including eight examples involving densely functionalized medicinal chemistry building blocks.

Abiotic peptides as carriers of information for the encoding of small-molecule library synthesis

2026-03-14T05:51:45Z

Abiotic peptides as carriers of information for the encoding of small-molecule library synthesis Rössler, Simon L; Grob, Nathalie M; Buchwald, Stephen L; Pentelute, Bradley L Encoding small-molecule information in DNA has been leveraged to accelerate the discovery of ligands for therapeutic targets such as proteins. However, oligonucleotide-based encoding is hampered by inherent limitations of information stability and density. In this study, we establish abiotic peptides for next-generation information storage and apply them for the encoding of diverse small-molecule synthesis. The chemical stability of the peptide-based tag allows the use of palladium-mediated reactions to efficiently synthesize peptide-encoded libraries (PELs) with broad chemical diversity and high purity. We demonstrate the successful de novo discovery of small-molecule protein ligands from PELs by affinity selection against carbonic anhydrase IX and the oncogenic protein targets BRD4(1) and MDM2. Collectively, this work establishes abiotic peptides as carriers of information for the encoding of small-molecule synthesis, leveraged herein for the discovery of protein ligands.

Studying Regioisomer Formation in the Pd‐Catalyzed Fluorination of Cyclic Vinyl Triflates: Evidence for in situ Ligand Modification

2026-03-14T05:51:43Z

Studying Regioisomer Formation in the Pd‐Catalyzed Fluorination of Cyclic Vinyl Triflates: Evidence for in situ Ligand Modification Ye, Yuxuan; Kim, Seoung‐Tae; King, Ryan P; Baik, Mu‐Hyun; Buchwald, Stephen L Pd-catalyzed nucleophilic fluorination reactions are important methods for the synthesis of fluoroarenes and fluoroalkenes. However, these reactions can generate a mixture of regioisomeric products that are often difficult to separate. While investigating the Pd-catalyzed fluorination of cyclic vinyl triflates, we observed that the addition of a substoichiometric quantity of TESCF3 significantly improved the regioselectivity of the reaction. Herein, we report a combined experimental and computational study on the mechanism of this transformation focusing on the role of TESCF3. The poor regioselectivity of the reaction in the absence of additives results from the formation of LPd-cyclohexyne complexes (L=biaryl monophosphine ligand). When TESCF3 is added to the reaction mixture, the generation of the Pd-cyclohexyne complexes is diminished by an unexpected pathway involving the dearomatization of the ligand by nucleophilic attack from a trifluoromethyl anion (CF3−).

Room-Temperature Cu-Catalyzed Amination of Aryl Bromides Enabled by DFT-Guided Ligand Design

2026-03-14T05:51:46Z

Room-Temperature Cu-Catalyzed Amination of Aryl Bromides Enabled by DFT-Guided Ligand Design Kim, Seoung-Tae; Strauss, Michael J; Cabré, Albert; Buchwald, Stephen L Ullmann-type C–N coupling reactions represent an important alternative to well-established Pd-catalyzed approaches due to the differing reactivity and the lower cost of Cu. While the design of anionic Cu ligands, particularly those by Ma, has enabled the coupling of various classes of aryl halides and alkyl amines, most methods require conditions that can limit their utility on complex substrates. Herein, we disclose the development of anionic N1,N2-diarylbenzene-1,2-diamine ligands that promote the Cu-catalyzed amination of aryl bromides under mild conditions. Guided by DFT calculations, these ligands were designed to (1) increase the electron density on Cu, thereby increasing the rate of oxidative addition of aryl bromides, and (2) stabilize the active anionic CuI complex via a π-interaction. Under optimized conditions, structurally diverse aryl and heteroaryl bromides and a broad range of alkyl amine nucleophiles, including pharmaceuticals bearing multiple functional groups, were efficiently coupled at room temperature. Combined computational and experimental studies support a mechanism of C–N bond formation that follows a catalytic cycle akin to the well-explored Pd-catalyzed variants. Modification of the ligand structure to include a naphthyl residue resulted in a lower energy barrier to oxidative addition, providing a 30-fold rate increase relative to what is seen with other ligands. Collectively, these results establish a new class of anionic ligands for Cu-catalyzed C–N couplings, which we anticipate may be extended to other Cu-catalyzed C–heteroatom and C–C bond-forming reactions.

Stereoselective Synthesis of Trisubstituted Alkenes via Copper Hydride-Catalyzed Alkyne Hydroalkylation

2026-03-14T05:51:40Z

Stereoselective Synthesis of Trisubstituted Alkenes via Copper Hydride-Catalyzed Alkyne Hydroalkylation Kutateladze, Dennis A; Mai, Binh Khanh; Dong, Yuyang; Zhang, Yu; Liu, Peng; Buchwald, Stephen L Alkenes are ubiquitous in organic chemistry, yet many classes of alkenes remain challenging to access by current synthetic methodology. Herein, we report a copper hydride-catalyzed approach for the synthesis of Z-configured trisubstituted alkenes with high stereo- and regioselectivity via alkyne hydroalkylation. A DTBM-dppf-supported Cu catalyst was found to be optimal, providing a substantial increase in product yield compared to reactions conducted with dppf as the ligand. DFT calculations show that the DTBM substitution leads to the acceleration of alkyne hydrocupration through combined ground and transition state effects related to preventing catalyst dimerization and enhancing catalyst–substrate dispersion interactions, respectively. Alkyne hydroalkylation was successfully demonstrated with methyl and larger alkyl tosylate electrophiles to produce a variety of (hetero)aryl-substituted alkenes in moderate to high yields with complete selectivity for the Z stereochemically configured products. In the formation of the key C–C bond, computational studies revealed a direct SN2 pathway for alkylation of the vinylcopper intermediate with in situ-formed alkyl iodides.

Room‐Temperature Copper‐Catalyzed Etherification of Aryl Bromides

2026-03-13T03:07:34Z

Room‐Temperature Copper‐Catalyzed Etherification of Aryl Bromides Strauss, Michael J; Greaves, Megan E; Kim, Seoung‐Tae; Teijaro, Christiana N; Schmidt, Michael A; Scola, Paul M; Buchwald, Stephen L We disclose the development of a Cu-catalyzed C−O coupling method utilizing a new N1,N2-diarylbenzene-1,2-diamine ligand, L8. Under optimized reaction conditions, structurally diverse aryl and heteroaryl bromides underwent efficient coupling with a variety of alcohols at room temperature using an L8-based catalyst. Notably, the L8-derived catalyst exhibited enhanced activity when compared to the L4-based system previously disclosed for C−N coupling, namely the ability to functionalize aryl bromides containing acidic functional groups. Mechanistic studies demonstrate that C−O coupling utilizing L8 ⋅ Cu involves rate-limiting alkoxide transmetallation, resulting in a mechanism of C−O bond formation that is distinct from previously described Pd-, Cu-, or Ni-based systems. This lower energy pathway leads to rapid C−O bond formation; a 7-fold increase relative to what is seen with other ligands. The results presented in this report overcome limitations in previously described C−O coupling methods and introduce a new ligand that we anticipate may be useful in other Cu-catalyzed C-heteroatom bond-forming reactions.

CuH-Catalyzed Regio- and Enantioselective Formal Hydroformylation of Vinyl Arenes

2026-03-13T03:07:21Z

CuH-Catalyzed Regio- and Enantioselective Formal Hydroformylation of Vinyl Arenes Garhwal, Subhash; Dong, Yuyang; Mai, Binh Khanh; Liu, Peng; Buchwald, Stephen L A highly enantioselective formal hydroformylation of vinyl arenes enabled by copper hydride (CuH) catalysis is reported. Key to the success of the method was the use of the mild Lewis acid zinc triflate to promote the formation of oxocarbenium electrophiles through the activation of diethoxymethyl acetate. Using the newly developed protocol, a broad range of vinyl arene substrates underwent efficient hydroacetalization reactions to provide access to highly enantioenriched α-aryl acetal products in good yields with exclusively branched regioselectivity. The acetal products could be converted to the corresponding aldehydes, alcohols, and amines with full preservation of the enantiomeric purity. Density functional theory studies support that the key C–C bond-forming event between the alkyl copper intermediate and the oxocarbenium electrophile takes place with inversion of configuration of the Cu–C bond in a backside SE2-type mechanism.

Cu-Catalyzed Amination of Base-Sensitive Aryl Bromides and the Chemoselective N- and O-Arylation of Amino Alcohols

2026-03-13T03:07:37Z

Cu-Catalyzed Amination of Base-Sensitive Aryl Bromides and the Chemoselective N- and O-Arylation of Amino Alcohols Strauss, Michael J; Liu, Kaylee X; Greaves, Megan E; Dahl, Jakob C; Kim, Seoung-Tae; Wu, Yong-Jin; Schmidt, Michael A; Scola, Paul M; Buchwald, Stephen L We report a general and functional-group-tolerant method for the Cu-catalyzed amination of base-sensitive aryl bromides including substrates possessing acidic functional groups and small five-membered heteroarenes. The results presented herein substantially expand the scope of Cu-catalyzed C–N coupling reactions. The combination of L8, an anionic N1,N2-diarylbenzene-1,2-diamine ligand, along with the mild base NaOTMS leads to the formation of a stable yet reactive catalyst that resists deactivation from coordination to heterocycles or charged intermediates. This system enables the use of low catalyst and ligand loadings. Exploiting the differences in nucleophile deprotonation in C–O and C–N coupling reactions catalyzed by Cu·L8 we developed a method to chemoselectively N- and O-arylate a variety of amino alcohol substrates. Employing NaOt-Bu as the base resulted exclusively in C–O coupling when the amino alcohols featured primary alcohols and more hindered amines or aniline groups. Utilizing NaOTMS enabled the ability to override the steric-based selectivity of these reactions completely and exclusively promoted C–N coupling regardless of the structure of the amino alcohol. The ability to invert the observed chemoselectivity is distinct from previously described methods that require protecting group manipulations or rely entirely on steric effects to control reactivity. These results substantially improve the scope of Cu-catalyzed C–N coupling reactions using N1,N2-diarylbenzene-1,2-diamine ligands and introduce a new chemoselective method to arylate amino alcohols.

Copper-Catalyzed Amination of Aryl Chlorides under Mild Reaction Conditions

2026-03-13T03:07:29Z

Copper-Catalyzed Amination of Aryl Chlorides under Mild Reaction Conditions Ai, Han-Jun; Kim, Seoung-Tae; Liu, Cecilia; Buchwald, Stephen L We report a mild method for the copper-catalyzed amination of aryl chlorides. Key to the success of the method was the use of highly sterically encumbered N¹,N²-diaryl diamine ligands which resist catalyst deactivation, allowing reactions to proceed at significantly lower temperatures and with a broader scope than current protocols. A sequence of highly chemoselective C-N and C-O cross-coupling reactions were demonstrated, and mechanistic studies indicate that oxidative addition of the Cu catalyst to the aryl chlorides is rate-limiting. We anticipate that the design principles disclosed herein will help motivate further advances in Cu-catalyzed transformations of aryl chlorides.

Development of a Deactivation-Resistant Dialkylbiarylphosphine Ligand for Pd-Catalyzed Arylation of Secondary Amines

2026-03-13T03:07:27Z

Development of a Deactivation-Resistant Dialkylbiarylphosphine Ligand for Pd-Catalyzed Arylation of Secondary Amines Feng, Kaibo; Raguram, Elaine Reichert; Howard, James R; Peters, Ellyn; Liu, Cecilia; Sigman, Matthew S; Buchwald, Stephen L Despite the prevalence of N-heteroarenes in small-molecule pharmaceuticals, Pd-catalyzed C-N cross-coupling reactions of aryl halides and amines containing these rings remain challenging due to their ability to displace the supporting ligand via coordination to the metal center. To address this limitation, we report the development of a highly robust Pd catalyst supported by a new dialkylbiarylphosphine ligand, FPhos. The FPhos-supported catalyst effectively resists N-heteroarene-mediated catalyst deactivation to readily promote C-N coupling between a wide variety of Lewis-basic aryl halides and secondary amines, including densely functionalized pharmaceuticals. Mechanistic and structural investigations, as well as principal component analysis and density functional theory, elucidated two key design features that enable FPhos to overcome the limitations of previous ligands. First, the ligated Pd complex is stabilized through its conformational preference for the O-bound isomer, which likely resists coordination by N-heteroarenes. Second, 3',5'-disubstitution on the non-phosphorus-containing ring of FPhos creates the ideal steric environment around the Pd center, which facilitates binding by larger secondary amines while mitigating the formation of off-cycle palladacycle species.

Kinetic Modeling Enables Understanding of Off-Cycle Processes in Pd-Catalyzed Amination of Five-Membered Heteroaryl Halides

2026-03-13T03:07:30Z

Kinetic Modeling Enables Understanding of Off-Cycle Processes in Pd-Catalyzed Amination of Five-Membered Heteroaryl Halides Raguram, Elaine Reichert; Dahl, Jakob C; Jensen, Klavs F; Buchwald, Stephen L The mechanism of Pd-catalyzed amination of five-membered heteroaryl halides was investigated by integrating experimental kinetic analysis with kinetic modeling through predictive testing and likelihood ratio analysis, revealing an atypical productive coupling pathway and multiple off-cycle events. The GPhos-supported Pd catalyst, along with the moderate-strength base NaOTMS, was previously found to promote efficient coupling between five-membered heteroaryl halides and secondary amines. However, slight deviations from the optimal concentration, temperature, and/or solvent resulted in significantly lower yields, contrary to typical reaction optimization trends. We found that the coupling of 4-bromothiazole with piperidine proceeds through an uncommon mechanism in which the NaOTMS base, rather than the amine, binds first to the oxidative addition complex; the resulting OTMS-bound Pd species is the resting state. Formation of the Pd-amido complex via base/amine exchange was identified as the turnover-limiting step, unlike other reported catalyst systems for which reductive elimination is turnover-limiting. We determined that the amine-bound Pd complex, usually an on-cycle intermediate, is instead a reversibly generated off-cycle species, and that base-mediated decomposition of 4-bromothiazole is the primary irreversible catalyst deactivation pathway. Predictive testing and kinetic modeling were key to the identification of these off-cycle processes, providing insight into minor mechanistic pathways that are difficult to observe experimentally. Collectively, this report reveals the unique enabling features of the Pd-GPhos/NaOTMS system, implementing mechanistic insights to improve the yields of particularly challenging coupling reactions. Moreover, these findings highlight the utility of applying predictive tests to kinetic models for the rapid evaluation of mechanistic possibilities in small-molecule catalytic systems.

Characterization of the response of radiochromic film to quasi-monoenergetic x rays through a cross-calibration with image plates

2026-03-13T03:07:23Z

Characterization of the response of radiochromic film to quasi-monoenergetic x rays through a cross-calibration with image plates Buschmann, BI; Cufari, M; Vanderloo, N; Vargas, J; Foo, BC; DeVault, A; Dannhoff, SG; Evans, TE; Johnson, TM; Kunimune, JH; Lawrence, Y; Pearcy, JA; Reichelt, BL; Wink, CW; Russell, L; Gatu Johnson, M; Petrasso, RD; Frenje, JA Radiochromic film (RCF) and image plates (IPs) are both commonly used detectors in diagnostics fielded at inertial confinement fusion (ICF) and high-energy-density physics (HEDP) research facilities. Due to the intense x-ray background in all ICF/HEDP experiments, accurately calibrating the optical density of RCF as a function of x-ray dose, and the photostimulated luminescence per photon of IPs as a function of x-ray energy, is necessary for interpreting experimental results. Various measurements of the sensitivity curve of different IPs to x rays have been performed [Izumi et al., Proc. SPIE 8850, 885006 (2013) and Rosenberg et al., Rev. Sci. Instrum. 90(1), 013506 (2019)]; however, calibrating RCF is a tedious process that depends on factors such as the orientation in which the RCF is scanned in the film scanner and the batch of RCF used. These issues can be mitigated by cross-calibrating RCF with IPs to enable the use of IPs for the determination of dose on the RCF without scanning the RCF. Here, the first cross-calibration of RCF with IPs to quasi-monoenergetic titanium, copper, and molybdenum K-line x rays is presented. It is found that the IP-inferred dose rates on the RCF for the Ti and Mo x rays agree well with the measured dose rates, while the IP-inferred dose rate for the Cu x rays is larger than the measured dose rate by ∼2×. Explanations for this discrepancy and plans for future work are discussed.

Determination of the response for the National Ignition Facility particle time of flight (PTOF) detector using single particle counting

2026-03-13T03:07:36Z

Determination of the response for the National Ignition Facility particle time of flight (PTOF) detector using single particle counting Lawrence, Y; Reichelt, BL; Wink, CW; Rigon, G; Johnson, M Gatu; Li, CK; Frenje, JA The Particle Time of Flight (PTOF) detector is a chemical vapor deposition diamond-based detector used to measure bang times in low-yield (≲ 1015 neutrons) experiments at the National Ignition Facility (NIF) at Lawrence Livermore National Laboratory (LLNL). Historically, the impulse response for PTOF diamond detectors has been obtained from x-ray timing shots on the NIF and shots on the MegaRay pulsed electron accelerator at LLNL. The impulse response may alternatively be obtained using single particle interactions with the detector, at substantially less cost and higher frequency compared to NIF timing shots, which typically occur months apart. Here, the response of a PTOF detector setup is characterized by statistically averaging a large number of single particle waveforms. A high fidelity instrument response function can be constructed in this way. This is confirmed by comparison of the single particle counting-constructed response to the impulse response function measured for the same detector at LLNL’s MegaRay facility.

Development of a compact magnetic spectrometer for use at the OMEGA Laser Facility and the National Ignition Facility

2026-03-13T03:07:32Z

Development of a compact magnetic spectrometer for use at the OMEGA Laser Facility and the National Ignition Facility Pearcy, JA; Russell, L; Kabadi, NV; Johnson, TM; Adrian, PA; Gatu-Johnson, M; Casco, E; Palmisano, K; Gates, G; Burgett, T; Scott, M; Petrasso, RD; Li, CK; Frenje, J Measurement of proton spectra is an important diagnostic for a variety of high energy density physics experiments. Current diagnostics are either not designed to capture the spectrum of low-energy protons or are unsuitable for high debris experiments. To bridge the gap, a new CR-39 based compact magnetic spectrometer (MagSpec) has been developed to measure proton spectra in the 1–20 MeV energy range, with a particular focus on the low-energy (1–6 MeV) spectrum, for use in experiments at the OMEGA Laser Facility and the National Ignition Facility (NIF). In the MagSpec diagnostic, protons of different energies are dispersed as they pass through a magnetic field before impinging on a differentially filtered CR-39 surface, resulting in a spatial distribution of CR-39 tracks that corresponds to the energy spectrum. In this paper, we discuss details of the design and implementation of MagSpec on the NIF and OMEGA.

Temperature stabilization of a lab space at 10 mK-level over a day

2026-03-13T03:07:26Z

Temperature stabilization of a lab space at 10 mK-level over a day Fife, Dylan; Shin, Dong-Chel; Sudhir, Vivishek Temperature fluctuations over long time scales (≳ 1 h) are an insidious problem for precision measurements. In optical laboratories, the primary effect of temperature fluctuations is drifts in optical circuits over spatial scales of a few meters and temporal scales extending beyond a few minutes. We present a lab-scale environment temperature control system approaching 10 mK-level temperature instability across a lab for integration times above an hour and extending to a day. This is achieved by passive isolation of the laboratory space from the building walls using a circulating air gap and an active control system feeding back to heating coils at the outlet of the laboratory’s Heating-Ventilation-Air-Conditioning (HVAC) unit. These techniques together result in 20 dB suppression of the temperature power spectrum across the lab at 10−4 Hz—approaching the limit set by statistical coherence of the temperature field—and 10 mK Allan deviation around 15 °C after an hour of averaging, which is an order of magnitude better than any previous report for a full laboratory.

Incentives to Comply with the Minimum Wage in the United States and the United Kingdom

2026-03-13T03:07:19Z

Incentives to Comply with the Minimum Wage in the United States and the United Kingdom Stansbury, Anna There is substantial evidence of minimum wage non-compliance in the United States and the United Kingdom. In this article, the author compiles new, comprehensive data on the costs that minimum wage violators incur when non-compliance is detected. In both countries, the costs violators face are often little more than the money they saved by underpaying. To have an incentive to comply under existing penalty regimes, typical US firms would thus have to expect a 47% to 83% probability of detection by the Department of Labor (DOL), or a 25% probability of a successful Fair Labor Standards Act (FLSA) suit. In the United Kingdom, typical firms would have to expect a 44% to 56% probability of detection. Actual probabilities of detection are substantially lower than this for many firms and would likely remain so even with realistic increases in enforcement capacity. Improved enforcement alone is thus insufficient: Expected penalties must also substantially increase to ensure that most firms have an incentive to comply.

Non-Euclidean motion planning with graphs of geodesically convex sets

2026-03-13T03:07:12Z

Non-Euclidean motion planning with graphs of geodesically convex sets Cohn, Thomas; Petersen, Mark; Simchowitz, Max; Tedrake, Russ Computing optimal, collision-free trajectories for high-dimensional systems is a challenging and important problem. Sampling-based planners struggle with the dimensionality, whereas trajectory optimizers may get stuck in local minima due to inherent nonconvexities in the optimization landscape. The use of mixed-integer programming to encapsulate these nonconvexities and find globally optimal trajectories has recently shown great promise, thanks in part to tight convex relaxations and efficient approximation strategies that greatly reduce runtimes. These approaches were previously limited to Euclidean configuration spaces, precluding their use with mobile bases or continuous revolute joints. In this paper, we handle such scenarios by modeling configuration spaces as Riemannian manifolds, and we describe a reduction procedure for the zero-curvature case to a mixed-integer convex optimization problem. We further present a method for obtaining approximate solutions via piecewise-linear approximations that is applicable to manifolds of arbitrary curvature. We demonstrate our results on various robot platforms, including producing efficient collision-free trajectories for a PR2 bimanual mobile manipulator.

Extended state infrastructure power in an age of networked competition: The cases of Thailand and Taiwan

2026-03-13T03:07:28Z

Extended state infrastructure power in an age of networked competition: The cases of Thailand and Taiwan Stokols, Andrew; Kollar, Justin Scholars have highlighted the emergence of infrastructure as a key domain in the struggle over network centrality in what some call the ‘Second Cold War’ between the U.S. and China. We qualify this ‘infrastructural turn’ by drawing attention to the contingent nature of state infrastructural power as depending on key domestic firms that often serve as intermediaries between domestic infrastructure and global supply chains or international partners. Utilising empirical case studies based on field research conducted between 2021 and 2023 in Thailand and Taiwan, we analyse the ways in which state infrastructure power is exercised through strategic negotiation between national politics of the state and territorial investment decisions of multinational and major domestic firms within global supply chains. The study highlights how outcomes of state projects to foster connectivity or centrality in networks are shaped by contingent and sometimes ad-hoc coalitions between state agencies and domestic and multinational companies with their own interests and agency. In the case of Taiwan, the centrality of Taiwan Semiconductor Manufacturing Company (TSMC) to global supply chains makes it an important player amidst continued U.S.-China tension. In Thailand, CP Group’s connections to China have afforded it a role as an interlocutor between Thailand and China, allowing it to obtain state infrastructure contracts. Through comparative case studies the paper complicates both ‘globalist’ and methodologically nationalist perspectives on the ‘infrastructural turn’, and introduces the concept of ‘extended state infrastructural power’ to account for this complex, networked exercise of state authority.

Uncovering Patterns in Overdose Deaths: An Analysis of Spike Identification in Fatal Drug Overdose Data in Massachusetts, 2017-2023

2026-03-13T03:07:33Z

Uncovering Patterns in Overdose Deaths: An Analysis of Spike Identification in Fatal Drug Overdose Data in Massachusetts, 2017-2023 Lee, Hannah; Otero-Leon, Daniel; Dong, Huiru; Stringfellow, Erin J; Jalali, Mohammad S Objectives: Yearly rolling aggregate trends or rates are commonly used to analyze trends in overdose deaths, but focusing on long-term trends can obscure short-term fluctuations (eg, daily spikes). We analyzed data on spikes in daily fatal overdoses and how various spike detection thresholds influence the identification of spikes. Materials and Methods: We used a spike detection algorithm to identify spikes among 16 660 drug-related overdose deaths (from any drug) reported in Massachusetts’ vital statistics from 2017 through 2023. We adjusted the parameters of the algorithm to define spikes in 3 distinct scenarios: deaths exceeding 2 adjusted moving SDs above the 7-, 30-, and 90-day adjusted moving average. Results: Our results confirmed the on-the-ground observation that there are days when many more people die of overdoses than would be expected based on fluctuations due to differences among people alone. We identified spikes on 5.8% to 20.6% of the days across the 3 scenarios, annually, constituting 11.1% to 31.6% of all overdose deaths. The absolute difference in percentage points of days identified as spikes varied from 5.2 to 11.5 between 7- and 30-day lags and from 0 to 4.6 between 30- and 90-day lags across years. When compared with the adjusted moving average across the 3 scenarios, in 2017 an average of 3.9 to 5.5 additional deaths occurred on spike days, while in 2023 the range was 3.7 to 6.0. Practice Implications: A substantial percentage of deaths occurred annually on spike days, highlighting the need for effectively monitoring short-term overdose trends. Moreover, our study serves as a foundational analysis for future research into exogenous events that may contribute to spikes in overdose deaths, aiming to prevent future deaths.

Experiential learning amid disequilibrium: Attuning to student emotions

2026-03-13T03:07:24Z

Experiential learning amid disequilibrium: Attuning to student emotions O’Flanagan, Sinead E; Y Jester, Michellana Educators recognize the significant role emotions play in experiential learning (EL), particularly in how they support students through the inherent emotion work. However, the traditional design of experiential learning theory (ELT) in higher education (HE) often presupposes a stable environment, which overlooks the impact of unpredictable external factors on students’ emotions and learning. Despite its critical importance, emotion work in EL remains underexplored, with emotional dynamics often obscured or dismissed as isolated incidents. This study sheds light on the heightened emotional challenges that arise during periods of sustained disequilibrium, such as the COVID-19-induced restrictions. It provides novel insights into the dynamic interplay of emotions and learning progression within EL frameworks, drawing on perspectives from EL educators, advisors, and students. The research underscores the importance of emotion-focused dialogue, educator-student connection, and assimilating autonomy needs in EL amid disequilibrium. It also identifies often-neglected elements in EL frameworks, such as students “sharing struggles” or “valuing work efforts,” alongside educator strategies like “personal anchoring.” The findings contribute to ELT by proposing adaptive strategies that integrate emotion work into pedagogical frameworks, enhancing reflection and conceptualization practices, and extending ELT’s applicability across diverse educational and work-based management learning settings.

Report to the President for year ended June 30, 2025, Associate Vice President for Research Administration

2026-03-12T07:26:11Z

Report to the President for year ended June 30, 2025, Associate Vice President for Research Administration White, Anne This report contains the following sections: Overview, Research Administration Services, OSATT Core, Technology Licensing Office, Corporate Relations, Research Systems & Support, and Cost Analysis.

Insulin Delivery Pumps for Human Spaceflight: Steps Toward an Accessible Space Future

2026-03-12T07:25:04Z

Insulin Delivery Pumps for Human Spaceflight: Steps Toward an Accessible Space Future Horn, Kyle J; Hoffman, Jeffrey A Commercially available insulin pumps for treatment of diabetes mellitus are currently not qualified to operate in the space environment. This work rigorously tested the fluid delivery performance of a Tandem t:slim X2 insulin pump in both micro- and hypergravity during a parabolic microgravity research flight. The parabolic research flight environment serves as an analogue to the types of transient gravitational loadings experienced during human-led missions, which provides a foundation to expand testing to suborbital and orbital flights in addition to other extreme environmental tests for wilderness dependency. The results of the flight data showed no significant difference between fluid delivery performance at 0, 1, and 2g acceleration regimes, nor at the transitions between gravity environments. Recommendations are made for further experimentation and qualification tests before use in future spaceflight missions.

On the nonlinear Eshelby inclusion problem and its isomorphic growth limit

2026-03-12T07:24:57Z

On the nonlinear Eshelby inclusion problem and its isomorphic growth limit Bonavia, Joseph E; Chockalingam, S; Cohen, Tal In the late 1950s, Eshelby’s linear solutions for the deformation field inside an ellipsoidal inclusion and, subsequently, the infinite matrix in which it is embedded were published. The solutions’ ability to capture the behavior of an orthotropically symmetric shaped inclusion made it invaluable in efforts to understand the behavior of defects within, and the micromechanics of, metals and other stiff materials throughout the rest of the 20th century. Over half a century later, we wish to understand the analogous effects of microstructure on the behavior of soft materials, both organic and synthetic, but in order to do so, we must venture beyond the linear limit, far into the nonlinear regime. However, no solutions to these analogous problems currently exist for non-spherical inclusions. In this work, we present an accurate semi-inverse solution for the elastic field in an isotropically growing spheroidal inclusion embedded in an infinite matrix, both made of the same incompressible neo-Hookean material. We also investigate the behavior of such an inclusion as it grows infinitely large, demonstrating the existence of a non-spherical asymptotic shape and an associated asymptotic pressure. We call this the isomorphic limit, and the associated pressure the isomorphic pressure.

Evaluating Risk for Astronaut Involvement in In-Space Manufacturing: Analog Field Testing and Future Planetary Surface Procedures

2026-03-12T07:24:54Z

Evaluating Risk for Astronaut Involvement in In-Space Manufacturing: Analog Field Testing and Future Planetary Surface Procedures MacRobbie, Madelyn; Patel, Palak B. Introduction A key objective of the NASA Artemis program is to establish a sustained human presence on the Moon, along with its international and commercial partners. NASA aims to establish a lunar economy, increasing the need for infrastructure to support human habitation and facilitate growth. In-space manufacturing (ISM) coupled with in situ resource utilization (ISRU) can reduce launch mass and reduce the dependency on Earth resupply for long-term habitation, enabling rapid expansion. However, the space environment introduces unique challenges compared to Earth, such as the absence of an atmosphere, reduced gravity levels, and high consequences of human-machine interactions given the barrier to evacuating an astronaut injured in a manufacturing accident on the Moon, necessitating new safety standards for ISM processes. Methods This study proposes the application of a modified analytical hierarchy process (AHP) to identify high-risk aspects of crew procedures in molten regolith electrolysis (MRE) for both Earth-based analog testing and lunar production. Results The modified AHP assists in pinpointing areas needing hazard mitigation to protect crew members, enabling the improvement of safety standards for MRE in both environments. Conclusion Findings will inform the development of robust safety protocols for ISM, crucial for the success of NASA's Artemis missions and the broader goal of sustained human presence on the Moon and Mars.

Origins of Face Responses in the Human Cortex: fNIRS and fMRI Evidence From Infants

2026-03-12T07:25:03Z

Origins of Face Responses in the Human Cortex: fNIRS and fMRI Evidence From Infants Saxe, Rebecca; Kosakowski, Heather L In adults, cortical regions in the fusiform face area (FFA), superior temporal sulcus (STS), and medial prefrontal cortex (MPFC) respond selectively to faces but underlie distinct perceptual and social processes. When do each of these regions, and their distinctive functions, develop? We reviewed recent studies of awake human infants’ cortical responses to faces using functional near-infrared spectroscopy (fNIRS) and functional MRI (fMRI). The results converged and do not support a slow, sequential posterior-to-anterior development of face-selective responses. Instead, cortical face-selective responses arise very early and simultaneously in infancy and may reflect distinctively social processes from the start.

The Effect of Basketball Analytics Investment on National Basketball Association (NBA) Team Performance

2026-03-12T07:25:02Z

The Effect of Basketball Analytics Investment on National Basketball Association (NBA) Team Performance Wang, Henry; Sarker, Arnab; Hosoi, Anette In the National Basketball Association (NBA), basketball data and analytics is an area of significant financial investment for all 30 franchises, despite there being little quantitative evidence demonstrating analytics adoption actually improves team-level performance. This study seeks to measure the return on investment of analytics on NBA team success in a time of great demand for analytical front office personnel. Using a two-way fixed effects modeling approach, we identify the causal effect of analytics department headcounts on regular season wins using 12 years of season-level data for each team. We find a positive and statistically significant effect, suggesting clubs that invest more in analytics tend to outperform competitors when controlling for roster characteristics, injuries, difficulty of schedule, and team-specific and time-specific effects. This research contributes to the body of literature affirming the value of data analytics for organizational performance and supports current investments in analytics being made by NBA teams.

Workshop on Noninvasive Glucose Monitoring 2024

2026-03-12T07:25:00Z

Workshop on Noninvasive Glucose Monitoring 2024 Kang, Jeon Woong; Arnold, Mark A; Steenkamp, Devin; Tapsak, Mark A; Mäntele, Werner; Khang, Yoonho; Jue, Miyeon; So, Peter TC This first workshop on noninvasive glucose monitoring (NIGM) was held at the Massachusetts Institute of Technology (MIT) on October 30, 2024. Six invited speakers, representing industry, academia, and clinics, gave presentations that covered (1) an overview of the NIGM technologies, (2) the state of the art in NIGM technologies, such as near-infrared (NIR), mid-infrared (IR), photoacoustic, and Raman spectroscopies, (3) minimally invasive implantable continuous glucose monitoring (CGM) sensors, and (4) a clinician’s perspective on the impact of the current CGM devices for patient care.

The community test tube of American civilization: Burt and Ethel Aginsky’s Social Science Field Laboratory, 1939–47

2026-03-12T07:25:01Z

The community test tube of American civilization: Burt and Ethel Aginsky’s Social Science Field Laboratory, 1939–47 Kapsalakis, Lauren The Social Science Field Laboratory (SSFL, 1939–47), a field school in the Ukiah Valley that trained students in social scientific and anthropological methodology, sheds light on a period in anthropology when methods were shifting from objective empiricism to meaningful participation. As analytic tools for framing the study of society failed to keep pace with social change, sociopolitical trends inside and outside anthropology situated a valley in northern California as the opportune place to gather a sample of ‘American history in vitro’. Founded by Columbia-trained anthropologists Burt and Ethel Aginsky, the SSFL responded to trends inside and outside anthropology. As the Great Depression directed anthropologists’ attention to the study of practical, modern problems in complex American communities—such as race relations, immigration, modernization, and urbanization—funding agencies strengthened the relations between sociology and anthropology and encouraged the development of interdisciplinary approaches. The Aginskys conceived of the Ukiah Valley as a ‘community test-tube of American civilization’, where scientists from all disciplines ‘can come for a convenient sample of the United States, past and present’. In teaching students how to collect data in the field, the Aginskys pierced the widely held notion that ethnographic technique cannot be taught but must be experienced by the lone individual in the field.

Exploring How Organizational Actors Experience Evaluation and Its Influence: A Q-Methodological Study

2026-03-12T07:24:59Z

Exploring How Organizational Actors Experience Evaluation and Its Influence: A Q-Methodological Study Kelly, Catherine This article contributes to research on evaluation by examining how organizational actors respond to and use evaluation imposed on them within an evaluation system. Drawing on Henry and Mark's theory of evaluation influence, this study uses Q-methodology to explore how staff within English higher education providers experience evaluation and its influence on their widening participation practice and strategy decision-making. The experiences of organizational actors are examined and classified into four types: strategic practitioners, pragmatic practitioners, staff with indirect involvement in widening participation, and evaluation enthusiasts. Through analyzing these experiences, the findings illustrate the diverse ways organizational actors are influenced by evaluation within evaluation systems. To deepen our understanding of evaluation influence in the contexts of evaluation systems, this article recommends explicitly embedding organizational theories into future theories of evaluation influence and provides suggestions for future research on the topic.

Mapping the Caregiver Experience: Predicting Dimensions of Caregiver Strain Through Task-Based Profiles

2026-03-12T07:24:55Z

Mapping the Caregiver Experience: Predicting Dimensions of Caregiver Strain Through Task-Based Profiles Brady, Samantha; Ashebir, Sophia; D’Ambrosio, Lisa; Balmuth, Alexa; Felts, Adam; Lee, Chaiwoo Objective: Family caregiving is a prevalent, diverse, and often challenging experience. We develop caregiving activity profiles to better understand how sets of care-tasks contribute to various aspects of strain. Methods: Using diary data from a survey of 213 family caregivers in the U.S., we perform latent class analysis to group commonly occurring care-related tasks into activity profiles. We then use these classifications to predict physical, financial, and emotional strain. Main Findings: We identified 4 unique activity profiles based on a set of 36 daily caregiving activities performed. Activity profiles varied significantly across the three analyzed strain dimensions. Conclusion: Activity profiles present opportunities to better understand how caregiving tasks are related to specific types kinds of caregiving strain.

Your home is not a school: The limits of homeschooling as a political practice

2026-03-12T07:24:53Z

Your home is not a school: The limits of homeschooling as a political practice Pavel, Sonia Maria; Cynamon, Jeremy Kingston Homeschooling is on the rise. It appeals to very different perspectives and ideologies that tend not to have common ground, from classical conservative to radical progressive. But the justifications for the practice are weak. In this paper, we build a case against the “home school” as a political practice using the existing commitments of liberal, conservative, and democratic theories of education. Whether education should aim at the cultivation of children's autonomy, their formation as members of cultural communities, or their training as democratic citizens, there are reasons to doubt that the practice of homeschooling can fulfill our educational goals. As such, we argue that liberals, conservatives, and democrats each have their own motivations to oppose homeschooling as an institutional alternative to traditional schools. Through our critiques, we also advance a metatheoretical argument in favor of centering the aims of education in our philosophical and political debates.

Qualitative Assessment of Terrestrial Care Settings to Inform Self-sufficient Spaceflight Medical Care

2026-03-12T07:24:52Z

Qualitative Assessment of Terrestrial Care Settings to Inform Self-sufficient Spaceflight Medical Care Porter, Allison; Arquilla, Katya; Stankovic, Aleksandra Introduction Long communication latencies in exploration spaceflight will necessitate in situ resolution to medical problems. Integrating automation into the care paradigm can address challenges posed by resource gaps inherent to spaceflight operations. However, it is not clear what aspects of exploration care are most well suited for automation integration. Methods To probe the potential role of automation in spaceflight medicine, we began by decomposing the human-automation system to first characterize the work domain(s) of the human tasks. Using the lens of point-of-care ultrasound, we leveraged existing analogous Earth medical domains to conduct in situ observations in a hospital emergency department to understand how clinicians process contextual information to provide urgent care using ultrasound and semistructured interviews with specialists to identify key procedural information components for automation. Results This investigation allowed us to characterize the dynamic system surrounding a task that does not exist in its intended—currently inaccessible—use case (ie, point-of-care ultrasound on Mars) to guide future human-automation systems development. Conclusion We conclude that specific aspects of the care environment that influence the result of a task or process (“mediating factors”) from candidate work domains call for distinct, targeted guidance for automation support and are valuable in providing system developers with tunable automation level and implementation guidelines within and/or between those work domains. Such evidence-based design practice is directly translatable to automation assistance for medical providers in resource-limited environments as well as to any situation where a person's sensory processing, perception, decision making, or response selection could be aided by automation to accomplish a task.

Solutions and Challenges for Addressing Misinformation

2026-03-12T07:24:47Z

Solutions and Challenges for Addressing Misinformation Martel, Cameron; Rand, David G Research on mitigating the effects of misinformation has contributed to the development of multiple feasible interventions designed to reduce belief in, and sharing of, falsehoods. The authors review these interventions and discuss challenges and open questions for future research. First, they provide an overview of content-neutral and content-based interventions. Next, they discuss two practical challenges to deploying and assessing these interventions in the field: scalability and pushback against content moderation efforts due to perceived political bias. Finally, they highlight several open theoretical questions and common pitfalls of research on misinformation. In particular, they argue for critical evaluation of how interventions may be effective across different types of misinformative content, different key subpopulations, and different media and environmental contexts.

Atomic Transactions

2026-03-12T13:19:44Z

Atomic Transactions Lynch, Nancy; Merritt, Michael; Weihl, William; Fekete, Alan

A design of a low-pressure steam turbine

2026-03-11T03:05:17Z

A design of a low-pressure steam turbine Jones, Bradley. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1910

Computational study of static replication for barrier options

2026-03-11T03:04:39Z

Computational study of static replication for barrier options Sun, Hai Po. Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 1997; Includes bibliographical references (leaves 75-76).

Modeling, control and experimentation of a two dimensional linear motor

2026-03-11T03:04:33Z

Modeling, control and experimentation of a two dimensional linear motor Castañeda Vega, José Israel. Thesis: M.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1997; Includes bibliographical references (leaf 118).

A study of anode dimensions in mercury-vapour thermionic rectifiers

2026-03-11T03:04:42Z

A study of anode dimensions in mercury-vapour thermionic rectifiers Fussell, Lewis. Thesis: M.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1932; Includes bibliographical references (leaf 50).

Cloud-chamber study of cosmic ray showers in lead plates

2026-03-11T03:02:21Z

Cloud-chamber study of cosmic ray showers in lead plates Fussell, Lewis. Thesis: Sc. D., Massachusetts Institute of Technology, Department of Electrical Engineering, 1938; Includes bibliographical references (leaves [113]-[118]).

Development of a high-speed light source suitable for photoelastic studies

2026-03-11T03:05:14Z

Development of a high-speed light source suitable for photoelastic studies Wyle, Frank S. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1941; Includes bibliographical references (leaf 25).

Boiling and spreading rates of instantaneous liquid methane spills on water

2026-03-11T03:04:37Z

Boiling and spreading rates of instantaneous liquid methane spills on water Chatlos, David Joseph. Thesis: M.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1982; Supervised by Robert C. Reid.; Includes bibliographical references (leaves 86-88).

Manipulation and measurement of charge transfer kinetics at chemically modified electrodes

2026-03-11T03:02:10Z

Manipulation and measurement of charge transfer kinetics at chemically modified electrodes Lewis, Nathan S. (Nathan Saul) Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemistry, 1981; Includes bibliographical references.

Application of Fourier transform spectroscopy to the absolute determination of the chemical shift of protons.

2026-03-11T03:04:45Z

Application of Fourier transform spectroscopy to the absolute determination of the chemical shift of protons. Wright, Francine Elaine. Thesis: M.S., Massachusetts Institute of Technology, Department of Physics, 1975; Vita.; Bibliography: leaves 65-66.

Double valves

2026-03-11T03:05:04Z

Double valves Faunce, Linus. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1877

Recirculation through western boundary currents varies nonlinearly with the ocean basin's aspect ratio

2026-03-11T03:08:31Z

Recirculation through western boundary currents varies nonlinearly with the ocean basin's aspect ratio Gianchandani, Kaushal Recirculation gyres adjacent to western boundary currents (WBCs) in the ocean enhance the poleward transport of these currents. While it is well-established that the WBC in a barotropic ocean strengthens with increase in basin's aspect ratio (the meridional-to-zonal extent ratio), how intensity of the recirculation through the western boundary layer varies with this parameter remains unexplored. I address this using the non-dimensional form of the nonlinear, wind-driven Stommel–Munk model of westward intensification that comprises three parameters—the aspect ratio (δ), the damping coefficient (ϵ), and the β-Rossby number (Rβ). Here, ϵ is set by the ratio of Rayleigh friction coefficient (or eddy viscosity) to the meridional gradient of the Coriolis frequency and the basin's zonal dimension, while Rβ is proportional to wind stress amplitude and quantifies the strength of nonlinearity. In the weak-to-moderate nonlinearity limit (Rβ<∼ϵ), perturbation analysis reveals that recirculation varies concavely with aspect ratio, suggesting existence of an optimal aspect ratio (δopt) for which the recirculation is maximum and for typical values of ϵ (10−3−10−2), δopt follows the power-law relation δopt=4.3ϵ. Numerical simulations further validate the existence of δopt. For large ϵ (>5×10−3), the power-law predicts δopt for the numerical solutions rather accurately, but does not hold for smaller ϵ (2×10−3) due to increased importance of nonlinear terms. Nevertheless, the nonlinear variation in recirculation through the western boundary layer with aspect ratio is observed for all ϵ values and may contribute to the heterogeneous increase in the WBC's transport across different ocean basins in a warming climate.

Streamlining Physics Problem Generation to Support Physics Teachers in Using Generative Artificial Intelligence

2026-03-11T03:08:30Z

Streamlining Physics Problem Generation to Support Physics Teachers in Using Generative Artificial Intelligence El-Adawy, Shams; Liao, Isaac; Lad, Vedang; Abdelhafez, Mohamed; Dourmashkin, Peter The rapid advancement of large language models (LLMs) presents a unique opportunity for educators to find ways to include artificial intelligence (AI) in physics course design. By critically engaging with LLMs to help with the task of generating problems, physics teachers can not only model a potentially effective way to use LLMs for other teachers, but also showcase to students ways to productively engage with LLMs. This article presents a workflow with two different starting points to generate physics problems using ChatGPT 3.5. The first initialization involves interacting with ChatGPT in a conversational manner, guiding iterative problem creation by breaking tasks into smaller tasks. The second initialization harnesses ChatGPT’s generative abilities, aligning problem generation with established problem styles by instructing the model to emulate contexts from question banks. We discuss the implications of this workflow for other physics instructors exploring productive ways to incorporate the use of AI in their own course design.

Ion optical design of the magnetic proton recoil neutron spectrometer for the SPARC tokamak

2026-03-11T03:08:29Z

Ion optical design of the magnetic proton recoil neutron spectrometer for the SPARC tokamak Mackie, S; Wink, CW; Dalla Rosa, M; Berg, GPA; Ball, JL; Wang, X; Carmichael, J; Tinguely, RA; Rigamonti, D; Tardocchi, M; Raj, P; Frenje, J; Rice, J A magnetic proton recoil (MPR) neutron spectrometer is being designed for SPARC, a high magnetic field (BT = 12 T), compact (R0 = 1.85 m, a = 0.57 m) tokamak currently under construction in Devens, MA, USA. MPR neutron spectrometers are versatile tools for making high fidelity ab initio calibrated measurements of fusion neutron flux spectra and have been used to infer fusion power, ion temperature, fuel ion ratio, and suprathermal fuel populations at several high performance fusion experiments. The performance of an MPR neutron spectrometer is in large part determined by the design of the magnetic field, which disperses and focuses recoil protons. This article details the ion optical design of a high-resolution MPR neutron spectrometer, including the amelioration of image aberrations due to nonlinear effects. An optimized design is presented that achieves ion optical energy resolution δE/E < 1% and focal plane properties that enable straightforward integration with the hodoscope detector array.

Performance predictions of the SPARC x-ray crystal spectrometers for ion temperature and toroidal rotation measurements

2026-03-11T03:08:34Z

Performance predictions of the SPARC x-ray crystal spectrometers for ion temperature and toroidal rotation measurements Perks, C; Vezinet, D; Rice, JE; Reinke, ML SPARC will be outfitted with three systems of x-ray crystal spectrometer arrays. Two of these are designed using cylindrically bent crystals to achieve high spectral-resolution for ion temperature and toroidal velocity measurements via imaging He-like Kr and Ne-like Xe. The last acts as a spectral survey system to monitor Ne-like W and nearby H- and He-like emission from Cr, Fe, Co, Ni, and Cu. Line radiation intensities are calculated using the Flexible Atomic Code for atomic data and ColRadPy for collisional-radiative modeling, then convoluted with a Voigt line shape. Free–free, free-bound, and two-photon continuum radiation is also included. The ToFu code is used to perform volume-of-sight integration to produce synthetic detector images. In addition, presented is cross-validation performed using the XICSRT Monte Carlo ray-tracing code. Ion temperature and toroidal velocity profiles are reconstructed using ToFu via tomographic inversion.

Edge scanning reflectometry for density profile measurement on the SPARC tokamak

2026-03-11T03:08:33Z

Edge scanning reflectometry for density profile measurement on the SPARC tokamak Lin, Y; Nikolaeva, V; Hachmeister, D; Kowalski, E; Reinke, ML Edge scanning reflectometry (ESRL) on the SPARC tokamak aims to measure the electron density profile from the far scrape-off layer to the top of the typical H-mode pedestal and provide real-time data for plasma control. ESRL uses a standard frequency-modulated continuous wave technique from 18 to 90 GHz. By implementing both the O-mode and left-hand-cutoff X-mode, it covers densities from ∼4 × 1018 to ∼4 × 1020 m−3 at B0 ∼12 T. A voltage-controlled oscillator acts as the frequency sweep source. Phase-locked dielectric resonator oscillators and bandpass filters generate base signals ∼9–15 GHz. The signals are then frequency multiplied and amplified to reach the K (18–26 GHz), Ka (26–40 GHz), U (40–60 GHz), and E (60–90 GHz) bands. Multi-band signals are combined via the quasi-optical technique. ESRL plans to use oversized waveguides (∼20 m one-way) and a bi-static arrangement to minimize signal losses and distortions while allowing system flexibility. A COMSOL Multiphysics RF model in 2D has been set up to simulate the reflectometry process and help decide the layout of the horn antennas. Engineering analyses of the key parts of the system have been carried out in support of its preliminary design.

Neutronics simulations for the design of neutron flux monitors in SPARC

2026-03-11T03:08:27Z

Neutronics simulations for the design of neutron flux monitors in SPARC Wang, X; Gocht, R; Ball, J; Mackie, S; Panontin, E; Tinguely, RA; Raj, P; Holmes, I; Saltos, AA; Johnson, A; Grieve, A This paper presents the development and application of high-fidelity neutronic models of the SPARC tokamak for the design of neutron flux monitors (NFM) for application during plasma operations. NFMs measure the neutron flux in the tokamak hall, which is related to fusion power via calibration. We have explored Boron-10 gamma-compensated ionization chambers (ICs) and parallel-plate Uranium-238 fission chambers (FCs). We plan for all NFMs to be located by the wall in the tokamak hall and directly exposed to neutrons streaming through a shielded opening in a midplane port. Our simulations primarily use a constructive solid geometry-based OpenMC model based on the true SPARC geometry. The OpenMC model is benchmarked against a detailed CAD-based MCNP6 model. The B10 ICs are equipped with high-density polyethylene (HDPE) sleeves, borated HDPE housings, and borated aluminum covers to shield out scattered neutrons, optimize detector response levels, and make calibration robust against changes in the tokamak hall. The B10 neutron absorption branching ratio may cause the detectors’ responses to be non-linear to neutron flux >200 keV. However, our simulations unveil that, in the SPARC environment and with the proposed housings and sleeves, >99% of the detector responses are induced by <100 keV neutrons. U238’s insensitivity to slow neutrons makes this FC a promising candidate for direct fusion neutron measurements. Along with a borated HDPE sleeve, about 60% of the FCs’ responses are induced by direct neutrons.

Image plate multi-scan response to fusion protons in the range of 1–14 MeV

2026-03-11T03:08:25Z

Image plate multi-scan response to fusion protons in the range of 1–14 MeV Vanderloo, N; Cufari, M; Russell, L; Johnson, TM; Vargas, J; Foo, BC; Buschmann, BI; Dannhoff, SG; DeVault, A; Evans, TE; Kunimune, JH; Lawrence, Y; Pearcy, JA; Reichelt, BL; Wink, CW; Gatu Johnson, M; Petrasso, RD; Frenje, JA; Li, CK Image plates (IPs) are a quickly recoverable and reusable radiation detector often used to measure proton and x-ray fluence in laser-driven experiments. Recently, IPs have been used in a proton radiography detector stack on the OMEGA laser, a diagnostic historically implemented with CR-39, or radiochromic film. The IPs used in this and other diagnostics detect charged particles, neutrons, and x-rays indiscriminately. IPs detect radiation using a photo-stimulated luminescence (PSL) material, often phosphor, in which electrons are excited to metastable states by ionizing radiation. Protons at MeV energies deposit energy deeper into the IP compared with x rays below ∼20 keV due to the Bragg peak present for protons. This property is exploited to discriminate between radiation types. Doses of mono-energetic protons between 1.7 and 14 MeV are applied to IPs using the MIT linear electrostatic ion accelerator. This paper presents the results from consecutive scans of IPs irradiated with different proton energies. The PSL ratios between subsequent scans are shown to depend on proton energy, with higher energy protons having lower PSL ratios for each scan. This finding is separate from the known energy dependence in the absolute sensitivity of IPs. The results can be compared to complimentary work on x rays, showing a difference between protons and x rays, forging a path to discriminate between proton and x-ray fluence in mixed radiation environments.

A compact and portable gamma-ray spectrometer (GRASP) for inertial confinement fusion and basic science experiments

2026-03-11T03:08:24Z

A compact and portable gamma-ray spectrometer (GRASP) for inertial confinement fusion and basic science experiments Dannhoff, SG; Wink, CW; Mackie, S; Berg, GPA; Frenje, JA A compact and portable gamma-ray spectrometer has been designed to diagnose different components of the inertial confinement fusionrelevant γ-ray spectrum with energies between ∼3.7–17.9 MeV. The system is designed to be as compact as possible for convenient transportation and fielding in diagnostic ports on the OMEGA laser, the National Ignition Facility, and other photon-source facilities. The system consists of a conversion foil for Compton scattering in front of four magnetic spectrometer “arms,” each covering a different energy range and constructed out of cylindrical permanent magnet Halbach arrays. Monte Carlo simulations have been used to optimize and assess the performance of the conversion foil, and COSY INFINITY ion-optical simulations have been used to optimize the spectrometer magnets. The performance of the design is assessed for a simulated direct-drive γ-ray spectrum. Spanning its total γ-ray energy bandwidth and using a 1.7 mm thick boron conversion foil, the system’s total energy resolution and efficiency are ∼15.8%–4.5% and 5.4 × 10−7 –3.7 × 10−7 e − /γ, respectively, with room for improvement. Spectral γ-ray measurements will provide guidance to the inertial confinement fusion program toward achieving high-energy gain relevant to inertial fusion energy and enable new measurement capabilities for basic discovery science.

Characterization of the image plate multi-scan response to mono-energetic x-rays

2026-03-11T03:08:21Z

Characterization of the image plate multi-scan response to mono-energetic x-rays Cufari, M; Vanderloo, N; Buschmann, BI; DeVault, A; Foo, BC; Vargas, J; Dannhoff, SG; Evans, TE; Johnson, TM; Kunimune, J; Lawrence, Y; Pearcy, JA; Reichelt, BL; Russell, L; Wink, CW; Gatu Johnson, M; Petrasso, RD; Frenje, JA Image plates (IPs), or phosphor storage screens, are a technology employed frequently in inertial confinement fusion (ICF) and high energy density plasma (HEDP) diagnostics because of their sensitivity to many types of radiation, including, x rays, protons, alphas, beta particles, and neutrons. Prior studies characterizing IPs are predicated on the signal level remaining below the scanner saturation threshold. Since the scanning process removes some signal from the IP via photostimulated luminescence, repeatedly scanning an IP can bring the signal level below the scanner saturation threshold. This process, in turn, raises concerns about the signal response of IPs after an arbitrary number of scans and whether such a process yields, for example, a constant ratio of signal between the nth and n + 1st scan. Here, the sensitivity of IPs is investigated when scanned multiple times. It is demonstrated that the ratio of signal decay is not a constant with the number of scans and that the signal decay depends on the x-ray energy. As such, repeatedly scanning an IP with a mixture of signal types (e.g., x ray, neutron, and protons) enables ICF and HEDP diagnostics employing IPs to better isolate a particular signal type.

Microbially-enhanced dissolution of calcite in sinking marine particles

2026-03-10T03:07:40Z

Microbially-enhanced dissolution of calcite in sinking marine particles Borer, Benedict; Subhas, Adam V.; Hayden, Matthew G.; Woosley, Ryan J.; Babbin, Andrew R. Evidence for the shallow cycling of calcium carbonate in the global ocean is mounting, but the mechanisms driving the dissolution of thermodynamically stable polymorphs, like aragonite and calcite, in the surface ocean remain unconstrained. Here, we quantify how microbial metabolism creates acidic microenvironments in marine particles that enhance the local dissolution of calcite despite supersaturated conditions in bulk waters. A temporal decoupling of particle deoxygenation and acidification suggests that respiration-derived carbon dioxide is not the sole driver of the observed undersaturation. Rapid dissolution occurs in particles exhibiting bacterial growth, with rates exceeding abiotic dissolution at the same bulk saturation by more than an order of magnitude. We observe the highest particle-associated dissolution rates at intermediate settling velocities, indicating that a trade-off between elevated mass transfer due to settling and bacterial respiration governs the ensuing dissolution rates. Translation of our experiments to the water column suggests that microbially driven undersaturation in marine particles may dissolve sufficient calcite in the mesopelagic ocean to extend particle transit times by eliminating this vital ballast mineral, reducing the efficiency of organic carbon sequestration.

Process cost analysis of performance challenges and their mitigations in sodium-ion battery cathode materials

2026-03-10T03:07:47Z

Process cost analysis of performance challenges and their mitigations in sodium-ion battery cathode materials Munjal, Mrigi; Prein, Thorben; Ramadan, Mahmoud M.; Smith, Hugh B.; Venugopal, Vineeth; Rupp, Jennifer L.M.; Abate, Iwnetim I.; Olivetti, Elsa A.; Huang, Kevin J. The success of sodium-ion batteries (SIBs) hinges on mitigating underperformance in ways that are cost effective, manufacturable, and scalable. This work investigates interfacial, morphological, and bulk interventions to enhance the performance of layered metal oxide cathode active materials (CAMs) for SIBs. We mapped the full space of literature-reported SIB CAM challenges and their mitigations. We then estimated the manufacturing costs for a diverse and representative set of mitigation approaches. Adding sacrificial salts can be cost effective, given low materials costs and minimal process changes. By contrast, many methods are reported to tune CAM morphology. Several are likely challenging at scale due to process throughput and yield limitations. Finally, bulk modifications can mitigate the moisture sensitivity of some CAMs, a likely less costly route than expanding stringent atmosphere controls during manufacturing. We end by discussing the limits and promise of process cost analysis, given the current state of battery reporting in the literature.

Policy Analytics for Cybersecurity of Cyber-Physical Systems

2026-03-06T03:00:51Z

Policy Analytics for Cybersecurity of Cyber-Physical Systems Choucri, Nazli Mounting concerns about safety and security have resulted in an intricate ecosystem system of guidelines, compliance measures, directives and policy reports for cybersecurity of all critical infrastructure. The policy paradox is that the text form of policy documents is an impediment to the implementation of policies and directives and creates potentially powerful opportunity costs. As a general practice, guidelines, directives and policy documents are presented in text form, page-by-page and word-by-word all supported by figures, diagrams and tables as needed. By definition text obscures properties of both policy and system-target in terms of dynamic relationships, feedback, “drill-down”, leads and lags, and so forth. The challenge is to develop analytics for cybersecurity policy of cyber physical systems. We begin with constructing (a) a structured system model of the system, in order to (b) identify major policydefined system-wide parameters, (c) situate system vulnerabilities, (d) map security requirements to security objectives, and (e) advance research on how system properties respond to diverse policy controls for security of cyber physical systems. This Project addresses the hard problem of policy-governed secure collaboration related to cyberphysical security of critical infrastructure (focusing on a generic and fundamental feature, namely smart grid of electric power systems). The purpose is to (a) reduce, if not eliminate barriers to full understanding of policy text as transmitted by the source, (b) explore system-wide or targeted implications, (c) help contextualize generic directives for specific applications, and (d) facilitate contingency analysis, as needed. This Compilation is based on the Quarterly Research Reports submitted by MIT to the Cyber- Physical Systems Organization of Vanderbilt University. The Compilation is the first of several Reports highlighting the research process and products of the MIT Project on Policy Analytics for Cybersecurity of Cyber-Physical Systems. Gaurav Agarwal [a.k.a. Gaurav], MIT alumnus, served as Lead Researcher for the Proof-of-Concept case presented here.

Summary of the Fourth ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2016 April 3-8

2026-03-06T03:01:04Z

Summary of the Fourth ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2016 April 3-8 Matthews, Lynn D.; Crew, G. B.; Fish, Vincent The primary objectives of the fourth APP CSV campaign were twofold: (1) to execute Very Long Baseline Interferometry (VLBI) observing mode (VOM) observations using Schedule Blocks (SBs); (2) to carry out the first end-to-end testing of intercontinental VLBI sessions in both Bands 3 and 6. While intercontinental VLBI fringes with ALMA have already been obtained during previous CSV campaigns (Matthews & Crew 2016b, c), those sessions were not conducted in a manner identical to future VLBI science campaigns (i.e., they used manual execution of observing commands rather than SBs) and did not involve observations of a full suite of ALMA and VLBI calibrators. Secondary objectives of the fourth CSV mission included further development work on an ALMA Phasing System (APS) graphical user interface (GUI), additional testing of the fast phasing loop (under a wider variety of weather conditions), tests of the phasing system in Band 7 (in support of an ongoing North America ALMA Study award), and the training of ALMA staﬀ in the operation of the VOM and the APS hardware and software. This is a report of the activities of this campaign. This report was prepared to report on APP commissioning progress, provided as ALMA Technical Note #19.

Summary of the Third ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 July 28-August 3

2026-03-06T03:01:12Z

Summary of the Third ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 July 28-August 3 Matthews, Lynn D.; Crew, G. B. The primary objective for the third APP CSV campaign was to perform intercontinental Very Long Baseline Interferometry (VLBI) fringe tests between phased ALMA and remote stations in Band 3, Band 6, and (conditions permitting) Band 7. Secondary objectives included preparations for making the APS available to the community for ALMA Cycle 4 and training ALMA staff in the operation of the ALMA Phasing System (APS). This is a report of those activities. This report was prepared to report on APP commissioning progress, provided as ALMA Technical Note #18.

Summary of the Second ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 March 24-30

2026-03-06T03:01:14Z

Summary of the Second ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 March 24-30 Matthews, Lynn D.; Crew, G. B. The primary objective for the second APP CSV campaign was to test and characterize the phasing system, including the recent changes in the handling of the front-end delays. Secondary goals were to repeat the local VLBI test between ALMA and an antenna at the Operations Support Facility (OSF) (also attempted during the January mission) and to obtain a short VLBI recording on a calibrator source with ALMA and one or more stations operating as part of the Event Horizon Telescope (EHT) network, thus allowing demonstration of an intercontinental VLBI fringe. This is a report of the week's activities. This report was prepared to report on APP commissioning progress, provided as ALMA Technical Note #17.

Summary of the First ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 January 6-13

2026-03-06T03:01:08Z

Summary of the First ALMA Phasing Project (APP) Commissioning and Science Verification Mission: 2015 January 6-13 Matthews, Lynn D.; Crew, G. B. The first Commissioning and Science Verification (CSV) mission for the ALMA Phasing Project (APP) took place during the ALMA EOC Week from 2015 January 6-13. The formal commencement of APP CSV activities followed the provisional acceptance of the APP hardware during a formal review by JAO that took place on 2014 December 11. This is a report of activities during the week. This report was prepared to report on APP commissioning progress, provided as ALMA Technical Note #16.

ALMA North America Cycle 4 Study Project Final Report: Diversifying the Applications of the ALMA Phasing System

2026-03-06T03:01:06Z

ALMA North America Cycle 4 Study Project Final Report: Diversifying the Applications of the ALMA Phasing System Matthews, Lynn D.; Crew, G.; Hecht, M. H. The Atacama Millimeter/submillimeter Array (ALMA) Phasing Project (APP) produced the hardware and software modifications necessary to bring Very Long Baseline Interferometry (VLBI) capabilities to ALMA. The resulting VLBI observing mode was introduced to the science community in ALMA Cycle 4 (2017), and two VLBI science campaigns have now been carried out successfully at ALMA. The current Cycle 4 ALMA North America (NA) Study was proposed to lay the groundwork for a variety of enhancements to the ALMA Phasing System (APS) that were not within the scope of the original APP project. These include: (1) devising an improved method for the handling of baseband delays; (2) development of procedures for use of the APS on fainter astronomical sources than is presently possible; (3) development of data acquisition and correlation techniques to allow the APS to be used for spectral line VLBI experiments. These tasks were intended as preparatory steps for a future full-scale implementation project (if approved). Formal approval of this implementation work has now been granted and is funded through an ALMA Cycle 5 NA Development project known as APP “Phase 2” (APP-2). As a result, efforts to implement capabilities designed and explored under the current Cycle 4 Study, as well as a previous Cycle 3 Study award, are now underway. This report provides a status summary of Cycle 4 activities and outlines follow-on work that is continuing as part of the ongoing Cycle 5 Development efforts.

ALMA North America Cycle 3 Study Project Final Report: Extensions and Enhancements to the ALMA Phasing System

2026-03-06T03:01:09Z

ALMA North America Cycle 3 Study Project Final Report: Extensions and Enhancements to the ALMA Phasing System Matthews, L.; Crew, G.; Hecht, M. H. The Atacama Millimeter/submillimeter Array (ALMA) Phasing Project (APP) has successfully brought Very Long Baseline Interferometry (VLBI) to ALMA. Nine VLBI science projects were observed in 2017 during ALMA’s in augural VLBI campaign as part of Cycle 4. This marked the culmination of an international 5-year effort that involved both hardware and software contribu- tions from the APP Team to the ALMA Observatory. A Cycle 3 ALMA North America (NA) Study was proposed to enable ongoing support of VLBI at ALMA and the investigation of enhancements to the ALMA Phasing System (APS) that were not within the scope of the original APP project. These included: (1) an extension of phasing capabilities to the submilleter (Band 7); (2) an exploration of correlation techniques to compensate for the mismatch in sampling rates between ALMA and other VLBI stations; (3) prescriptions for optimization of ALMA baseband delay application; (4) defining and documenting data calibration and analysis pathways for experiments utilizing phased ALMA data. This report summarizes outcomes from the Cycle 3 Study. Work on the APP remains ongoing under a Cycle 4 Study award and will continue under a pending ALMA NA Cycle 5 Development Project that is expected to enable full implementation of the capabilities explored under the Cycle 3 and Cycle 4 Studies.

Pulsars, Magnetars, and Transients with Phased ALMA, Final Report

2026-03-06T03:01:05Z

Pulsars, Magnetars, and Transients with Phased ALMA, Final Report Cordes, James; Blackburn, Lindy; Chatterjee, Shami; Crew, Geoffrey; Devignes, Gregory; Doeleman, Shep; Kramer, Michael; Lazio, Joe; Liu, Kuo; Ransom, Scott The present study developed fast time-domain capability for the ALMA phased-array system that is needed for observations of compact objects in the Galactic center and elsewhere in the Galaxy. ALMA can provide unparalleled sensitivity to a spectral region that has been poorly explored for neutron stars. Observations at mm and sub-mm wavelengths have the potential for providing decisive observational constraints on emission processes from the magnetospheres of neutron stars. ALMA is also important for surveys for pulsars and transients in the Galactic center. This is a final report of the work performed.

Final Report: ALMA Phasing Project Augmentation

2026-03-06T03:01:10Z

Final Report: ALMA Phasing Project Augmentation Matthews, Lynn D. This report provides a summary of activities carried out under the NA ALMA Development Fund award to augment the National Science Foundation MRI award for the ALMA Beamformer proposal, performed as the ALMA Phasing Project.

MRI: Development of an ALMA Beamformer for Ultra High Resolution VLBI and High Frequency Phased Array Science

2026-03-06T03:01:11Z

MRI: Development of an ALMA Beamformer for Ultra High Resolution VLBI and High Frequency Phased Array Science Matthews, Lynn D. This is the final report covering the activities of the National Science Foundation MRI Award number 1126433.

Cycle 11 VLBI Acceptance: Delay Fix Final Report

2026-03-06T03:01:07Z

Cycle 11 VLBI Acceptance: Delay Fix Final Report Crew, Geoffrey B.; Matthews, Lynn D. This report presents details of changes to the VLBI software system made on the path to the Cycle 11 Acceptance. The principal new feature is the long-awaited "delay fix" to the APS that is presented in considerable detail here. This was proposed for a new ADF implementation project, APP2. The delay fix is a technology that allows the full 2-GHz continuum band to be used in active phasing, resulting results in a lower flux density limit for direct observation of science targets. It also allows a greater range of passive phasing targets to support weaker targets. Work on the delay fix began during Cycle 3 and is only now concluded in time for additional testing as desired in Cycle 11 and then full use in Cycle 12. Other than the delay fix, a new software device was delivered to support the new hydrogen maser, and there was the usual round of minor scripting updates in the SSR component. This report is written to present a largely self-contained history of the delay fix effort and its fruit, and it also covers the other, minor development items which are formally part of the Cycle 11 Acceptance. As the software phasing engine which supports this mode is also now the recommended engine for the APS, this document also serves to present relatively complete documentation on the final implementation of the SSR as well as TelCal sides of the APS. The underlying VOM remains as it was deployed and used in Cycle 4.

ObsMode2022 Cycle 10 Go/No-Go Report for VLBI Capabilities

2026-03-06T03:01:13Z

ObsMode2022 Cycle 10 Go/No-Go Report for VLBI Capabilities Matthews, Lynn D.; Crew, Geoff; Fish, Vincent; Messias, Hugo; Titus, Mike; Krichbaum, Thomas We present here a report on the Very Long Baseline Interferometry (VLBI) development efforts under consideration at ALMA as new offerings in Cycle 10. These activities are being carried out under the ALMA North America Development Project known as the ALMA Phasing Project Phase 3 (APP3). The two VLBI priorities previously identified for Cycle 10 by the ObsMode process are: (1) spectral line VLBI with flexible tuning; (2) panchromatic VLBI for spectral line and continuum (Bands 1, 3, 6 and 7, with provisions for extension to any other band). This document provides an overview of the development, testing, and readiness of these capabilities. Some updates on minor development efforts are also provided. This report was prepared for the formal acceptance of the software required for Cycle 10. Notionally it is ALMA Technical Note #25, but not published (yet) as such.

Cycle 8 (2021–2022) VLBI Delta Acceptance Report

2026-03-06T03:01:03Z

Cycle 8 (2021–2022) VLBI Delta Acceptance Report Crew, Geoff; Vila-Vilaro, Baltasar This report summarizes the acceptance process for VLBI which was carried out in 2021 as part of the normal Cycle 8 Acceptance and later, through the 2021–2022 preparations for the 2022 VLBI Campaigns. It reviews the hardware setup and checks that must be made at various times prior to any observations with VLBI peers. Then there is a suite of offline tests of the SB-generation that include observation simulation. There is also a suite of on-sky regression tests that exercise the ALMA Phasing System (APS). The final step of the acceptance has historically been the execution of a short “dress rehearsal” (DR) usually in January of the cycle year with the Event Horizon Telescope (EHT) to provide end-to-end validation of the system via fringes to remote peers. Collectively these tests establish that the science VLBI projects may proceed without issue. This report was prepared for the formal acceptance of the software required for Cycle 8. Notionally, it is ALMA Technical Note #24, but not published (yet) as such.

Design of a diamond-based in-vessel soft x-ray detector for the SPARC tokamak

2026-03-06T03:09:00Z

Design of a diamond-based in-vessel soft x-ray detector for the SPARC tokamak Normile, S; Vezinet, D; Perks, C; Bombarda, F; Verona-Rinati, G; Rice, JE; Verona, C; Raso, AM; Angelone, M The in-vessel silicon diode arrays that are used for soft x-ray detection in many tokamaks are sensitive to neutron damage, making them unsuitable for burning plasma devices such as SPARC. In such a device, the silicon diodes would need to be placed far from the plasma—limiting their field of view—or an alternative detector could be used. Here, we present the design of a camera containing an array of chemical vapor deposition single-crystal diamonds, which will be placed in the upper and lower port plugs of the SPARC tokamak with a large enough view of the poloidal cross section to enable tomographic inversion. The camera design presented here is optimized to provide a wide field of view of the poloidal cross section. Simulated plasma conditions are used to estimate the x-ray signal that this detector array will receive and to fine-tune the camera placement within the tokamak.

Development of the prototype for the SPARC hard X-ray monitor

2026-03-06T03:08:57Z

Development of the prototype for the SPARC hard X-ray monitor Panontin, E; Tinguely, RA; Hartwig, ZS; Saltos, AA; Vezinet, D; Rice, J The SPARC tokamak will be equipped with a hard X-ray (HXR) monitor system capable of measuring the bremsstrahlung emission from runaway electrons with photon energies in excess of about 100 keV. This diagnostic will detect the formation of runaway electron beams during plasma start-up and inform the plasma control system to terminate the discharge early to protect the machine. In this work, we present a 0D estimate of the HXR emission in SPARC during plasma start-up. Then we discuss the characterization of a prototype of the HXR monitor. The detector mounts a 1 × 1-in.2 LaBr3 inorganic scintillator coupled with a photomultiplier tube and has been tested with γ-ray sources to find its dynamic range. Finally, two possible modes of operation for spectroscopic and current mode measurements on SPARC are proposed.

Perspectives on pilot-wave hydrodynamics

2026-03-06T03:08:58Z

Perspectives on pilot-wave hydrodynamics Bush, John WM; Frumkin, Valeri; Sáenz, Pedro J We present a number of fresh perspectives on pilot-wave hydrodynamics, the field initiated in 2005 by Couder and Fort's discovery that millimetric droplets self-propelling along the surface of a vibrating bath can capture certain features of quantum systems. A recurring theme will be that pilot-wave hydrodynamics furnishes a classical framework for reproducing many quantum phenomena and allows one to rationalize such phenomena mechanistically, from a local realist perspective, obviating the need to appeal to quantum nonlocality. The distinction is drawn between hydrodynamic pilot-wave theory and its quantum counterparts, Bohmian mechanics, the Bohm–Vigier stochastic pilot-wave theory, and de Broglie's theory of the double-solution. Each of these quantum predecessors provide a valuable touchstone as we take the physical picture engendered in the walking droplets and extend it into the quantum realm via theoretical modeling. Emphasis is given to recent developments in the field, both experimental and conceptual, and to forecasting potentially fruitful new directions.

Automated transient grating spectroscopy mapping and signal control for large samples

2026-03-06T03:08:55Z

Automated transient grating spectroscopy mapping and signal control for large samples Weaver, Colin; Stapelberg, Myles; Short, Michael P; Wylie, Angus; Artalejo, Elena Botica We present developments for the mapping of large areas using transient grating spectroscopy (TGS) that allow for smoother, larger, autonomous measurements of material samples. The addition of a precise linear stage in the direction parallel to laser sampling coupled with signal optimizing control allows for hands free, self-correcting measurements. In addition, the simplification of the sample holding design to a form that is small enough to mount directly to the linear stage exhibits a straightforward, low-cost solution for automated TGS applications. This capability is demonstrated by taking large uninterrupted maps of gradient wafers, and the results are validated on calibrated tungsten samples and control TGS samples from gradient wafers.

Manipulating the duration of picoinjection controls the injected volume of individual droplets

2026-03-06T03:09:01Z

Manipulating the duration of picoinjection controls the injected volume of individual droplets Thakur, R.; Weitz, D. The ability to add reagents into droplets is required in many microfluidic workflows. Picoinjection can address this need; however, it is unable to control the injection volume for each individual droplet. Here, we present an improved picoinjection method that can inject controlled volumes into individual droplets. We achieve this by adjusting the injection duration for each picoinjection event. This improved picoinjection method can be used to create complex microfluidic workflows that are able to control the biochemical composition of individual droplets.

Multitask methods for predicting molecular properties from heterogeneous data

2026-03-06T03:08:49Z

Multitask methods for predicting molecular properties from heterogeneous data Fisher, KE; Herbst, MF; Marzouk, YM Data generation remains a bottleneck in training surrogate models to predict molecular properties. We demonstrate that multitask Gaussian process regression overcomes this limitation by leveraging both expensive and cheap data sources. In particular, we consider training sets constructed from coupled-cluster (CC) and density functional theory (DFT) data. We report that multitask surrogates can predict at CC-level accuracy with a reduction in data generation cost by over an order of magnitude. Of note, our approach allows the training set to include DFT data generated by a heterogeneous mix of exchange–correlation functionals without imposing any artificial hierarchy on functional accuracy. More generally, the multitask framework can accommodate a wider range of training set structures—including the full disparity between the different levels of fidelity—than existing kernel approaches based on Δ-learning although we show that the accuracy of the two approaches can be similar. Consequently, multitask regression can be a tool for reducing data generation costs even further by opportunistically exploiting existing data sources.

7.344 Cellular Metabolism and Cancer: Nature or Nurture?, Fall 2018

2026-03-09T18:17:48Z

7.344 Cellular Metabolism and Cancer: Nature or Nurture?, Fall 2018 Lau, Allison; Lien, Evan In this course we will explore how altered metabolism drives cancer progression. Students will learn (1) how to read, discuss, and critically evaluate scientific findings in the primary research literature, (2) how scientists experimentally approach fundamental issues in biology and medicine, (3) how recent findings have challenged the traditional “textbook” understanding of metabolism and given us new insight into cancer, and (4) how a local pharmaceutical company is developing therapeutics to target cancer metabolism in an effort to revolutionize cancer therapy.

7.343 Single-Molecule Imaging: Capturing Nanoscale Cellular Machines in Action, Fall 2021

2026-03-09T18:16:55Z

7.343 Single-Molecule Imaging: Capturing Nanoscale Cellular Machines in Action, Fall 2021 Kose, Hazal B. Did you know that we have approximately 2 meters of DNA packed in our cells, which are less than 10 μm diameter? Or that to replicate DNA it is copied at a rate of 70,000 basepairs per second by a cellular apparatus that coordinates at least six different enzymes? Or that microtubules form greater than 1 meter long “railways” upon which molecular machines transport cargo within nerve cells? In this course, we will explore how single-molecule imaging techniques capture the mega-cellular machines working in real-time. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.342 Immune Cell Migration: On the Move in Response to Pathogens and Cancer Immunotherapy, Fall 2021

2026-03-09T18:16:00Z

7.342 Immune Cell Migration: On the Move in Response to Pathogens and Cancer Immunotherapy, Fall 2021 Fessenden, Timothy The mammalian immune system is sometimes called a “liquid organ,” capable of rapidly initiating and then resolving potent responses to pathogens at almost any location in the organism. What protein machinery drives immune cells’ rapid migration? How do cells make pathfinding decisions around barriers? How do they find rare pathogens or target cells in complex environments? This course will begin by examining the general immunological functions of two major immune cell types—T cells and dendritic cells. Through our readings and discussions, we will examine the connections between immunotherapy as an emerging treatment modality for a variety of cancers and the migration of immune cells. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.341 Turning Evolutionary Dials: Directed Evolution Techniques for Climate Change and Beyond, Spring 2022

2026-03-09T18:15:19Z

7.341 Turning Evolutionary Dials: Directed Evolution Techniques for Climate Change and Beyond, Spring 2022 Kizer, Megan; Wilson, Robbie This course will cover the many ways in which we have realized evolution in the laboratory toward functional biomolecules, such as protein and nucleic-acid-based therapeutics, enzymes that catalyze production of synthetic drugs, and carbon-dioxide capture molecules to lessen the impact of climate change. Students will both become familiar with the field of directed molecular evolution and learn how to critically analyze primary research papers, design research experiments, and present data relating to molecular biology and evolution. The importance of directed evolution in biomedical and biotechnological careers, both academic and industrial, will be highlighted. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.342 Synapse Remodeling in Health and Disease, Fall 2022

2026-03-09T18:14:38Z

7.342 Synapse Remodeling in Health and Disease, Fall 2022 Ordonez, Dalila; Boivin, Josiah Our brains are remarkably adaptable throughout our lives. Individual brain cells called neurons form synapses, sites of physical connection and communication between neurons, and then repeatedly rewire those connections in response to new experiences or to neuronal cell death caused by injury, disease, or aging. In this course, we will explore how neurons establish their synapses in the healthy brain during childhood and later in life, and how this process goes awry in disease states. More specifically, we will discuss how the brain forms its synapses early in life, stabilizes a subset of those synapses for long-term maintenance, and continues to add and remove synapses throughout life. We will then explore synapse dysfunction in diseases such as autism and Alzheimer’s disease, which involve abnormal increases or losses of synaptic connections, respectively. We will also consider synapse remodeling, a process of adding and removing synaptic connections to optimize our brain network, in the context of neuroinflammation, recovery from traumatic brain injury, and psychological trauma following prolonged stress. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.342 How To Build An Animal: Cell Fate and Identity in Development and Disease, Fall 2017

2026-03-09T18:14:04Z

7.342 How To Build An Animal: Cell Fate and Identity in Development and Disease, Fall 2017 Blanton, Laura V In this course, we will explore how animals determine and maintain cell fate. We will discuss changes to DNA structure and packaging, special proteins (known as "master regulators") with the ability to alter cell fate via transcription, cell-cell signaling, and RNA localization. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.341 DNA's Sister Does All the Work: The Central Roles of RNA in Gene Expression , Spring 2019

2026-03-09T18:13:29Z

7.341 DNA's Sister Does All the Work: The Central Roles of RNA in Gene Expression , Spring 2019 Fiszbein, Ana; Jens, Marvin This course will explore the current frontiers of the world of RNA biology with primary research papers to trace how the original odd detail sometimes leads to major discoveries. As we discuss the different transcripts and processing events that enable this exciting diversity of RNA functions, we invite you to read landmark papers with us, think critically, and ask new questions, as we marvel at the wonders of RNA. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.341 The Microbiome and Drug Delivery: Cross-species Communication in Health and Disease, Spring 2018

2026-03-09T18:12:34Z

7.341 The Microbiome and Drug Delivery: Cross-species Communication in Health and Disease, Spring 2018 Beyzavi, Ali; Jimenez, Miguel There are more microbes permanently living in our gut than there are cells in the human body. This rich community of bacteria, fungi and viruses, called the microbiome, plays a central role in human health and disease. Recent research has linked this passenger community to nutrition, circadian rhythms, infectious disease, inflammatory disease, cancer, diabetes, arthritis and even immune system and nervous system development. How can we analyze such a complex system? Can we exploit the microbiome to improve human health? Can interactions with microbes be harnessed for drug delivery? In this course, we will learn to critically assess the primary scientific literature to find answers to these questions and learn to distinguish between correlation and causality. We will learn how mechanistic insights and emerging tools, such as synthetic biology and microfluidics, together are transforming microbiome research, and might lead to new types of therapeutics and drug delivery for improving human health. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.341 Microbes at War: The Mechanisms That Drive Infectious Diseases, Fall 2022

2026-03-09T18:11:52Z

7.341 Microbes at War: The Mechanisms That Drive Infectious Diseases, Fall 2022 McLellan, Lisa How can a tick bite cause a meat allergy? And does cranberry juice do anything to help cure a urinary tract infection? To answer these and other questions, we are going to take a dive into the molecular world of microbes. In this class, we will use the primary research literature to explore the molecular interactions between pathogens and their hosts that allow microbes to cause infectious diseases. We will examine the factors that pathogens use to colonize a host and how the host response can impact the outcome of the infection. By the end of the class, students will have both developed critical scientific skills in evaluating scientific literature and an appreciation of the microbes influencing our lives and health every day. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.341 Biomaterials and Devices for Disease Diagnosis and Therapy, Fall 2018

2026-03-09T18:11:08Z

7.341 Biomaterials and Devices for Disease Diagnosis and Therapy, Fall 2018 McHugh, Kevin; Beyzavi, Ali Students will learn about the use of biomaterials to create advanced diagnostic tools for detection of infectious and chronic diseases, restore insulin production to supplement lost pancreatic function in diabetes, provide cells with appropriate physical, mechanical, and biochemical cues to direct tissue regeneration, and enhance the efficacy of cancer immunotherapy. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.342 The Seeds and the Soil: Roles of Tumor Heterogeneity and the Tumor Microenvironment in Cancer Metastasis, Fall 2020

2026-03-09T18:09:46Z

7.342 The Seeds and the Soil: Roles of Tumor Heterogeneity and the Tumor Microenvironment in Cancer Metastasis, Fall 2020 Lambert, Arthur; Zhang, Yun Metastatic disease is responsible for the vast majority of deaths associated with cancer, yet our understanding of how metastases arise is still developing. In this course, we will introduce various concepts and models that have been proposed to explain how cancer cells disseminate from a primary tumor to distant anatomical sites. We’ll learn about the critical factors that influence cancer metastasis frontiers through analysis and discussion of relevant primary research articles, with an emphasis on mechanisms of metastasis that can be applied across different cancer types. Students will gain a broad understanding of the field of cancer metastasis, including state-of-the-art techniques that are being used to address pressing questions in the field. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.343 Microbial Megaproducers: Discovery, Biosynthesis, Engineering and Applications of Natural Products, Fall 2020

2026-03-09T18:08:41Z

7.343 Microbial Megaproducers: Discovery, Biosynthesis, Engineering and Applications of Natural Products, Fall 2020 Ulrich, Emily C; Hetrick, Kenton The natural world is a mega-factory of small molecules, peptides, fatty acids, phospholipids, and a host of other compounds, known as natural products (NPs). Immensely diverse in structure and function, NPs have strongly influenced how we treat infectious disease, cancer, pain, and a host of other conditions. Roughly half of the drugs that have been approved in the past 30 years are NPs, derivatives of NPs or NP-inspired. In this discussion-based course, we will delve into research on discovering NPs from producing organisms, investigating the biochemistry of NP production, and using synthetic biology to create NP derivatives—all with a particular emphasis on how genomic data guides and informs all these studies. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

7.342 A Double-Edged Sword: Cellular Immunity in Health and Disease, Fall 2018

2026-03-09T18:09:08Z

7.342 A Double-Edged Sword: Cellular Immunity in Health and Disease, Fall 2018 Ma, Haiting Immune cells protect our bodies from both self-derived threats and exogenous pathogens, while keeping peace with normal cells and non-harmful commensal microbiota. They have various mechanisms to perform these tasks, a capacity that is essential for maintaining homeostasis. However, these same mechanisms can backfire, resulting in severe disorders such as immunodeficiency, chronic inflammation, allergy, degenerative diseases, and cancer. This course discusses the connections between normal physiology and disease by examining the developmental relationship between innate and adaptive immune cells as well as the functions and malfunctions of immune cells. The course familiarizes students with both basic biological principles (such as cell death and immune cell signaling) and clinical applications (such as immune checkpoint blockade). More generally, students learn to identify relevant primary research literature, critically evaluate experimental data, and reach their own conclusions based on primary data. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

Automated Flow Synthesis of Artificial Heme Enzymes for Enantiodivergent Biocatalysis

2026-03-05T06:13:21Z

Automated Flow Synthesis of Artificial Heme Enzymes for Enantiodivergent Biocatalysis Fittolani, Giulio; Kutateladze, Dennis A; Loas, Andrei; Buchwald, Stephen L; Pentelute, Bradley L The remarkable efficiency with which enzymes catalyze small-molecule reactions has driven their widespread application in organic chemistry. Here, we employ automated fast-flow solid-phase synthesis to access catalytically active full-length enzymes without restrictions on the number and structure of noncanonical amino acids incorporated. We demonstrate the total syntheses of iron-dependent Bacillus subtilis myoglobin (BsMb) and sperm whale myoglobin (SwMb). The synthetic enzymes displayed excellent enantioselectivity and yield in carbene transfer reactions. Absolute control over enantioselectivity in styrene cyclopropanation was achieved using synthetic L- and D-BsMb mutants, which delivered each enantiomer of cyclopropane product in identical and opposite enantiomeric enrichment. BsMb mutants outfitted with noncanonical amino acids were used to facilitate detailed structure–activity relationship studies, revealing a previously unrecognized hydrogen-bonding interaction as the primary driver of enantioselectivity in styrene cyclopropanation. We anticipate that our approach will advance biocatalysis by providing reliable and rapid access to fully synthetic enzymes possessing noncanonical amino acids.

Development of a Ligand for Cu-Catalyzed Amination of Base-Sensitive (Hetero)aryl Chlorides

2026-03-05T06:13:11Z

Development of a Ligand for Cu-Catalyzed Amination of Base-Sensitive (Hetero)aryl Chlorides Ai, Han-Jun; Mai, Binh Khanh; Liu, Cecilia; Liu, Peng; Buchwald, Stephen L We report a new N1,N2-diarylbenzene-1,2-diamine ligand, L6, that supports a copper catalyst capable of coupling base-sensitive aryl chlorides and amines that were previously unsuccessful substrates for Cu-catalyzed C–N coupling. A detailed structure–activity relationship study, combined with density functional theory (DFT) calculations, was used to uncover two key structural features that contribute to the efficacy of the catalyst derived from L6. First, steric repulsion caused by a methyl substituent induces a conformational change that opens up additional space for ligand deprotonation and oxidative addition. Second, the trifluoromethyl groups create electrostatic interactions between the ligand and aryl chloride substrates that facilitate oxidative addition via through-space ligand–substrate interaction.

Ligand Design Enables Cu-Catalyzed Etherification of Aryl Bromides Using Mild Bases

2026-03-05T06:13:20Z

Ligand Design Enables Cu-Catalyzed Etherification of Aryl Bromides Using Mild Bases Strauss, Michael J; Greaves, Megan E; Kim, Seoung-Tae; Schmidt, Michael A; Scola, Paul M; Buchwald, Stephen L We report a Cu-catalyzed method for the efficient coupling of base-sensitive aryl bromides and alcohols utilizing a newly developed N1,N2-diarylbenzene-1,2-diamine ligand, L15. This ligand was developed to increase the Lewis acidity of the Cu center, thereby permitting the use of a substantially milder base (NaOTMS or NaOPh) relative to those required in a previous iteration of this methodology (NaOMe or NaOt-Bu). Under the optimized reaction conditions, several classes of previously incompatible aryl bromides were efficiently transformed, including base-sensitive heterocycles and those containing acidic functional groups. Kinetic analyses support that C–O coupling proceeds via a mechanism involving binding/deprotonation of alcohol nucleophiles, that the pKa of the base influences the overall rate law, and that substoichiometric quantities of strong base can be utilized to accelerate ligand activation and thereby increase the overall rate of the transformation.

Time-Resolved Line Shapes of Single Quantum Emitters via Machine Learned Photon Correlations

2026-03-05T06:13:24Z

Time-Resolved Line Shapes of Single Quantum Emitters via Machine Learned Photon Correlations Proppe, Andrew H; Lee, Kin Long Kelvin; Kaplan, Alexander EK; Ginterseder, Matthias; Krajewska, Chantalle J; Bawendi, Moungi G Solid-state single-photon emitters (SPEs) are quantum light sources that combine atomlike optical properties with solid-state integration and fabrication capabilities. SPEs are hindered by spectral diffusion, where the emitter's surrounding environment induces random energy fluctuations. Timescales of spectral diffusion span nanoseconds to minutes and require probing single emitters to remove ensemble averaging. Photon correlation Fourier spectroscopy (PCFS) can be used to measure time-resolved single emitter line shapes, but is hindered by poor signal-to-noise ratio in the measured correlation functions at early times due to low photon counts. Here, we develop a framework to simulate PCFS correlation functions directly from diffusing spectra that match well with experimental data for single colloidal quantum dots. We use these simulated datasets to train a deep ensemble autoencoder machine learning model that outputs accurate, noiseless, and probabilistic reconstructions of the noisy correlations. Using this model, we obtain reconstructed time-resolved single dot emission line shapes at timescales as low as 10 ns, which are otherwise completely obscured by noise. This enables PCFS to extract optical coherence times on the same timescales as Hong-Ou-Mandel two-photon interference, but with the advantage of providing spectral information in addition to estimates of photon indistinguishability. Our machine learning approach is broadly applicable to different photon correlation spectroscopy techniques and SPE systems, offering an enhanced tool for probing single emitter line shapes on previously inaccessible timescales.

Uncovering temperature-dependent exciton-polariton relaxation mechanisms in hybrid organic-inorganic perovskites

2026-03-05T06:13:18Z

Uncovering temperature-dependent exciton-polariton relaxation mechanisms in hybrid organic-inorganic perovskites Laitz, Madeleine; Kaplan, Alexander EK; Deschamps, Jude; Barotov, Ulugbek; Proppe, Andrew H; García-Benito, Inés; Osherov, Anna; Grancini, Giulia; deQuilettes, Dane W; Nelson, Keith A; Bawendi, Moungi G; Bulović, Vladimir Hybrid perovskites have emerged as a promising material candidate for exciton-polariton (polariton) optoelectronics. Thermodynamically, lowthreshold Bose-Einstein condensation requires efficient scattering to the polariton energy dispersion minimum, and many applications demand precise control of polariton interactions. Thus far, the primary mechanisms by which polaritons relax in perovskites remains unclear. In this work, we perform temperature-dependent measurements of polaritons in low-dimensional perovskite wedged microcavities achieving a Rabi splitting of _ΩRabi = 260 ± 5 meV. We change the Hopfield coefficients by moving the optical excitation along the cavity wedge and thus tune the strength of the primary polariton relaxation mechanisms in this material. We observe the polariton bottleneck regime and show that it can be overcome by harnessing the interplay between the different excitonic species whose corresponding dynamics are modified by strong coupling. This work provides an understanding of polariton relaxation in perovskites benefiting from efficient, material-specific relaxation pathways and intracavity pumping schemes from thermally brightened excitonic species. Springer Science and Business Media LLC

ObsMode2021 Cycle 9 Go/No-Go Report for VLBI Capabilities

2026-03-05T18:45:41Z

ObsMode2021 Cycle 9 Go/No-Go Report for VLBI Capabilities Matthews, Lynn D.; Crew, Geoff; Goddi, Ciriaco; Marti-Vidal, Ivan; Titus, Mike; Fish, Vincent; Wagner, Jan; Rottmann, Helge; Pridiprihora, Yurii; Krichbaum, Thomas; Liu, Kuo; Kramer, Michael We present a status summary of the primary Very Long Baseline Interferometry (VLBI) development efforts which are under consideration at ALMA as new offerings in Cycle 9. These activities are being carried out under the ALMA North America Development Project known as the ALMA Phasing Project Phase 2 (APP2). The two VLBI priorities previously identified for Cycle 9 by the ObsMode process are: (1) a submillimeter (Band 7) VLBI observing capability and (2) a prototype spectral line VLBI mode (in Band 3 only). This document provides an overview of the development, testing, and readiness of these capabilities. Updates on other APP2 development efforts of relevance for future cycles, including the phased-array mode offered for the first time in Cycle 8, are also provided. Warning: This report contains material that is considered proprietary to the Event Horizon Telescope Collaboration (EHTC). These results may not be publicly posted, cited, or shared in any form. They are presented here with permission from EHTC Management solely for the purpose of allowing an evaluation of ALMA's performance as a VLBI station for 345 GHz (Band 7) VLBI. This report also contains ALMA EOC (test) data which are subject to ALMA's standard policies on the use of such data. We are working to ensure that the EHTC follows the ALMA guidelines (Carpenter et al., 2019) in the appropriate publication and release of data to allow follow-up work for science. This report was prepared for the formal acceptance of the software required for Cycle 9. Notionally it is ALMA Technical Note #23, but not published (yet) as such.

A Self-Directed, Home-Like XR System for Sustained Intangible Cultural Heritage Practice: An Ikebana Case Study

2026-03-05T06:13:11Z

A Self-Directed, Home-Like XR System for Sustained Intangible Cultural Heritage Practice: An Ikebana Case Study Wu, Yu; Li, Manxueying; Mai, Gelei Sustained Intangible Cultural Heritage (ICH) practices for novices depend more on curiosity and creative agency than on procedural training. Yet, most extended reality (XR) systems for ICH emphasize guided instruction or exhibitions, limiting self-direction and continuity beyond the device. Using Ikebana as a case study, we present a self-directed, home-like virtual reality (VR) experience built with 3D Gaussian Splatting (3D GS) and natural hand tracking, complemented by an augmented reality (AR) revisiting feature that exports creations for real-world placement and sharing. In a study with 11 novices, pre-post questionnaires showed gains in interest, likelihood to continue offline, and understanding (p ≤.01). Interviews indicated that domestic realism reduced intimidation, natural gestures supported immersion, and AR revisiting extended reflection and engagement. We contribute (1) a home-like, self-directed XR design for ICH practice and (2) evidence that approachability, autonomy, and cross-reality continuity enhance motivation beyond the virtual world. VRCAI ’25, Macau, China

CS Ed. in Prisons and Jails: Evidence of Computer Programming Self-Efficacy Growth Across Multiple Course Offerings

2026-03-05T06:13:06Z

CS Ed. in Prisons and Jails: Evidence of Computer Programming Self-Efficacy Growth Across Multiple Course Offerings Fishberg, Andrew; Gaetz, Marisa; Nisser, Martin; Cafferty, Carole; Perlman, Lee; Soicher, Raechel N.; Long, Joshua Incarcerated students enrolled in education programs in prisons and jails experience a range of benefits, from reduced recidivism to improved psychosocial well-being. With respect to computer science education, little is still known about how courses impact incarcerated students' experiences, though recent work has explored fears and confidence of incarcerated students enrolled in computer science courses. Our work investigates incarcerated students' changes in self-efficacy over multiple iterations of four different classes. Our findings showed that all subscales of computer programming self-efficacy (algorithm, control, cooperation, debugging, and logic), but not generalized self-efficacy, were statistically significantly increased at the end of the courses relative to the beginning (p < 0.001, n = 36). A similar pattern of results across the full sample (n = 188) adds additional support for the veracity of the effects found in the subset of paired data. Additionally, we share students' qualitative data to add nuance to our findings and emphasize the importance of these educational experiences for incarcerated students' personal and professional development.

HyProf: A Profiler for Programming Students that Offers Hypotheses about Performance Bugs

2026-03-05T06:13:25Z

HyProf: A Profiler for Programming Students that Offers Hypotheses about Performance Bugs Dargan, Hope; Hartz, Adam; Miller, Robert Programming students often struggle to find and fix performance bugs in their code. To provide students additional performance debugging support, as well as expose them to profiling tools, we developed Hypothesis Profiler (HyProf). HyProf automatically profiles a slow student submission and produces a profile visualization suitable for learners. In addition to showing individual function and line times, HyProf shows details about the call graph, lines that made recursive calls or did not execute, and hypotheses about possible causes of slow performance, formulated by comparing the slow profile against fast submissions from other students. We deployed HyProf in a 400-student Python course and evaluated it through web logs, office hour observations, and surveys, which showed that 75% of respondents successfully used HyProf to find or fix a performance issue and 85% would recommend it to others. SIGCSE TS 2026, St. Louis, MO, USA

KANELÉ: Kolmogorov–Arnold Networks for Efficient LUT-based Evaluation

2026-03-05T06:13:27Z

KANELÉ: Kolmogorov–Arnold Networks for Efficient LUT-based Evaluation Hoang, Duc; Gupta, Aarush; Harris, Philip C Low-latency, resource-efficient neural network inference on FPGAs is essential for applications demanding real-time capability and low power. Lookup table (LUT)-based neural networks are a common solution, combining strong representational power with efficient FPGA implementation. In this work, we introduce KANELÉ, a framework that exploits the unique properties of Kolmogorov–Arnold Networks (KANs) for FPGA deployment. Unlike traditional multilayer perceptrons (MLPs), KANs employ learnable one-dimensional splines with fixed domains as edge activations, a structure naturally suited to discretization and efficient LUT mapping. We present the first systematic design flow for implementing KANs on FPGAs, co-optimizing training with quantization and pruning to enable compact, high-throughput, and low-latency KAN architectures. Our results demonstrate up to a 2700x speedup and orders of magnitude resource savings compared to prior KAN-on-FPGA approaches. Moreover, KANELÉ matches or surpasses other LUT-based architectures on widely used benchmarks, particularly for tasks involving symbolic or physical formulas, while balancing resource usage across FPGA hardware. Finally, we showcase the versatility of the framework by extending it to real-time, power-efficient control systems. FPGA ’26, Seaside, CA, USA

AI Séance: Recounts from designing artificial intelligence for transcendence, interpretive lenses and chance

2026-03-05T06:13:22Z

AI Séance: Recounts from designing artificial intelligence for transcendence, interpretive lenses and chance Schroeder, Hope; Smith, Amy; Epstein, Ziv As AI becomes a prism through which we reflect, see, and make sense of the world, the way we create creative, transcendent experiences around AI can shape our relationship to it. Drawing inspiration from the ritual structures of Spiritualist séances and creative art-making séances of Hilma af Klint, we present reflections from a series of participatory experiments we called AI Séances. These gatherings brought together artists, technologists, and spiritual practitioners to engage with generative models in contexts shaped by ritual, randomness, and collaborative interpretation. We found that creative production with AI can yield transcendent user experiences (TUX), different communities bring distinct interpretive lenses to AI outputs, and increased technical control can paradoxically diminish serendipity and transcendence. Through our experiences, we suggest that reclaiming interpretive agency over AI outputs in the creative and spiritual context, rather than treating models as machines that produce answers, opens up new avenues for critical and creative engagement with these technologies and is critical to preserving our humanity. The AI Séance offers a model for human-centered interaction with generative systems where magic lies not in the machine’s capabilities, but in our collective ability to create meaning.

2.008 Design and Manufacturing II, Spring 2003

2026-03-04T18:04:45Z

2.008 Design and Manufacturing II, Spring 2003 Dow, David; Sachs, Emanuel; Chun, Jung-Hoon; McAtamney, Patrick; Sarma, Sanjay Integration of design, engineering, and management disciplines and practices for analysis and design of manufacturing enterprises. Emphasis is on the physics and stochastic nature of manufacturing processes and systems, and their effects on quality, rate, cost, and flexibility. Topics include process physics and control, design for manufacturing, and manufacturing systems. Group project requires design and fabrication of parts using mass-production and assembly methods to produce a product in quantity.

2.008 Design and Manufacturing II, Spring 2004

2026-03-04T18:05:26Z

2.008 Design and Manufacturing II, Spring 2004 Chun, Jung-Hoon; Kim, Sang-Gook This course introduces you to modern manufacturing with four areas of emphasis: manufacturing processes, equipment/control, systems, and design for manufacturing. The course exposes you to integration of engineering and management disciplines for determining manufacturing rate, cost, quality and flexibility. Topics include process physics, equipment design and automation/control, quality, design for manufacturing, industrial management, and systems design and operation. Labs are integral parts of the course, and expose you to various manufacturing disciplines and practices.

NeuSE: Neural SE(3)-equivariant embedding for long-term object-based simultaneous localization and mapping

2026-03-05T06:13:16Z

NeuSE: Neural SE(3)-equivariant embedding for long-term object-based simultaneous localization and mapping Fu, Jiahui; Du, Yilun; Singh, Kurran; Tenenbaum, Joshua B; Leonard, John J We present NeuSE, a novel Neural SE(3)-Equivariant Embedding for objects, and illustrate how it supports object-based Simultaneous Localization and Mapping (SLAM) for consistent spatial understanding with long-term scene changes. NeuSE is a set of latent object embeddings created from partial object observations. It serves as a compact point cloud surrogate for complete object models, encoding the full shape, scale, and transform information about an object. In addition, the inferred latent code is both SE(3) and scale equivariant, enabling strong generalization to objects of both unseen sizes and different SE(3) poses. This makes NeuSE particularly effective in real-world scenarios where objects may vary in size or spatial configuration. With NeuSE, relative frame transforms can be directly derived from inferred latent codes. Our proposed SLAM paradigm, using NeuSE for object shape, size, and pose characterization, can operate independently or in conjunction with typical SLAM systems. It directly infers SE(3) camera pose constraints that are compatible with general SLAM pose graph optimization, while maintaining a lightweight, object-centric map that adapts to real-world changes. Our evaluation is conducted on synthetic and real-world sequences with changes in both controlled and uncontrolled settings, featuring multi-category objects of various shapes and sizes. Our approach demonstrates improved localization capability and change-aware mapping consistency when working either independently or as a complement to common SLAM pipelines.

TravelAgent: Generative agents in the built environment

2026-03-05T06:13:07Z

TravelAgent: Generative agents in the built environment Noyman, Ariel; Hu, Kai; Larson, Kent Understanding human behavior in the built environment is critical for designing highly-functional, human-centered urban spaces. Traditional approaches, such as manual observations, surveys, and simple simulations, often struggle to capture the complexity and nuance of real-world human behavior and experience. Here we introduce TravelAgent, a novel agentic simulation platform that models pedestrian navigation, activity, and human-like decision-making in the built environment. TravelAgent is proposed to help design teams and decision-makers understand how different users might experience diverse built environments under varying environmental conditions. TravelAgent integrates Generative Agents, multi-modal sensory inputs, and virtual environments, enabling agents to perceive, navigate, and interact with their surroundings, with tasks ranging from goal-oriented navigation to free exploration. We share analysis from 200 simulations with 3364 decision points and task completion rate of ∼80%, across diverse spatial layouts and agent archetypes. We present spatial, linguistic, and sentiment analysis, and show how agents react and experience their surroundings. Finally, we suggest TravelAgent as a new paradigm for designing, simulating, and understanding human experiences in urban environments.

Cleaning a dark matter detector: A case of ontological and normative elusiveness

2026-03-05T06:13:29Z

Cleaning a dark matter detector: A case of ontological and normative elusiveness de Swart, Jaco; Mol, Annemarie Laboratory sciences crucially depend on the cleanliness of experiments. But what is clean? In this article, we show that the salience of the valuation clean emerges through its relation to a particular ontological repertoire. Our case is the XENONnT experiment in the Gran Sasso Mountains of Italy, designed to detect dark matter in the form of hypothetical WIMPs (Weakly Interacting Massive Particles). In this experiment, dirt presents a significant disruption, as contaminations can mimic the signals of WIMPs, and electronegative molecules risk erasing such signals. The ideosyncratic cleanliness required makes the practice of cleaning the XENONnT detector exceedingly difficult. So far, the ontological question ‘do WIMPs exist?’ remains open, which means that the normative question ‘is the detector clean enough?’ cannot be answered either. In addition, more cleaning will make the detector sensitive to a background of unremovable neutrinos—hence irredeemably dirty. With the normative goal of a ‘clean detector’ out of reach, the ontological question ‘do WIMPs exist?’ is bound to remain open as well. Alternative experiments therefore hunt for different hypothetical dark matter candidates, with different equipment, requiring different kinds of cleanliness. At the same time, the XENONnT experiment must navigate tensions between its own cleanliness goals and rules meant to ensure the environmental cleanliness of the Gran Sasso National Park. Cleaning turns out to be dirty. This leads us to ask: Which goods deserve to be cherished, and, intertwined with that, which realities deserve to be cared for?

Multi-fidelity reinforcement learning for time-optimal quadrotor re-planning

2026-03-05T06:12:59Z

Multi-fidelity reinforcement learning for time-optimal quadrotor re-planning Ryou, Gilhyun; Wang, Geoffrey; Karaman, Sertac High-speed online trajectory planning for UAVs poses a significant challenge due to the need for precise modeling of complex dynamics while also being constrained by computational limitations. This paper presents a multi-fidelity reinforcement learning method (MFRL) that aims to effectively create a realistic dynamics model and simultaneously train a planning policy that can be readily deployed in real-time applications. The proposed method involves the co-training of a planning policy and a reward estimator; the latter predicts the performance of the policy’s output and is trained efficiently through multi-fidelity Bayesian optimization. This optimization approach models the correlation between different fidelity levels, thereby constructing a high-fidelity model based on a low-fidelity foundation, which enables the accurate development of the reward model with limited high-fidelity experiments. The framework is further extended to include real-world flight experiments in reinforcement learning training, allowing the reward model to precisely reflect real-world constraints and broadening the policy’s applicability to real-world scenarios. We present rigorous evaluations by training and testing the planning policy in both simulated and real-world environments. The resulting trained policy not only generates faster and more reliable trajectories compared to the baseline snap minimization method, but it also achieves trajectory updates in 2 ms on average, while the baseline method takes several minutes.

Designing Distribution Network Tariffs in the US with an Application to Increased Electric Vehicle Adoption

2026-03-05T06:13:23Z

Designing Distribution Network Tariffs in the US with an Application to Increased Electric Vehicle Adoption Turk, Graham; Schittekatte, Tim; Duenas-Martinez, Pablo; Joskow, Paul L; Schmalensee, Richard Time-of-use (TOU) tariffs that vary the cost per kWh to reflect wide variations in generation and wholesale market costs give incentives to shift all electric vehicle (EV) charging to low-price periods. As EV penetration increases, such tariffs would substantially raise the local kW demand in those low-priced periods, which eventually would lead to increasing network expansion costs. A straightforward way to mitigate this problem is to separate energy charges from network charges, with appropriate rate designs for each. This paper uses a realistic case study to investigate the implications of combining TOU energy charges with various network tariff designs in the face of increased EV penetration. Our results provide support for the adoption in the US of ex-ante subscribed capacity tariffs (subscription charges), which give consumers incentives to reduce their peak kW demands. Reducing costs of EV ownership (a priority for many US states) need not be pursued at the expense of broader affordability goals. JEL classification: L51, L94, L97, Q41, D40

Geopolitical ecologies of cloud capitalism: Territorial restructuring and the making of national computing power in the U.S. and China

2026-03-05T06:13:28Z

Geopolitical ecologies of cloud capitalism: Territorial restructuring and the making of national computing power in the U.S. and China Kollar, Justin; Stokols, Andrew As computing power becomes central to geopolitical rivalry, cloud infrastructure is increasingly framed as critical to national security, economic resilience and technological sovereignty. Current debates often focus on global competition – especially between the U.S. and China – highlighting strategic investments, export controls and infrastructure diplomacy abroad. Yet far less attention has been paid to the domestic territorial transformations that make such geopolitical projection possible. This paper argues that national strategies for AI and cloud dominance depend on the reorganization of land, energy and regulatory systems to sustain large-scale computation. Using a geopolitical ecology framework, we examine how the U.S. and China build national computing power as a strategic economic and military resource. In the U.S., cloud firms operate as state-aligned actors, drawing on fragmented regulatory authority, public subsidies and national security discourse to expand into rural and peri-urban regions. China pursues a more centralized strategy through its East Data, West Computing initiative, redistributing infrastructure to inland provinces under state-led development goals. Through comparative regional analysis, we show how domestic infrastructural expansion underpins geopolitical rivalry, producing new forms of territorial governance and socio-environmental inequality. Far from immaterial, the cloud is grounded in enclosure, extraction and the spatial foundations of techno-industrial power.

Unlikely Organizers: The Rise of Tech Worker Labor Activism

2026-03-05T06:13:09Z

Unlikely Organizers: The Rise of Tech Worker Labor Activism Tan, JS; Luka, Natalia; Mazo, Emily Tech workers—professionals in the technology industry, such as software engineers, product managers, and UX designers—are not normally associated with labor activism. Yet, since 2017, there has been a significant rise in workplace activism over “bread-and-butter” issues among this group. Using an original data set, the authors demonstrate how, in the case of tech workers, periods of intense workplace social activism preceded later periods of heightened labor activism. Regression analysis confirms that participation in social activism increases the likelihood of labor activism six months to one year later at the same company. Extending Rick Fantasia’s cultures of solidarity to professional workers, the authors highlight a new mechanism by which professionals engage in labor organizing: First, tech workers, guided by their professional interest in socially beneficial work, engage in workplace social activism. This action generates solidarity among employee-participants but also creates conflict with management and leads to the emergence of labor activism among professionals.

Fast detection of liver fibrosis with collagen-binding single-nanometer iron oxide nanoparticles via T1-weighted MRI

2026-03-04T03:07:55Z

Fast detection of liver fibrosis with collagen-binding single-nanometer iron oxide nanoparticles via T1-weighted MRI Zhang, Juanye; Ning, Yingying; Zhu, Hua; Rotile, Nicholas J; Wei, He; Diyabalanage, Himashinie; Hansen, Eric C; Zhou, Iris Y; Barrett, Stephen C; Sojoodi, Mozhdeh; Tanabe, Kenneth K; Humblet, Valerie; Jasanoff, Alan; Caravan, Peter; Bawendi, Moungi G SNIO–CBP, a single-nanometer iron oxide (SNIO) nanoparticle functionalized with a type I collagen-binding peptide (CBP), was developed as a T1-weighted MRI contrast agent with only endogenous elements for fast and noninvasive detection of liver fibrosis. SNIO–CBP exhibits 6.7-fold higher relaxivity compared to a molecular gadolinium-based collagen-binding contrast agent CM-101 on a per CBP basis at 4.7 T. Unlike most iron oxide nanoparticles, SNIO–CBP exhibits fast elimination from the bloodstream with a 5.7 min half-life, high renal clearance, and low, transient liver enhancement in healthy mice. We show that a dose of SNIO–CBP that is 2.5-fold lower than that for CM-101 has comparable imaging efficacy in rapid (within 15 min following intravenous injection) detection of hepatotoxin-induced liver fibrosis using T1-weighted MRI in a carbon tetrachloride–induced mouse liver injury model. We further demonstrate the applicability of SNIO–CBP in detecting liver fibrosis in choline-deficient L-amino acid-defined high-fat diet mouse model of nonalcoholic steatohepatitis. These results provide a platform with potential for the development of high relaxivity, gadolinium-free molecular MRI probes for characterizing chronic liver disease.

Quantum Shell in a Shell: Engineering Colloidal Nanocrystals for a High-Intensity Excitation Regime

2026-03-04T03:07:49Z

Quantum Shell in a Shell: Engineering Colloidal Nanocrystals for a High-Intensity Excitation Regime Harankahage, Dulanjan; Cassidy, James; Beavon, Jacob; Huang, Jiamin; Brown, Niamh; Berkinsky, David B; Marder, Andrew; Kayira, Barbra; Montemurri, Michael; Anzenbacher, Pavel; Schaller, Richard D; Sun, Liangfeng; Bawendi, Moungi G; Malko, Anton V; Diroll, Benjamin T; Zamkov, Mikhail Many optoelectronic processes in colloidal semiconductor nanocrystals (NCs) suffer an efficiency decline under high-intensity excitation. This issue is caused by Auger recombination of multiple excitons, which converts the NC energy into excess heat, reducing the efficiency and life span of NC-based devices, including photodetectors, X-ray scintillators, lasers, and high-brightness light-emitting diodes (LEDs). Recently, semiconductor quantum shells (QSs) have emerged as a promising NC geometry for the suppression of Auger decay; however, their optoelectronic performance has been hindered by surface-related carrier losses. Here, we address this issue by introducing quantum shells with a CdS-CdSe-CdS-ZnS core-shell-shell-shell multilayer structure. The ZnS barrier inhibits the surface carrier decay, which increases the photoluminescence (PL) quantum yield (QY) to 90% while retaining a high biexciton emission QY of 79%. The improved QS morphology allows demonstrating one of the longest Auger lifetimes reported for colloidal NCs to date. The reduction of nonradiative losses in QSs also leads to suppressed blinking in single nanoparticles and low-threshold amplified spontaneous emission. We expect that ZnS-encapsulated quantum shells will benefit many applications exploiting high-power optical or electrical excitation regimes.

Synthesis of Zwitterionic CsPbBr3 Nanocrystals with Controlled Anisotropy using Surface-Selective Ligand Pairs

2026-03-04T03:07:31Z

Synthesis of Zwitterionic CsPbBr3 Nanocrystals with Controlled Anisotropy using Surface-Selective Ligand Pairs Zhu, Hua; Kick, Matthias; Ginterseder, Matthias; Krajewska, Chantalle J; Šverko, Tara; Li, Ruipeng; Lu, Yongli; Shih, Meng‐Chen; Van Voorhis, Troy; Bawendi, Moungi G Mechanistic studies of the morphology of lead halide perovskite nanocrystals (LHP‐NCs) are hampered by a lack of generalizable suitable synthetic strategies and ligand systems. Here, the synthesis of zwitterionic CsPbBr3 NCs is presented with controlled anisotropy using a proposed “surface‐selective ligand pairs” strategy. Such a strategy provides a platform to systematically study the binding affinity of capping ligand pairs and the resulting LHP morphologies. By using zwitterionic ligands (ZwL) with varying structures, majority ZwL‐capped LHP NCs with controlled morphology are obtained, including anisotropic nanoplatelets and nanorods, for the first time. Combining experiments with density functional theory calculations, factors that govern the ligand binding on the different surface facets of LHP‐NCs are revealed, including the steric bulkiness of the ligand, the number of binding sites, and the charge distance between binding moieties. This study provides guidance for the further exploration of anisotropic LHP‐NCs.

Theory of Photoluminescence Spectral Line Shapes of Semiconductor Nanocrystals

2026-03-04T03:07:30Z

Theory of Photoluminescence Spectral Line Shapes of Semiconductor Nanocrystals Lin, Kailai; Jasrasaria, Dipti; Yoo, Jason J; Bawendi, Moungi; Utzat, Hendrik; Rabani, Eran Single-molecule photoluminescence (PL) spectroscopy of semiconductor nanocrystals (NCs) reveals the nature of exciton-phonon interactions in NCs. Understanding the homogeneous spectral line shapes and their temperature dependence remains an open problem. Here, we develop an atomistic model to describe the PL spectrum of NCs, accounting for excitonic effects, phonon dispersion relations, and exciton-phonon couplings. We validate our model using single-NC measurements on CdSe/CdS NCs from T = 4 to 290 K, and we find that the slightly asymmetric main peak at low temperatures is comprised of a narrow zero-phonon line (ZPL) and acoustic phonon sidebands. Furthermore, we identify the specific phonon modes that give rise to the optical phonon sidebands. At temperatures above 200 K, the spectral line width shows a stronger dependence upon the temperature, which we demonstrate to be correlated with higher order exciton-phonon couplings. We also identify the line width dependence upon reorganization energy, NC core sizes, and shell thicknesses.

Ultrafast dense DNA functionalization of quantum dots and rods for scalable 2D array fabrication with nanoscale precision

2026-03-04T03:07:50Z

Ultrafast dense DNA functionalization of quantum dots and rods for scalable 2D array fabrication with nanoscale precision Chen, Chi; Luo, Xin; Kaplan, Alexander EK; Bawendi, Moungi G; Macfarlane, Robert J; Bathe, Mark Scalable fabrication of two-dimensional (2D) arrays of quantum dots (QDs) and quantum rods (QRs) with nanoscale precision is required for numerous device applications. However, self-assembly–based fabrication of such arrays using DNA origami typically suffers from low yield due to inefficient QD and QR DNA functionalization. In addition, it is challenging to organize solution-assembled DNA origami arrays on 2D device substrates while maintaining their structural fidelity. Here, we reduced manufacturing time from a few days to a few minutes by preparing high-density DNA-conjugated QDs/QRs from organic solution using a dehydration and rehydration process. We used a surface-assisted large-scale assembly (SALSA) method to construct 2D origami lattices directly on solid substrates to template QD and QR 2D arrays with orientational control, with overall loading yields exceeding 90%. Our fabrication approach enables the scalable, high fidelity manufacturing of 2D addressable QDs and QRs with nanoscale orientational and spacing control for functional 2D photonic devices.

Rational Design of a Chemical Bath Deposition Based Tin Oxide Electron‐Transport Layer for Perovskite Photovoltaics

2026-03-04T03:07:44Z

Rational Design of a Chemical Bath Deposition Based Tin Oxide Electron‐Transport Layer for Perovskite Photovoltaics Lu, Yongli; Shih, Meng‐Chen; Tan, Shaun; Grotevent, Matthias J; Wang, Lili; Zhu, Hua; Zhang, Ruiqi; Lee, Joo‐Hong; Lee, Jin‐Wook; Bulović, Vladimir; Bawendi, Moungi G Chemical bath deposition (CBD) is widely used to deposit tin oxide (SnOx) as an electron-transport layer in perovskite solar cells (PSCs). The conventional recipe uses thioglycolic acid (TGA) to facilitate attachments of SnOx particles onto the substrate. However, nonvolatile TGA is reported to harm the operational stability of PSCs. In this work, a volatile oxalic acid (OA) is introduced as an alternative to TGA. OA, a dicarboxylic acid, functions as a chemical linker for the nucleation and attachment of particles to the substrate in the chemical bath. Moreover, OA can be readily removed through thermal annealing followed by a mild H2O2 treatment, as shown by FTIR measurements. Synergistically, the mild H2O2 treatment selectively oxidizes the surface of the SnOx layer, minimizing nonradiative interface carrier recombination. EELS (electron-energy-loss spectroscopy) confirms that the SnOx surface is dominated by Sn4+, while the bulk is a mixture of Sn2+ and Sn4+. This rational design of a CBD SnOx layer leads to devices with T85 ≈1500 h, a significant improvement over the TGA-based device with T80 ≈250 h. The champion device reached a power conversion efficiency of 24.6%. This work offers a rationale for optimizing the complex parameter space of CBD SnOx to achieve efficient and stable PSCs.

Reduced recombination via tunable surface fields in perovskite thin films

2026-03-04T03:07:54Z

Reduced recombination via tunable surface fields in perovskite thin films deQuilettes, Dane W; Yoo, Jason J; Brenes, Roberto; Kosasih, Felix Utama; Laitz, Madeleine; Dou, Benjia Dak; Graham, Daniel J; Ho, Kevin; Shi, Yangwei; Shin, Seong Sik; Ducati, Caterina; Bawendi, Moungi G; Bulović, Vladimir The ability to reduce energy loss at semiconductor surfaces through passivation or surface field engineering is an essential step in the manufacturing of efficient photovoltaic (PV) and optoelectronic devices. Similarly, surface modification of emerging halide perovskites with quasi-two-dimensional (2D) heterostructures is now ubiquitous to achieve PV power conversion efficiencies (PCEs) >25%, yet a fundamental understanding to how these treatments function is still generally lacking. Here we use a unique combination of depth-sensitive nanoscale characterization techniques to uncover a tunable passivation strategy and mechanism found in perovskite PV devices that were the first to reach the >25% PCE milestone. Namely, treatment with hexylammonium bromide leads to the simultaneous formation of an iodide-rich 2D layer along with a Br halide gradient that extends from defective surfaces and grain boundaries into the bulk three-dimensional (3D) layer. This interface can be optimized to extend the charge carrier lifetime to record values >30 μs and to reduce interfacial recombination velocities to values as low as <7 cm s−1.

Solution-phase sample-averaged single-particle spectroscopy of quantum emitters with femtosecond resolution

2026-03-04T03:07:48Z

Solution-phase sample-averaged single-particle spectroscopy of quantum emitters with femtosecond resolution Shi, Jiaojian; Shen, Yuejun; Pan, Feng; Sun, Weiwei; Mangu, Anudeep; Shi, Cindy; McKeown-Green, Amy; Moradifar, Parivash; Bawendi, Moungi G; Moerner, WE; Dionne, Jennifer A; Liu, Fang; Lindenberg, Aaron M The development of many quantum optical technologies depends on the availability of single quantum emitters with near-perfect coherence. Systematic improvement is limited by a lack of understanding of the microscopic energy flow at the single-emitter level and ultrafast timescales. Here we utilize a combination of fluorescence correlation spectroscopy and ultrafast spectroscopy to capture the sample-averaged dynamics of defects with single-particle sensitivity. We employ this approach to study heterogeneous emitters in two-dimensional hexagonal boron nitride. From milliseconds to nanoseconds, the translational, shelving, rotational and antibunching features are disentangled in time, which quantifies the normalized two-photon emission quantum yield. Leveraging the femtosecond resolution of this technique, we visualize electron–phonon coupling and discover the acceleration of polaronic formation on multi-electron excitation. Corroborated with theory, this translates to the photon fidelity characterization of cascaded emission efficiency and decoherence time. Our work provides a framework for ultrafast spectroscopy in heterogeneous emitters, opening new avenues of extreme-scale characterization for quantum applications.

Additive‐Free Oxidized Spiro‐MeOTAD Hole Transport Layer Significantly Improves Thermal Solar Cell Stability

2026-03-04T03:07:51Z

Additive‐Free Oxidized Spiro‐MeOTAD Hole Transport Layer Significantly Improves Thermal Solar Cell Stability Grotevent, Matthias J; Lu, Yongli; Šverko, Tara; Shih, Meng‐Chen; Tan, Shaun; Zhu, Hua; Dang, Tong; Mwaura, Jeremiah K; Swartwout, Richard; Beiglböck, Finn; Kothe, Linda; Bulović, Vladimir; Bawendi, Moungi G Perovskite solar cells are among the most promising new solar technologies, already surpassing polycrystalline silicon solar cell efficiencies. The stability of the highest efficiency devices at elevated temperature is, however, poor. These cells typically use Spiro‐MeOTAD as the hole transporting layer. It is generally believed that additives, required for enhancing electrical conductivity and optimizing energy level alignment, are responsible for the reduced stability—inferring that Spiro‐MeOTAD based hole transporting layers are intrinsically unstable. Here, a reliable noble metal free synthesis of Spiro‐MeOTAD (bis(trifluoromethane)sulfonimide)4 is presented which is used as the oxidizing agent. No additives are added to the partially oxidized Spiro‐MeOTAD hole‐transporting layer. Device efficiencies up to 24.2% are achieved. Electrical conductivity is largely developed by the first 1% oxidation. Further oxidation shifts the energy levels away from the vacuum level, which allows tuning of the energy level alignment without the use of additives—contradicting the current understanding of this system. Without additives, devices demonstrate significant improvement in stability at elevated temperatures up to 85 °C under one sun over 1400 h continuous illumination. The remaining degradation is pinpointed to ion migration and reactions in the perovskite layer which may be further suppressed with compositional engineering and adequate ion barrier layers.

Bright and Fast Emission from Robust Supramolecular J-Aggregate Nanostructures through Silica-Encapsulation

2026-03-04T03:07:46Z

Bright and Fast Emission from Robust Supramolecular J-Aggregate Nanostructures through Silica-Encapsulation Thanippuli Arachchi, Dimuthu H; Barotov, Ulugbek; Perkinson, Collin F; Šverko, Tara; Kaplan, Alexander EK; Bawendi, Moungi G We introduce a two-step silica-encapsulation procedure to optimize both the optical efficiency and structural robustness of 5,5',6,6'-tetrachloro-1,1'-diethyl-3,3'-di(4-sulfobutyl)-benzimidazolocarbocyanine (TDBC), a two-dimensional sheet-like J-aggregate. We report a fluorescence quantum yield of ∼98%, the highest quantum yield recorded for any J-aggregate structure at room temperature, and a fast, emissive lifetime of 234 ps. Silica, as an encapsulating matrix, provides optical transparency, chemical inertness, and robustness to dilution, while rigidifying the J-aggregate structure. Our in situ encapsulation process preserves the excitonic structure in TDBC J-aggregates, maintaining their light absorption and emission properties. The homogeneous silica coating has an average thickness of 0.5-1 nm around J-aggregate sheets. Silica encapsulation permits extensive dilutions of J-aggregates without significant disintegration into monomers. The narrow absorbance and emission line widths exhibit further narrowing upon cooling to 79 K, which is consistent with J-type coupling in the encapsulated aggregates. This silica TDBC J-aggregate construct signifies (1) a bright, fast, and robust fluorophore system, (2) a platform for further manipulation of J-aggregates as building blocks for integration with other optical materials and structures, and (3) a system for fundamental studies of exciton delocalization, transport, and emission dynamics within a rigid matrix.

Toward biophysical markers of depression vulnerability

2026-03-04T03:07:22Z

Toward biophysical markers of depression vulnerability Pinotsis, DA; Fitzgerald, S; See, C; Sementsova, A; Widge, AS A major difficulty with treating psychiatric disorders is their heterogeneity: different neural causes can lead to the same phenotype. To address this, we propose describing the underlying pathophysiology in terms of interpretable, biophysical parameters of a neural model derived from the electroencephalogram. We analyzed data from a small patient cohort of patients with depression and controls. Using DCM, we constructed biophysical models that describe neural dynamics in a cortical network activated during a task that is used to assess depression state. We show that biophysical model parameters are biomarkers, that is, variables that allow subtyping of depression at a biological level. They yield a low dimensional, interpretable feature space that allowed description of differences between individual patients with depressive symptoms. They could capture internal heterogeneity/variance of depression state and achieve significantly better classification than commonly used EEG features. Our work is a proof of concept that a combination of biophysical models and machine learning may outperform earlier approaches based on classical statistics and raw brain data.

Understanding Robustness and Generalization of Artificial Neural Networks Through Fourier Masks

2026-03-04T03:07:37Z

Understanding Robustness and Generalization of Artificial Neural Networks Through Fourier Masks Karantzas, Nikos; Besier, Emma; Ortega Caro, Josue; Pitkow, Xaq; Tolias, Andreas S.; Patel, Ankit B.; Anselmi, Fabio Despite the enormous success of artificial neural networks (ANNs) in many disciplines, the characterization of their computations and the origin of key properties such as generalization and robustness remain open questions. Recent literature suggests that robust networks with good generalization properties tend to be biased toward processing low frequencies in images. To explore the frequency bias hypothesis further, we develop an algorithm that allows us to learn modulatory masks highlighting the essential input frequencies needed for preserving a trained network's performance. We achieve this by imposing invariance in the loss with respect to such modulations in the input frequencies. We first use our method to test the low-frequency preference hypothesis of adversarially trained or data-augmented networks. Our results suggest that adversarially robust networks indeed exhibit a low-frequency bias but we find this bias is also dependent on directions in frequency space. However, this is not necessarily true for other types of data augmentation. Our results also indicate that the essential frequencies in question are effectively the ones used to achieve generalization in the first place. Surprisingly, images seen through these modulatory masks are not recognizable and resemble texture-like patterns.

Enhancement of Cyanobacterial Bloom Monitoring in Lake Taihu Using Dual Red-Edge Bands of GF-6/WFV: Multi-Dimensional Feature Combination and Extraction Accuracy Analysis

2026-03-04T03:07:41Z

Enhancement of Cyanobacterial Bloom Monitoring in Lake Taihu Using Dual Red-Edge Bands of GF-6/WFV: Multi-Dimensional Feature Combination and Extraction Accuracy Analysis Sun, Yunxiao; Zhang, Ruolin; Zhao, Chunhong; Meng, Qingyan; Sun, Zhenhui; Wang, Jialong; Wu, Jun; Wang, Yao; Gao, Decai; Guan, Huyi Cyanobacterial blooms pose a serious threat to freshwater ecosystems, necessitating accurate remote sensing monitoring. Although red-edge bands show potential in terrestrial monitoring, their multi-dimensional features (i.e., spectral, textural, and index-based characteristics) remain underutilized for aquatic blooms. This study leverages the dual red-edge bands (710 nm and 750 nm) of GF-6/WFV to enhance cyanobacterial bloom identification in Lake Taihu. Multi-temporal images from 2019–2023 were used to construct red-edge features in three dimensions: spectral (evaluated via adaptive band selection method) and Jeffries–Matusita–Bhattacharyya distance), texture (based on Gray Level Co-occurrence Matrix and principal component analysis), and indices (nine vegetation indices ranked by Random Forest importance). Twelve feature-combination schemes were designed and implemented with a Random Forest classifier. Results show that red-edge features consistently improve identification accuracy. Quantitatively, compared to the basic four-band (RGBN) combination, the 710 nm band improved spectral separability by an average of 9.63%, whereas the 750 nm band yielded a lower average improvement of 5.69%. Red-edge indices, especially the modified chlorophyll absorption reflectance index 1 (MCARI1) and normalized difference red-edge index (NDRE), exhibited higher importance than non-red-edge indices. All schemes incorporating red-edge features achieved mean overall accuracies of 92.8–94.9% and Kappa coefficients of 0.86–0.94, surpassing the basic four-band scheme. Among these features, red-edge indices contributed most significantly to accuracy gains, increasing the overall accuracy by an average of 0.36–6.06% and the Kappa coefficient by up to 0.06. The enhancement effect of the red-edge 710 nm band features was superior to that of the 750 nm band. This study demonstrates that multi-dimensional red-edge features effectively enhance the identification accuracy of cyanobacterial blooms and provides a methodological reference for operational GF-6 applications in water quality monitoring.

Biological Activity of Metal Complexes

2026-03-04T03:07:43Z

Biological Activity of Metal Complexes Sharma, Vinay K. Metal complexes play a fundamental role in biological systems and continue to attract sustained interest due to their remarkable potential in therapeutic, diagnostic, and biotechnological applications [1–8]. In recent years, the field of bioinorganic chemistry has advanced rapidly, driven by progress in coordination chemistry, spectroscopy, nanotechnology, and molecular biology [9–22]. These developments have enabled a deeper understanding of how metal ions and complexes interact with biomolecular targets and have opened new avenues for the rational design of metal-based agents for cancer therapy, antimicrobial treatment, imaging, and the study of metal-mediated biochemical processes [23–30].

Single Parameter Model for Galaxy Rotation Curves

2026-03-04T03:07:33Z

Single Parameter Model for Galaxy Rotation Curves Cisneros, Sophia N.; Ott, Rich; Crowley, Meagan; Roberts, Amy; Paz, Marcus One key piece of evidence for dark matter is the rotation-curve problem: the disagreement between measured galactic rotation curves and their luminous mass. A novel solution to this problem is presented here, in a model that predicts observed Doppler-shifted spectra based only on the luminous matter estimates and one free model parameter 𝛼. This model is applied to fit the rotation curves of the SPARC sample of 175 galaxies, yielding mass-to-light ratios, goodness of fit measurements, and 𝛼. The measured average 𝜒2 𝑟 =2.24 compares favorably with the Navarro-Frenk-White dark matter model’s average of 𝜒2 𝑟 =4.19 for the same data, and more galaxies are successfully fit by this model. The model provides a useful formulation linking luminous matter to the observed rotation curves, with the dark matter contribution to galaxies encoded in two transformation terms of the luminous mass. It also offers a lower-parameter characterization of the rotation curve problem, and a power law relationship between 𝛼 and galactic photometric quantities is observed, potentially removing the need for the free parameter.

Non-Clinical Safety of GRAd Vector-Based COVID-19 and HIV Vaccines Supports a Platform Regulatory Approach

2026-03-04T03:07:38Z

Non-Clinical Safety of GRAd Vector-Based COVID-19 and HIV Vaccines Supports a Platform Regulatory Approach Paalangara, Reji; Gohin, Stephanie; Menard, Alexis; Amy, Charlotte; Berrabah, Wahiba; Rogue, Alexandra; Getz, Matthew A.; Alrubayyi, Aljawharah; Battella, Simone; Raggioli, Angelo; Gentile, Michela; Di Rita, Anthea; Noto, Alessia; Miselli, Giuseppina; Grazioli, Fabiana; Napolitano, Federico; Sowcik, Dhurata; Soriani, Marco; Chmielewski, Benjamin; Molife, Lebohang Background/Objectives: The rapid development of safe and efficacious vaccines is often hindered by extensive, mandated non-clinical safety evaluations in animals. With the aim to provide scientific evidence supporting a “vaccine platform approach”, here we present the complete non-clinical studies for two investigational vaccines, GRAd-COV2 and GRAdHIVNE1, based on GRAd, a gorilla-derived group C adenoviral vector. Methods: The biodistribution of GRAd genomes following the intramuscular administration of the vaccines was assessed in rats by a sensitive qPCR method. Local tolerance and systemic toxic effects were evaluated in single- and repeated-dose toxicity studies in rabbits. Results: GRAd-COV2 and GRAdHIVNE1 were well-tolerated. Distribution was highly confined to the injection site and draining lymph nodes, and toxicity profile consisted of transient, non-adverse inflammatory responses, while the expected immune responses to the encoded antigens were successfully induced. Notably, both vaccines demonstrated a consistent safety profile despite transgene and backbone differences, comparable to other replication-defective adenoviral vectors. Conclusions: The established non-clinical safety profile of the GRAd platform provides a robust foundation for a more efficient and streamlined regulatory pathway. By leveraging this prior knowledge, future GRAd-based vaccines can achieve accelerated clinical development while fully adhering to the ethical principles of replacement, reduction, and refinement of animal use in research.

Design and User-Centered Field Evaluation of an Accessible Precision Irrigation Tool and Its Human–Machine Interaction on a Jordanian Farm

2026-03-04T03:07:34Z

Design and User-Centered Field Evaluation of an Accessible Precision Irrigation Tool and Its Human–Machine Interaction on a Jordanian Farm Van de Zande, Georgia D.; Sheline, Carolyn; Pratt, Shane R.; Winter V, Amos G. This work aims to demonstrate the successful, long-term human use of an automatic scheduling-manual operation (AS-MO) precision irrigation tool by farmers on a medium-scale Jordanian farm. Innovation in low-cost, accessible, and water-efficient irrigation technologies is critical as water resources become scarce, especially on resource-constrained farms in the drought-prone Middle East and North Africa (MENA) region. Prior work has shown that a proposed AS-MO decision support tool could bridge the gap between fully manual irrigation—a common practice on many MENA farms—and existing precision agriculture solutions, which are often too expensive or complex for medium-scale farmers to adopt. Recent developments have also demonstrated that the scheduling theory behind the proposed AS-MO tool uses up to 44% less water compared to fully manual irrigation. However, a functional design of the AS-MO tool has not been realized nor has it been demonstrated on a farm with farmer users. This work documents the detailed design of an AS-MO tool’s human–machine interaction (HMI) and validates the human execution of the tool in context. Through an 11-week case study conducted on a Jordanian farm, we show that farmers used a functional prototype of the AS-MO tool as intended. The functional tool prototype was designed to deliver a long-term AS-MO user experience to study participants. The prototype monitored local weather conditions, generated water-efficient schedules using an existing scheduling theory, and notified users’ phones when they should manually open or close valves. The irrigation practices of participants using the AS-MO prototype were measured, and participants demonstrated successful use of the tool. Users correctly confirmed 93% of the scheduled events using the tool’s HMI. Despite manual operation, a majority of confirmed irrigation event durations fell within 15% of the automatically scheduled durations; relative to the length of scheduled irrigation event durations, the medians of confirmed and scheduled durations were 102% and 88%, respectively. These results demonstrate the success of the tool’s decision support ability. Feedback from study participants can support the AS-MO tool’s next design iteration and can inform the development of other decision support systems designed for resource-constrained, medium-scale farms. This work presents an important step towards developing a precision irrigation tool that, if adopted at scale, could increase the adoption of water-efficient irrigation practices on resource-constrained farms that are not served by existing technology, improving sustainable agriculture in MENA.

The Effect of Genipin Matrix Augmentation on the Retention of Glycosaminoglycans in the Intervertebral Disc—A Pilot Study

2026-03-04T03:07:39Z

The Effect of Genipin Matrix Augmentation on the Retention of Glycosaminoglycans in the Intervertebral Disc—A Pilot Study Hedman, Thomas; Brown, Matthew; Slusarewicz, Pawel The degradation of intervertebral disc proteoglycans, including the loss or shortening of their hydrophilic glycosaminoglycan chains, causes a loss of disc hydration, leading to an increase in solid matrix stresses. This illustrates one aspect of the complex multifactorial relationship between tissue degradation and the resulting mechanical dysfunction. Genipin matrix augmentation has previously been evaluated with regard to its ability to improve mechanical properties of the disc, increasing joint stability and permeability. The study aim was to evaluate the ability of genipin augmentation to increase retention of glycosaminoglycans in disc specimens subjected to free swelling. Three different models were utilized: whole bovine caudal discs, partial annulus specimens from bovine, and human thoracic discs. Total glycosaminoglycan release to a surrounding bath was quantified using a modified dimethyl-methylene blue assay. Genipin solution injections reduced glycosaminoglycan loss by 44.0% in intact bovine discs compared to buffer-only controls (p = 0.027), by 75.8% in partial bovine annulus specimens (p = 0.0004), and by 51.9% in human annulus specimens (p = 0.017). The combination of increased permeability and glycosaminoglycans retention may produce beneficial effects on nutritional flow, diurnal irrigation, and reduction of matrix solid phase stress. Combining these effects with the ability to improve joint stability and augment tissue mechanical properties suggests this nano-scale device may be capable of arresting ongoing degeneration.

A Novel Recurrent Neural Network Framework for Prediction and Treatment of Oncogenic Mutation Progression

2026-03-04T03:07:35Z

A Novel Recurrent Neural Network Framework for Prediction and Treatment of Oncogenic Mutation Progression Parthasarathy, Rishab; Bhowmik, Achintya K. Despite significant medical advancements, cancer remains the second leading cause of death in the US, causing over 600,000 deaths per year. One emerging field, pathway analysis, is promising but still relies on manually derived wet lab data, which is time-consuming to acquire. This work proposes an efficient, effective, end-to-end framework for Artificial Intelligence (AI)-based pathway analysis that predicts both cancer severity and mutation progression in order to recommend possible treatments. The proposed technique involves a novel combination of time-series machine learning models and pathway analysis. First, mutation sequences were isolated from The Cancer Genome Atlas (TCGA) Database. Then, a novel preprocessing algorithm was used to filter key mutations by mutation frequency. This data was fed into a Recurrent Neural Network (RNN) that predicted cancer severity. The model probabilistically used the RNN predictions, information from the preprocessing algorithm, and multiple drug-target databases to predict future mutations and recommend possible treatments. This framework achieved robust results and Receiver Operating Characteristic (ROC) curves (a key statistical metric) with accuracies greater than 60%, similar to existing cancer diagnostics. In addition, preprocessing played a key role in isolating a few hundred key driver mutations per cancer stage, consistent with current research. Heatmaps based on predicted gene frequency were also generated, highlighting key mutations in each cancer. Overall, this work is the first to propose an efficient, cost-effective end-to-end framework for projecting cancer prognosis and providing possible treatments without relying on expensive, time-consuming wet lab work.

All-Pay Auctions with Different Forfeits

2026-03-04T03:07:42Z

All-Pay Auctions with Different Forfeits Kang, Benjamin; Unwin, James In an auction, each party bids a certain amount, and the one who bids the highest is the winner. Interestingly, auctions can also be used as models for other real-world systems. In an all-pay auction all parties must pay a forfeit for bidding. In the most commonly studied all-pay auction, parties forfeit their entire bid, and this has been considered as a model for expenditure on political campaigns. Here, we consider a number of alternative forfeits that might be used as models for different real-world competitions, such as preparing bids for defense or infrastructure contracts.

Spontaneous formation of robust two-dimensional perovskite phases

2026-03-03T03:07:04Z

Spontaneous formation of robust two-dimensional perovskite phases Tan, Shaun; Shih, Meng-Chen; Lu, Yongli; Choi, Seung-Gu; Dong, Yifan; Lee, Joo-Hong; Yavuz, Ilhan; Larson, Bryon W; Park, So Yeon; Kodalle, Tim; Zhang, Ruiqi; Grotevent, Matthias J; Lin, Yu-Kuan; Zhu, Hua; Bulović, Vladimir; Sutter-Fella, Carolin M; Park, Nam-Gyu; Beard, Matthew C; Lee, Jin-Wook; Zhu, Kai; Bawendi, Moungi G The two-dimensional on three-dimensional (2D/3D) perovskite bilayer heterostructure can improve the stability and performance of perovskite solar cells. We show that the 2D/3D perovskite stack in a device evolves dynamically during its end-of-life decomposition. Initially phase-pure 2D interlayers can evolve differently, resulting in different device stabilities. We show that a robust 2D interlayer can be formed using mixed solvents to regulate its crystallinity and phase purity. The resulting 2D/3D devices achieved 25.9% efficiency and had good durability, retaining 91% of their initial performance after 1074 hours at 85°C using maximum power point tracking.

Third-order photon correlations extract single-nanocrystal multiexciton properties in solution

2026-03-03T03:06:58Z

Third-order photon correlations extract single-nanocrystal multiexciton properties in solution Horowitz, Jonah R; Berkinsky, David B; Bendekgey, Henry C; Tye, Oliver J; Šverko, Tara; Shulenberger, Katherine E; Bawendi, Moungi G Colloidal semiconductor nanocrystals are considered promising materials for high-flux optical applications, including lasing, light-emitting diodes, biological imaging, and quantum optics. In high-flux applications, multiexcitons can significantly contribute to emission, influencing its brightness, spectral purity, and kinetics. As a result, understanding and controlling multiexciton emission in colloidal nanocrystal materials is of the utmost importance. In the past, single-nanocrystal photon correlation methods have been applied to understand biexciton and triexciton efficiencies, lifetimes, and spectra. While powerful, such methods suffer from user selection bias and require stable emission from single nanocrystals. To compensate for this shortcoming, second-order correlation methods were developed to extract sample-averaged biexciton properties from a solution of nanocrystals. Until now, however, the analogous third-order solution photon correlation methods remained unexplored. In this work, we present a pair of third-order photon correlation techniques to obtain the sample-averaged single-nanocrystal triexciton quantum yield and lifetime in a solution-phase experiment. These techniques derive from the relationship between the Poisson probability of nanocrystal photon absorption and the intrinsic probability of nanocrystal photon emission. We validate the theoretical background of these techniques by creating a numerical model to simulate the diffusion and emission of many nanocrystals in solution. Our simulations confirm that the average triexciton quantum yield and triexciton lifetime can be extracted from a solution of nanocrystals. These techniques will enable researchers to gain a better understanding of the fundamental multiexciton properties of colloidal nanocrystals.

Challenges of II‐VI and III‐V Blue Quantum Dot Light‐Emitting Diodes

2026-03-03T03:07:00Z

Challenges of II‐VI and III‐V Blue Quantum Dot Light‐Emitting Diodes Tan, Shaun; Horowitz, Jonah R; Tye, Oliver J; Bawendi, Moungi G Quantum dot light-emitting diodes (QD-LEDs) are electroluminescent devices where the emissive layer consists of inorganic colloidal quantum dots. Recent breakthroughs have enabled the development of bright and efficient blue-emitting QD-LEDs based on heavy metal-free compositions. However, challenges remain that hinder their practical application in electroluminescent displays and lighting technologies. The primary obstacle is their limited operational lifetimes which remain significantly below practical requirement standards, especially in comparison to the red- and green-emitting QD-LEDs. Another important issue is the low color purity and broad spectral linewidths of heavy metal-free blue quantum dot compositions. Additional problems include transient electroluminescent behaviors such as fluorescence intermittency and positive aging effects. This review examines the current understanding of the physical mechanisms underlying these challenges faced by blue QD-LEDs. Often, contradictory explanations are proposed to account for the same phenomenon. Here, potential interpretations are suggested that may help reconcile the conflicting reports. Recent advances are further examined that have contributed to the development of state-of-the-art blue QD-LEDs.

Impact of Processing Environment on Anti-Solvent Free FAPbI3 Films and Solar Cells

2026-03-03T03:06:51Z

Impact of Processing Environment on Anti-Solvent Free FAPbI3 Films and Solar Cells Wall, Elizabeth M; Lin, Yu‐Kuan; Bawendi, Moungi; Burlingame, Quinn C; Loo, Yueh‐Lin As perovskite solar cells approach commercialization, understanding the environmental sensitivities of perovskites during fabrication becomes increasingly important. In this work, the humidity-dependence of each deposition and annealing step in the anti-solvent-free two-step formamidinium lead iodide fabrication process is investigated in air and N2. In-situ grazing-incidence wide-angle X-ray scattering measurements during spin-coating indicate that humidity affects the formation and dynamics of intermediate phases in perovskite precursor films. These differences, and those induced by annealing in humidity, impact the structure, morphology, and composition of resultant perovskite films, though the initial performance of solar cells fabricated using these active layers is relatively insensitive to humidity across the range studied. In contrast, stability is maximized in devices with dry-processed active layers and those terminally annealed in humidity. Spin-coating of PbI2 is most environmentally sensitive—needle-like structures precipitate while spin-coating in 40% relative humidity leading to significantly reduced photovoltaic performance and device stability. Additionally, films and solar cells fabricated in air appear virtually identical to those fabricated in N2. Collectively, these results show that optimal performance and stability of two-step processed formamidinium lead iodide solar cells is achieved when fabricating active layers in a dry atmosphere or with some humidity during the final anneal.

Cognitive Reinforcement: Capturing Tacit Knowledge and Enhancing Expertise with a Biofeedback Interface for Visual Attention

2026-03-04T17:47:02Z

Cognitive Reinforcement: Capturing Tacit Knowledge and Enhancing Expertise with a Biofeedback Interface for Visual Attention Armengol-Urpi, Alexandre; Salazar-Gomez, Andres F.; Sinha, Pawan; Sarma, Sanjay E. Objective. Tacit or implicit knowledge refers to know-how that experts possess but often cannot articulate, codify, or explicitly transfer to others. This can present a significant challenge for learning, skill acquisition, and knowledge transfer across various domains, including those that rely on apprenticeships, craftsmanship, sports, and medical imaging diagnosis. This study explores whether expert tacit knowledge can be accessed and leveraged using an EEG and gaze-informed biofeedback interface to enhance expertise transfer and training. Approach. We designed an image classification task where novices were trained until they implicitly learned to classify images correctly, despite being unaware of which image regions or features guided their decisions. The task involved images with a hidden spatial asymmetry that even trained participants did not explicitly recognize. Using combined eye-tracking and EEG measures, we tracked both overt and covert visual attention to determine whether individuals unconsciously internalized this asymmetry during learning. We then investigated whether providing explicit gaze-informed feedback on their own implicit attention biases could further improve task performance of trained participants. Main Results. Our findings reveal that as participants became trained, their attention patterns —both overt and covert— consistently reflected an unconscious awareness of image asymmetry, with attention biased toward task-relevant image regions. Moreover, trained individuals who received explicit feedback derived from their own gaze behavior showed additional improvements in classification performance compared to an equally trained control group. Significance. These results open the door to novel uses of biofeedback interfaces to facilitate new forms of expertise transfer, training, and collective intelligence. By extracting and conveying tacit expert knowledge—ordinarily difficult to externalize—our interface enables its transmission to novices, trained individuals, or even machine learning systems. We refer to this process as cognitive reinforcement.

Higher Siegel-Weil formulae over function fields

2026-02-28T03:02:13Z

Higher Siegel-Weil formulae over function fields Mkrtchyan, Mikayel In their seminal work, Feng-Yun-Zhang introduced function field analogues of Kudla-Rapoport cycles for moduli spaces of unitary shtukas, and initiated the study of their intersection theory. They proved a higher Siegel-Weil formula in the case of non-degenerate Fourier coefficients, relating the degrees of these cycles to higher derivatives of Siegel-Eisenstein series. In this thesis, we generalize their result in two directions: we 1) prove a higher Siegel-Weil formula for unitary groups for corank-1 degenerate coefficients, and 2) introduce analogous cycles on moduli spaces of symplectic shtukas, and prove a higher Siegel-Weil formula for such cycles in the non-degenerate case, relating their degrees to derivatives of Siegel-Eisenstein series on split orthogonal groups.

A Diverse Array of Synthetic Strategies for Phosphorus Group Transfer Chemistry: From Phosphinidenes to Phosphates

2026-02-28T03:02:21Z

A Diverse Array of Synthetic Strategies for Phosphorus Group Transfer Chemistry: From Phosphinidenes to Phosphates Xin, Tiansi This thesis compiles the published scientific contributions of Tiansi Xin. Chapter 1 consists of a brief collection of eulogies from friends and colleagues, reflecting on his life and time at the Massachusetts Institute of Technology. The subsequent chapters describe the development of novel synthetic methods for the transfer of phosphorus-containing moieties, specifically metaphosphates and phosphinidenes. The work presented here has significant implications for both the fundamental understanding and practical advancement of synthetic inorganic and organic chemistry. Chapters 2 and 3 address the sustainable production and processing of phosphoruscontaining chemicals, focusing on mechanochemical methods to synthesize reduced phosphorus species while circumventing the need to access hazardous white phosphorus as an intermediate. In particular, Chapter 2 describes a solvent-free mechanochemical approach to producing phosphite (HPO₃²⁻) via hydride-mediated reduction of condensed phosphates. Using potassium hydride, a range of inorganic phosphate sources—including pyrophosphate, triphosphate, trimetaphosphate, fluorophosphate, and polyphosphate—were converted to phosphite in moderate to high yields. Mechanistic studies identified overreduction pathways leading to hypophosphite and other low-oxidation P-species. Chapter 3 similarly applies this mechanochemical approach to phosphorus–carbon bond formation, reporting the phosphorylation of acetylides with condensed phosphates to afford phosphonates. Biogenic polyphosphates were also shown to be viable precursors, a proof-of-concept to closing the modern phosphorus cycle using recycled inputs. These results demonstrate the possibility of accessing organophosphorus chemicals directly from condensed phosphates and may offer an opportunity toward a “greener” phosphorus industry. Chapters 4 and 5 shift focus to phosphinidene transfer chemistry and the synthesis of novel phosphorus-containing heterocycles. This expands on previously published studies from the Cummins group on the chemistry of dibenzo-7-phosphanorbornadiene “RPA” reagents. Chapter 4 reports the preparation and structural characterization of iron–phosphido complexes relevant to phosphinidene group transfer catalysis and describes the development of an improved catalytic system based on a simple diiron precursor (Fp₂), enabling efficient synthesis of phosphiranes from electron-deficient alkenes. The mechanism was thoroughly experimentally and computationally interrogated. Chapter 5 describes the novel synthesis of free, uncomplexed phosphet-2-ones via phosphinidene transfer to cyclopropenones, with experimental and theoretical studies supporting a mechanism involving ketene-derived intermediates and transformations to additional phosphorus heterocycles through subsequent reactions.

Host–Guest Complexation by β-Cyclodextrin Enhances the Solubility of an Esterified Protein

2026-02-27T04:16:04Z

Host–Guest Complexation by β-Cyclodextrin Enhances the Solubility of an Esterified Protein Cheah, Keith M; Jun, Joomyung V; Wittrup, K Dane; Raines, Ronald T The carboxyl groups of a protein can be esterified by reaction with a diazo compound, 2-diazo-2-(p-methylphenyl)-N,N-dimethylacetamide. This esterification enables the entry of the protein into the cytosol of a mammalian cell, where the nascent ester groups are hydrolyzed by endogenous esterases. The low aqueous solubility of the ensuing esterified protein is, however, a major practical challenge. Solubility screening revealed that β-cyclodextrin (β-CD) is an optimal solubilizing agent for esterified green fluorescent protein (est-GFP). Its addition can increase the recovery of est-GFP by 10-fold. α-CD, γ-CD, and cucurbit-7-uril are less effective excipients. 1H NMR titration experiments revealed that β-CD encapsulates the hydrophobic tolyl group of ester conjugates with Ka = 321 M–1. Combining l-arginine and sucrose with β-CD enables the nearly quantitative recovery of est-GFP. Thus, the insolubility of esterified proteins can be overcome with excipients.

Quantifying the Role of Kinematic and Behavioral Features in Driver-Pedestrian Interaction across Environments: An Inverse Reinforcement Learning Approach

2026-02-27T04:16:16Z

Quantifying the Role of Kinematic and Behavioral Features in Driver-Pedestrian Interaction across Environments: An Inverse Reinforcement Learning Approach Noonan, T Zach; Gershon, Pnina; Domeyer, Josh; Mehler, Bruce; Reimer, Bryan This study examined real-world driver-pedestrian encounters to identify key interaction features and assess how the importance of these features is mediated by protection afforded by the environment. Using inverse reinforcement learning, we estimated the utility functions to evaluate the relative importance of different aspects of the interaction for each road user and how they differ between undesignated (e.g., jaywalking) and designated (e.g., zebra crossings) crossings. Pedestrian pausing behavior and dynamic features like acceleration changes and time gaps were important at designated crossings, whereas undesignated crossings relied on distances and bidirectional gaze, highlighting reliance on non-verbal cues. This work builds on previous studies analyzing the role of environmental features on interaction, communication, and negotiation between drivers and pedestrians. Understanding driver-pedestrian communication and identifying the most important interaction features may enhance the design of effective and coordinated driver-pedestrian interaction strategies, especially in the context of automated driving systems.

Preoperative Function, Previous SERM Treatment, and Triple-Negative Tumor Status are Independently Associated With 3-Month Postoperative Function After Surgical Decompression of Metastatic Breast Cancer

2026-02-27T04:16:12Z

Preoperative Function, Previous SERM Treatment, and Triple-Negative Tumor Status are Independently Associated With 3-Month Postoperative Function After Surgical Decompression of Metastatic Breast Cancer Siraj, Layla; Duvall, Julia B.; Massaad, Elie; Fourman, Mitchell S.; Shin, John H. Background: The most common cancer in women worldwide, breast cancer most often metastasizes to the bone. Improved chemo- and radiotherapies and novel molecular therapies have prolonged survival in women with osseous metastatic breast cancer, but spinal metastases often cause cord compression that degrades their functional independence. Purpose: In women with breast cancer metastasized to the spine, we sought to (1) identify independent predictors of a functional deficit 3 months after surgical management and (2) assess the utility of existing metrics at highlighting patients at risk of a postoperative functional deficit. Methods: We performed a single-institution, retrospective analysis of 92 patients meeting our inclusion criteria between 2004 and 2021. Patients were classified by 3-month postoperative Eastern Cooperative Oncology Group (ECOG) scores into good/independent (ECOG 0 to 2) and poor/dependent (ECOG 3 to 5) functional outcome groups. Univariate and multivariate analyses were performed to identify patient and tumor factors associated with good vs. poor 3-month ECOG scores. Results: Preoperative use of selective estrogen receptor modulators (SERMs) was significantly associated with good postoperative functional outcomes. Poor preoperative function, the presence of visceral metastases at the time of surgery, and triple-negative primary or metastatic tumor status were independently associated with poor 3-month postoperative function. Host characteristics, sociodemographic factors, and indicators of surgical complexity, including estimated blood loss, front/back surgery, and corpectomy reconstruction, were not associated with 3-month ECOG score. A multivariate model including these significant univariate associations and normalized for patient demographics identified preoperative SERM use, poor preoperative function (ECOG score), and triple-negative primary or metastatic tumor status as independently associated with functional status 3 months after surgery. Conclusions: Our retrospective analysis found that preoperative SERM use was significantly associated with improved postoperative functional outcomes, while poor preoperative function and triple-negative tumor status were significantly associated with poor function 3 months after surgery. These factors may serve as indicators of function and independence after surgery for patients with metastatic breast cancer to the spine. Level of Evidence: Level IV: Prognostic Study

The Influence of Prior Semantic Knowledge in Noisy Channel Interpretation

2026-02-27T04:16:13Z

The Influence of Prior Semantic Knowledge in Noisy Channel Interpretation Chen, Sihan; Washington, Lia; Gibson, Edward How do comprehenders interpret semantically implausible sentences? Previous studies proposed a noisy-channel framework of sentence comprehension, where communication between a speaker and a comprehender happens in a noisy channel. The comprehender rationally adopts an interpretation of a sentence based on how likely the interpretation is (the semantic prior) and how likely is the interpretation corrupted into the perceived sentence because of noise (the likelihood). The theory predicted that comprehenders would be more likely to adopt a literal interpretation of an implausible sentence if their prior of implausible sentences were higher. To test this hypothesis, Gibson et al. manipulated the proportion of implausible test sentences in two sets of experiments, where participants read a number of sentences and answer a comprehension question following each sentence. Although their results supported the hypothesis, the experiment could be confounded (a) by participants’ adaptation effect (due to different experiment lengths) and (b) by different participants having different strategies to do the task (due to the between-subject design). In our study, we manipulated the semantic prior and controlled for these potential confounds. We found participants exposed to more implausible sentences were indeed more likely to interpret implausible sentences literally. Our results hence offer additional support for the noisy-channel framework.

The Politics of Engagement in Platform Governance

2026-02-27T04:16:11Z

The Politics of Engagement in Platform Governance Lewis, Becca; Christin, Angèle In recent years, the concept of user engagement has dominated debate over the governance of online platforms, and critics use the term to assign crass commercial interests to social media companies. We argue that social media engagement is a multifaceted ideal that serves both economic and ideological functions for platforms. We show how Facebook’s early leadership used the concept to reconcile the competing demands of expansion, revenue generation, and community-building. In doing so, they synthesized three distinct ideas: the Silicon Valley belief that network expansion correlated with network strength, the ad industry’s contention that media should promote emotional investment from viewers, and the academic claim that civic participation is the most important democratic virtue. Even as the contradictions that these claims yield have come to the foreground, the multiple logics of engagement have proven difficult to evade, and it continues to shape discussions of platform governance.

The Impact of Electrification and Partial Automation on Driver Speeding Behavior

2026-02-27T04:16:15Z

The Impact of Electrification and Partial Automation on Driver Speeding Behavior Gershon, Pnina; Noonan, T Zach; Lenneman, John As electric vehicles (EVs) and partial automation systems become increasingly prevalent, their impact on everyday driving behavior remains underexplored. This study utilizes real-world naturalistic data to examine how vehicle type, an electric versus an internal combustion engine (ICE), and the use of partial automation are associated with speeding behavior. Data were collected from 24 drivers over the course of a month each, comparing Tesla Model 3s with Autopilot (EV) and Cadillac CT6s with Super Cruise (ICE), covering about 38,000 miles of driving. Results indicate that EV drivers tended to speed for shorter durations on arterial roads but exhibited higher speeding magnitudes on residential and controlled access roads after their first week of driving. Notably, driving with partial automation, regardless of powertrain, was associated with significantly longer speeding durations and slightly greater speeding magnitudes compared to manual driving. These findings suggest that both electrification and automation contribute to evolving driver behaviors, changing speeding behavior in specific driving contexts. As drivers adapt to new vehicle technologies, understanding how these systems shape behavior is important. Insights from this study may inform the design of future in-vehicle systems and guide driver education strategies to promote safe driving practices in an evolving transportation landscape.

Cloud Capitalism and the AI Transition

2026-02-27T04:16:14Z

Cloud Capitalism and the AI Transition Tan, JS; Thelen, Kathleen This article explores the origins and implications of a new cloud business model that is powering the advance of AI. We document how this model emerged within a handful of the most dominant IT firms whose reach into all corners of the economy makes them a powerful node or “choke point” in the political economy as a whole. We then elaborate how the features of the cloud business model differ from the traditional platform model out of which it grew, as it evolved from asset-light to asset-heavy, from hierarchical organization to semivertical integration, from domination over to collaboration with partner firms, and from embracing consumer- to enterprise-facing strategies. A final section considers the technological, political, and distributional impacts of the rise of this new business model—showing how the current race to artificial general intelligence (AGI) has reinforced and accelerated its underlying dynamics (above all, intensifying the drive for scale and ever-greater asset intensity), analyzing the new techno-nationalist alliance between industry leaders and the state that the model's development has inspired, and considering the new power-distributional dynamics this model has produced.

Report to the President for year ended June 30, 2025, Vice President for Research

2026-02-27T04:17:20Z

Report to the President for year ended June 30, 2025, Vice President for Research Waitz, Ian A This report contains the following sections: Overview, Research Administration, Research Compliance, Office of Research Computing and Data, International Scholars Office, Postdoctoral Services, OVPR Administration, and Lab and Center Leadership.

Hexamethylbenzene Elimination Enables the Generation of Transient, Sterically Unhindered Multiply Bonded Boron Species

2026-02-27T04:16:10Z

Hexamethylbenzene Elimination Enables the Generation of Transient, Sterically Unhindered Multiply Bonded Boron Species Zhang, Chonghe; Dabringhaus, Philipp; Tra, Bi Youan E.; Gilliard, Robert J. Jr; Cummins, Christopher C. We present a method for the generation of boron-containing unsaturated small molecules via hexamethylbenzene elimination. The fragmentation precursors are obtained through bond insertion into phenyl boranorbornadiene (PhB(C6Me6), 1). Compound 1 undergoes 1,1-insertion with 2,6-xylyl isocyanide, affording a boron-doped bicyclo[2.2.2]octa-2,5-diene 2. Heating 2 in toluene results in the formation of a base-stabilized boraketenimine PhB(CNxyl)2 (i.e., borylene diisocyanide) as an intermediate via retro-Diels–Alder reaction. Surprisingly, PhB(CNxyl)2 dimerizes to give a boron-doped 6-membered ring (PhB)2C4(CNxyl)64. The reaction of 1 with trimethylamine N-oxide and phenyl azide yields triphenyl boroxine and a BN4 ring, respectively, implying the involvement of transient oxoborane (PhB[triple bond, length as m-dash]O) and iminoborane intermediates (PhB[triple bond, length as m-dash]NPh), respectively. Furthermore, boranorbornadiene also undergoes 2,3-insertion with mesityl isocyanate (MesNCO), affording a fused 6/5-membered heterocycle 11. This insertion profile is analogous to the insertion of phenyl azide into 1.

Mechanisms and Scale-up Potential of 3D Solar Interfacial-Evaporators

2026-02-27T04:16:06Z

Mechanisms and Scale-up Potential of 3D Solar Interfacial-Evaporators Zhang, James H.; Mittapally, Rohith; Oluwade, Abimbola; Chen, Gang Evaporation fluxes from porous evaporators under sunlight have been reported to exceed the solar-thermal limit, determined by relating the incoming solar energy to the latent and sensible heat of water, for applications in desalination and brine pond drying. Although flat two-dimensional (2D) evaporators exceeding the solar limit imply a non-thermal process, tall three-dimensional (3D) solar evaporators can exceed it by absorbing additional environmental heat into its cold sidewalls. Through modeling, we explain the physics and identify the critical heights in which a fin transitions from 2D to 3D evaporation and exceeds the solar-thermal limit. Our analyses illustrate that environmental heat absorption in 3D evaporators is determined by the ambient relative humidity and the airflow velocity. The model is then coarse-grained into a large-scale fin array device on the meters scale to analyze their scalability. We identify that these devices are unlikely to scale favorably in closed environment settings such as solar stills. Our modeling clearly illustrates the benefits and limitations of 3D evaporating arrays and pinpoints design choices in previous works that hinder the device's overall performance. This work illustrates the importance in distinguishing 2D from 3D evaporation for mechanisms underlying interfacial evaporation exceeding the solar-thermal limit.

Decarbonization Approaches for Ethylene Production: Comparative Techno-Economic and Life-Cycle Analysis

2026-02-27T04:16:09Z

Decarbonization Approaches for Ethylene Production: Comparative Techno-Economic and Life-Cycle Analysis Shin, Woojae; Lin, Bosong; Lai, Haoxiang; Ibrahima, Gasim; Zang, Guiyan Ethylene, a building block of the chemical industry, significantly contributes to global greenhouse gas (GHG) emissions, prompting interest in decarbonization approaches to align with recent carbon neutrality initiatives. This paper presents a comprehensive techno-economic analysis (TEA) and life cycle analysis (LCA) of GHG emissions, comparing conventional ethane-based ethylene plants with three decarbonization approaches. The study was conducted within the context of the U.S. average, with sensitivity analysis to identify key drivers affecting well-to-gate (WTG) GHG emissions and the levelized cost of ethylene (LCOE). The conventional plant exhibited a GHG emission of 869 kgCO2e per tonne-ethylene and a LCOE of $746 per tonne-ethylene. Substituting external natural gas fuels with grid or renewable electricity decreased the emissions to 806 and 717 kgCO2e per tonne-ethylene, respectively. The emissions of the grid-powered or renewable-powered electrically heated cracker that exports co-produced hydrogen to substitute conventional gray hydrogen were 1031 and −163 kgCO2e per tonne-ethylene, respectively. The application of CCS to purge gas showed 703 and 514 kgCO2e per tonne-ethylene emissions, respectively. The electric cracker showed lower emissions than the conventional plant below 380 kgCO2e per MW h electricity upstream, and at 60 kgCO2e per MW h, it achieved carbon neutrality. Regarding LCOE, when using a grid electricity source, no external natural gas, electric cracker, and adding CCS to purge gas showed $743, 833, and 771 per tonne-ethylene, respectively. When these plants adopt renewable electricity, their LCOEs will be $737, 746 and 757 per tonne-ethylene. Below $41.1 per MW h electricity price, the electric cracker had the lowest value among all cases. With hydrogen prices of $0.5–3.0 per kg-H2, the electric cracker's LCOE ranged from −$45(cost)–128(saving) per tonne-ethylene compared to the conventional concept.

Archerfish: A Retrofitted 3D Printer for High-throughput Combinatorial Experimentation via Continuous Printing

2026-02-27T04:16:03Z

Archerfish: A Retrofitted 3D Printer for High-throughput Combinatorial Experimentation via Continuous Printing Alexander E. Siemenn, Basita Das, Eunice Aissi, Fang Sheng, Lleyton Elliott, Blake Hudspeth, Marilyn Meyers, James Serdy and Tonio Buonassisi The maturation of 3D printing technology has enabled low-cost, rapid prototyping capabilities for mainstreaming accelerated product design. The materials research community has recognized this need, but no universally accepted rapid prototyping technique currently exists for material design. Toward this end, we develop Archerfish, a 3D printer retrofitted to dispense liquid with in situ mixing capabilities for performing high-throughput combinatorial printing (HTCP) of material compositions. Using this HTCP design, we demonstrate continuous printing throughputs of up to 250 unique compositions per minute, 100× faster than similar tools such as Opentrons that utilize stepwise printing with ex situ mixing. We validate the formation of these combinatorial “prototype” material gradients using hyperspectral image analysis and energy-dispersive X-ray spectroscopy. Furthermore, we describe hardware challenges to realizing reproducible, accurate, and precise composition gradients with continuous printing, including those related to precursor dispensing, mixing, and deposition. Despite these limitations, the continuous printing and low-cost design of Archerfish demonstrate promising accelerated materials screening results across a range of materials systems from nanoparticles to perovskites.

Borinine-FLP Ring Expansion: Isolation of Eight-Membered B-P Rings Bridged by µ2 Chalcogenide and Chloronium Ions

2026-02-27T04:15:52Z

Borinine-FLP Ring Expansion: Isolation of Eight-Membered B-P Rings Bridged by µ2 Chalcogenide and Chloronium Ions Frey, Nathan C.; Sarkar, Samir Kumar; Dickie, Diane A.; Molinoa, Andrew; Gilliard, Robert J. Jr Boron–phosphorus (B–P) frustrated Lewis pairs (FLPs) are an important class of compounds for activating various small molecules. Utilizing the ring expansion reactivity of 9-chloro-9-borafluorene, a borinine-based FLP was synthesized. Various five-membered main-group element heterocycles were obtained via the reaction of the FLP with Me3NO, S8, and Se. Subsequent reduction of these species yielded the ring-expanded compounds, each featuring bridging B–E–B (E = O, S, Se) bonds. Similarly, halide abstraction from the FLP with AgNTf2 led to the formation of a cationic ring-expanded compound with a bridging B–Cl–B motif. This motif constitutes one of the first examples of a boron-stabilized chloronium ion, as verified using in-depth bonding analysis methods. Mechanistic pathways for the reduction- and halide abstraction-mediated ring expansion reactions are proposed with the aid of density functional theory. Electronic structure computations were performed to determine the best representation of bonding interactions in each compound, suggesting phosophorus(V)–chalcogen double bonding and chalcogen–boron(III) dative interactions within the heterocycles.

High-resolution structure of Zn3(HOTP)2 (HOTP = hexaoxidotriphenylene), a three-dimensional conductive MOF

2026-02-27T04:16:01Z

High-resolution structure of Zn3(HOTP)2 (HOTP = hexaoxidotriphenylene), a three-dimensional conductive MOF Zhang, Kimberly J.; Chen, Tianyang; Oppenheim, Julius J.; Yang, Luming; Palatinus, Lukáš; Müller, Peter; Van Voorhisa, Troy; Dincă, Mircea Although two-dimensional (2D) electrically conducting metal–organic frameworks (cMOFs) have become prominent due to their numerous potential applications, their structures are often implied or assumed from rather crude powder X-ray diffraction data. Indeed, exceedingly few examples exist of atomic-level structural details coming from single crystal diffraction experiments. Most widely studied among cMOFs are materials based on triphenylene ligands, in particular M3(HOTP)2 (M = Cu, Zn) and [M3(HOTP)2][M3(HOTP)]2 (M = Mg, Ni, Co; H6HOTP = 2,3,6,7,10,11-hexahydroxytriphenylene), which are invariably described as 2D van der Waals materials with sheets of ligands connected by square planar or octahedral metal ions. Here, we employ electron diffraction to show that, unlike the Mg, Co, Ni, and Cu analogs, Zn3(HOTP)2 crystallizes into a three-dimensional network that is analogous to the structures of the lanthanide-based HOTP MOFs. Moreover, similar to the lanthanide frameworks, Zn3(HOTP)2 exhibits incommensurate modulation, likely originating from a frustration between the preferred π–π stacking distance and the Zn–O bond lengths, or from a Peierls distortion. This work reinforces the importance of employing single crystal diffraction measurements for the characterization of conductive MOFs, especially when trying to correlate electronic properties to structural details.

Intratumoral nanobody–IL-2 fusions that bind the tumor extracellular matrix suppress solid tumor growth in mice

2026-02-26T03:07:36Z

Intratumoral nanobody–IL-2 fusions that bind the tumor extracellular matrix suppress solid tumor growth in mice Lutz, Emi A; Jailkhani, Noor; Momin, Noor; Huang, Ying; Sheen, Allison; Kang, Byong H; Wittrup, K Dane; Hynes, Richard O Confining cytokine exposure to the tumors would greatly enhance cancer immunotherapy safety and efficacy. Immunocytokines, cytokines fused to tumor-targeting antibodies, have been developed with this intention, but without significant clinical success to date. A critical limitation is uptake by receptor-expressing cells in the blood, that decreases the dose at the tumor and engenders toxicity. Small-format immunocytokines, constructed with antibody fragments, are hypothesized to improve tumor specificity due to rapid systemic clearance. However, effective design criteria for small-format immunocytokines need further examination. Here, we engineer small interleukin-2 (IL-2) immunocytokines fused to nanobodies with nanomolar to picomolar affinities for the tumor-specific EIIIB domain of fibronectin (also known as EDB). Upon intravenous delivery into immunocompetent mice, such immunocytokines led to similar tumor growth delay as size-matched untargeted IL-2. Intratumoral (i.t.) delivery imparted improved survival dependent on affinity to EIIIB. I.t. administration offers a promising avenue to deliver small-format immunocytokines, given effective affinity for the tumor microenvironment.

Ablative radiotherapy improves survival but does not cure autochthonous mouse models of prostate and colorectal cancer

2026-02-26T03:07:27Z

Ablative radiotherapy improves survival but does not cure autochthonous mouse models of prostate and colorectal cancer Schmidt, Daniel R; Gramatikov, Iva Monique T; Sheen, Allison; Williams, Christopher L; Hurwitz, Martina; Dodge, Laura E; Holupka, Edward; Kiger, WS; Cornwall-Brady, Milton R; Huang, Wei; Mak, Howard H; Cormier, Kathleen S; Condon, Charlene; Dane Wittrup, K; Yilmaz, Ömer H; Stevenson, Mary Ann; Down, Julian D; Floyd, Scott R; Roper, Jatin; Vander Heiden, Matthew G Background Genetically engineered mouse models (GEMMs) of cancer are powerful tools to study mechanisms of disease progression and therapy response, yet little is known about how these models respond to multimodality therapy used in patients. Radiation therapy (RT) is frequently used to treat localized cancers with curative intent, delay progression of oligometastases, and palliate symptoms of metastatic disease. Methods Here we report the development, testing, and validation of a platform to immobilize and target tumors in mice with stereotactic ablative RT (SART). Xenograft and autochthonous tumor models were treated with hypofractionated ablative doses of radiotherapy. Results We demonstrate that hypofractionated regimens used in clinical practice can be effectively delivered in mouse models. SART alters tumor stroma and the immune environment, improves survival in GEMMs of primary prostate and colorectal cancer, and synergizes with androgen deprivation in prostate cancer. Complete pathologic responses were achieved in xenograft models, but not in GEMMs. Conclusions While SART is capable of fully ablating xenografts, it is unable to completely eradicate disease in GEMMs, arguing that resistance to potentially curative therapy can be modeled in GEMMs. Plain language summary Mice can be used to model the types of cancer seen in people to investigate the effects of cancer therapies, such as radiation. Here, we apply radiation therapy treatments that are able to cure cancer in humans to mice that have cancer of the prostate or colorectum. We show that the mice do not experience many side effects and that the tumours reduce in size, but in some cases show progression after treatment. Our study demonstrates that mice can be used to better understand how human cancers respond to radiation treatment, which can lead to the development of improved treatments and treatment schedules.

Anti–PD-1 and Extended Half-life IL2 Synergize for Treatment of Murine Glioblastoma Independent of Host MHC Class I Expression

2026-02-26T03:07:45Z

Anti–PD-1 and Extended Half-life IL2 Synergize for Treatment of Murine Glioblastoma Independent of Host MHC Class I Expression Tritz, Zachariah P; Ayasoufi, Katayoun; Wolf, Delaney M; Owens, Carley A; Malo, Courtney S; Himes, Benjamin T; Fain, Cori E; Goddery, Emma N; Yokanovich, Lila T; Jin, Fang; Hansen, Michael J; Parney, Ian F; Wang, Chensu; Moynihan, Kelly D; Irvine, Darrell J; Wittrup, K Dane; Diaz Marcano, Rosa M; Vile, Richard G; Johnson, Aaron J Glioblastoma (GBM) is the most common malignant brain tumor in adults, responsible for approximately 225,000 deaths per year. Despite preclinical successes, most interventions have failed to extend patient survival by more than a few months. Treatment with anti—programmed cell death protein 1 (anti–PD-1) immune checkpoint blockade (ICB) monotherapy has been beneficial for malignant tumors such as melanoma and lung cancers but has yet to be effectively employed in GBM. This study aimed to determine whether supplementing anti–PD-1 ICB with engineered extended half-life IL2, a potent lymphoproliferative cytokine, could improve outcomes. This combination therapy, subsequently referred to as enhanced checkpoint blockade (ECB), delivered intraperitoneally, reliably cures approximately 50% of C57BL/6 mice bearing orthotopic GL261 gliomas and extends median survival of the treated cohort. In the CT2A model, characterized as being resistant to CBI, ECB caused a decrease in CT2A tumor volume in half of measured animals similar to what was observed in GL261-bearing mice, promoting a trending survival increase. ECB generates robust immunologic responses, features of which include secondary lymphoid organ enlargement and increased activation status of both CD4 and CD8 T cells. This immunity is durable, with long-term ECB survivors able to resist GL261 rechallenge. Through employment of depletion strategies, ECB's efficacy was shown to be independent of host MHC class I–restricted antigen presentation but reliant on CD4 T cells. These results demonstrate ECB is efficacious against the GL261 glioma model through an MHC class I–independent mechanism and supporting further investigation into IL2-supplemented ICB therapies for tumors of the central nervous system.

Collagen-Anchored Interleukin-2 and Interleukin-12 Safely Reprogram the Tumor Microenvironment in Canine Soft-Tissue Sarcomas

2026-02-26T03:07:43Z

Collagen-Anchored Interleukin-2 and Interleukin-12 Safely Reprogram the Tumor Microenvironment in Canine Soft-Tissue Sarcomas Stinson, Jordan A; Sheen, Allison; Momin, Noor; Hampel, Jordan; Bernstein, Rebecca; Kamerer, Rebecca; Fadl-Alla, Bahaa; Samuelson, Jonathan; Fink, Elizabeth; Fan, Timothy M; Wittrup, K Dane Purpose: Cytokine therapies such as IL2 and IL12 suffer from impractically small therapeutic windows driven by their on-target, off-tumor activity, limiting their clinical potential despite potent antitumor effects. We previously engineered cytokines that bind and anchor to tumor collagen following intratumoral injection, and sought to test their safety and biomarker activity in spontaneous canine soft-tissue sarcomas (STS). Experimental Design: Collagen-binding cytokines were canine-ized to minimize immunogenicity and were used in a rapid dose-escalation study in healthy beagles to identify a maximum tolerated dose. Ten client-owned pet dogs with STS were then enrolled into trial, receiving cytokines at different intervals prior to surgical tumor excision. Tumor tissue was analyzed through IHC and NanoString RNA profiling for dynamic changes within treated tumors. Archived, untreated STS samples were analyzed in parallel as controls. Results: Intratumorally administered collagen-binding IL2 and IL12 were well tolerated by STS-bearing dogs, with only Grade 1/2 adverse events observed (mild fever, thrombocytopenia, neutropenia). IHC revealed enhanced T-cell infiltrates, corroborated by an enhancement in gene expression associated with cytotoxic immune function. We found concordant increases in expression of counter-regulatory genes that we hypothesize would contribute to a transient antitumor effect, and confirmed in mouse models that combination therapy to inhibit this counter-regulation can improve responses to cytokine therapy. Conclusions: These results support the safety and activity of intratumorally delivered, collagen-anchoring cytokines for inflammatory polarization of the canine STS tumor microenvironment. We are further evaluating the efficacy of this approach in additional canine cancers, including oral malignant melanoma. Translational Relevance Successful translation of novel cancer therapies could be accelerated through the inclusion of tumor models that accurately recapitulate natural evolution and malignant transformation processes operative in human tumor development. Spontaneous cancer in pet dogs provides an underutilized opportunity to assess the safety and activity of investigational cancer therapies in tumors that arise following years of immunoediting. Particularly for the evaluation of immunotherapies, canine tumors enable the assessment of clinical potential in the context of an experienced, and often senescent, immune background. Beyond efficacy, such evaluation provides meaningful insight into tumor resistance mechanisms that could influence eventual human clinical success. Herein, we characterize immune activities generated by intratumoral injections of engineered collagen-binding cytokines IL2 and IL12 into naturally occurring canine soft-tissue sarcomas, and demonstrate through comparative assessment in mouse tumors the differential learnings from each model and their combined role in guiding rational design of treatment combinations with greater expected efficacy.

Both intratumoral regulatory T cell depletion and CTLA-4 antagonism are required for maximum efficacy of anti-CTLA-4 antibodies

2026-02-26T03:07:38Z

Both intratumoral regulatory T cell depletion and CTLA-4 antagonism are required for maximum efficacy of anti-CTLA-4 antibodies Lax, Brianna M; Palmeri, Joseph R; Lutz, Emi A; Sheen, Allison; Stinson, Jordan A; Duhamel, Lauren; Santollani, Luciano; Kennedy, Alan; Rothschilds, Adrienne M; Spranger, Stefani; Sansom, David M; Wittrup, K Dane Anti-CTLA-4 antibodies have successfully elicited durable tumor regression in the clinic; however, long-term benefit is limited to a subset of patients for select cancer indications. The incomplete understanding of their mechanism of action has hindered efforts at improvement, with conflicting hypotheses proposing either antagonism of the CTLA-4:B7 axis or Fc effector-mediated regulatory T cell (Treg) depletion governing efficacy. Here, we report the engineering of a nonantagonistic CTLA-4 binding domain (b1s1e2) that depletes intratumoral Tregs as an Fc fusion. Comparison of b1s1e2-Fc to 9d9, an antagonistic anti-CTLA-4 antibody, allowed for interrogation of the separate contributions of CTLA-4 antagonism and Treg depletion to efficacy. Despite equivalent levels of intratumoral Treg depletion, 9d9 achieved more long-term cures than b1s1e2-Fc in MC38 tumors, demonstrating that CTLA-4 antagonism provided additional survival benefit. Consistent with prior reports that CTLA-4 antagonism enhances priming, treatment with 9d9, but not b1s1e2-Fc, increased the percentage of activated T cells in the tumor-draining lymph node (tdLN). Treg depletion with either construct was restricted to the tumor due to insufficient surface CTLA-4 expression on Tregs in other compartments. Through intratumoral administration of diphtheria toxin in Foxp3-DTR mice, we show that depletion of both intratumoral and nodal Tregs provided even greater survival benefit than 9d9, consistent with Treg-driven restraint of priming in the tdLN. Our data demonstrate that anti-CTLA-4 therapies require both CTLA-4 antagonism and intratumoral Treg depletion for maximum efficacy—but that potential future therapies also capable of depleting nodal Tregs could show efficacy in the absence of CTLA-4 antagonism.

Overcoming lung cancer immunotherapy resistance by combining nontoxic variants of IL-12 and IL-2

2026-02-26T03:07:35Z

Overcoming lung cancer immunotherapy resistance by combining nontoxic variants of IL-12 and IL-2 Horton, Brendan L; D’Souza, Alicia D; Zagorulya, Maria; McCreery, Chloe V; Abhiraman, Gita C; Picton, Lora; Sheen, Allison; Agarwal, Yash; Momin, Noor; Wittrup, K Dane; White, Forest M; Garcia, K Christopher; Spranger, Stefani Engineered cytokine-based approaches for immunotherapy of cancer are poised to enter the clinic, with IL-12 being at the forefront. However, little is known about potential mechanisms of resistance to cytokine therapies. We found that orthotopic murine lung tumors were resistant to systemically delivered IL-12 fused to murine serum albumin (MSA, IL12-MSA) because of low IL-12 receptor (IL-12R) expression on tumor-reactive CD8+ T cells. IL2-MSA increased binding of IL12-MSA by tumor-reactive CD8+ T cells, and combined administration of IL12-MSA and IL2-MSA led to enhanced tumor-reactive CD8+ T cell effector differentiation, decreased numbers of tumor-infiltrating CD4+ regulatory T cells, and increased survival of lung tumor-bearing mice. Predictably, the combination of IL-2 and IL-12 at therapeutic doses led to significant dose-limiting toxicity. Administering IL-12 and IL-2 analogs with preferential binding to cells expressing Il12rb1 and CD25, respectively, led to a significant extension of survival in mice with lung tumors while abrogating dose-limiting toxicity. These findings suggest that IL-12 and IL-2 represent a rational approach to combination cytokine therapy whose dose-limiting toxicity can be overcome with engineered cytokine variants.

Intratumoral aluminum hydroxide–anchored IL-12 drives potent antitumor activity by remodeling the tumor microenvironment

2026-02-26T03:07:32Z

Intratumoral aluminum hydroxide–anchored IL-12 drives potent antitumor activity by remodeling the tumor microenvironment Battula, Sailaja; Papastoitsis, Gregory; Kaufman, Howard L; Wittrup, K Dane; Schmidt, Michael M IL-12 is a potent cytokine that can promote innate and adaptive anticancer immunity, but its clinical development has been limited by toxicity when delivered systemically. Intratumoral (i.t.) administration can expand the therapeutic window of IL-12 and other cytokines but is in turn limited by rapid drug clearance from the tumor, which reduces efficacy, necessitates frequent administration, and increases systemic accumulation. To address these limitations, we developed an anchored IL-12 designated ANK-101, composed of an engineered IL-12 variant that forms a stable complex with the FDA-approved vaccine adjuvant aluminum hydroxide (Alhydrogel). Following i.t. administration of murine ANK-101 (mANK-101) in early intervention syngeneic mouse tumors, the complex formed a depot that was locally retained for weeks as measured by IVIS or SPECT/CT imaging, while unanchored protein injected i.t. was cleared within hours. One or 2 i.t. injections of mANK-101 induced single-agent antitumor activity across a diverse range of syngeneic tumors, including models resistant to checkpoint blockade at doses where unanchored IL-12 had no efficacy. Local treatment with mANK-101 further induced regressions of noninjected lesions, especially when combined with systemic checkpoint blockade. Antitumor activity was associated with remodeling of the tumor microenvironment, including prolonged IFN-γ and chemokine expression, recruitment and activation of T and NK cells, M1 myeloid cell skewing, and increased antigen processing and presentation. Subcutaneous administration of ANK-101 in cynomolgus macaques was well tolerated. Together, these data demonstrate that ANK-101 has an enhanced efficacy and safety profile and warrants future clinical development.

CD8+ T cell priming that is required for curative intratumorally anchored anti-4-1BB immunotherapy is constrained by Tregs

2026-02-26T03:07:40Z

CD8+ T cell priming that is required for curative intratumorally anchored anti-4-1BB immunotherapy is constrained by Tregs Palmeri, Joseph R; Lax, Brianna M; Peters, Joshua M; Duhamel, Lauren; Stinson, Jordan A; Santollani, Luciano; Lutz, Emi A; Pinney, William; Bryson, Bryan D; Dane Wittrup, K Although co-stimulation of T cells with agonist antibodies targeting 4-1BB (CD137) improves antitumor immune responses in preclinical studies, clinical success has been limited by on-target, off-tumor activity. Here, we report the development of a tumor-anchored ɑ4-1BB agonist (ɑ4-1BB-LAIR), which consists of a ɑ4-1BB antibody fused to the collagen-binding protein LAIR. While combination treatment with an antitumor antibody (TA99) shows only modest efficacy, simultaneous depletion of CD4+ T cells boosts cure rates to over 90% of mice. Mechanistically, this synergy depends on ɑCD4 eliminating tumor draining lymph node regulatory T cells, resulting in priming and activation of CD8+ T cells which then infiltrate the tumor microenvironment. The cytotoxic program of these newly primed CD8+ T cells is then supported by the combined effect of TA99 and ɑ4-1BB-LAIR. The combination of TA99 and ɑ4-1BB-LAIR with a clinically approved ɑCTLA-4 antibody known for enhancing T cell priming results in equivalent cure rates, which validates the mechanistic principle, while the addition of ɑCTLA-4 also generates robust immunological memory against secondary tumor rechallenge. Thus, our study establishes the proof of principle for a clinically translatable cancer immunotherapy.

Estimating Isotope Shifts for $^{227}$Th and $^{229}$Th in Th$^{3+}$: 5F$_{5/2} \rightarrow$ 6D$_{5/2}$ Transition

2026-02-26T03:01:06Z

Estimating Isotope Shifts for $^{227}$Th and $^{229}$Th in Th$^{3+}$: 5F$_{5/2} \rightarrow$ 6D$_{5/2}$ Transition Lam, P. Y. Ian; MohanMurthy, Prajwal \textbf{Motivation:} Accurate spectroscopic information for isotopes of $^{224-232}$Th are critical for experimental programs investigating such thorium isotopes as candidates for next-generation nuclear optical clocks and as platforms for searches of symmetry-violating effects. Direct experimental data on isotope shifts in the $\text{Th}^{3+}:5F_{5/2} \to 6D_{5/2}$ line at $690$ nm are sparse, with measurements available only for $^{229}\text{Th}$ and $^{230}\text{Th}$ relative to the reference isotope $^{232}\text{Th}$.\\ \textbf{Method:} To address this gap, we employed a King-plot analysis comparing the well-characterized isotope shifts of the $\text{Th}^{+}$ transition at $583.9$ nm to the limited data available for the $690$ nm transition of $\text{Th}^{3+}$. Using nuclear structure information on mean-square charge radii and nuclear quadrupole deformations, we extracted the field-shift constant $F_{690}$ and mass-shift constant $M_{690}$ for the $690$ nm transition. We subsequently calculated the missing isotope shifts by incorporating published values of $\delta\langle r^2\rangle$ where available and estimating $\delta\langle r^2\rangle$ for unmeasured isotopes using nuclear quadrupole deformation coefficients $\beta_2$ from the FRDM model.\\ \textbf{Key Results:} The calculated isotope shifts for the $\text{Th}^{3+}:5F_{5/2} \to 6D_{5/2}$ transition relative to $690~$nm transition of $^{232}\text{Th}$ are: \begin{align} \delta\nu^{224,232}_{690} &= -29296(5585) \text{MHz} \nonumber\\ \delta\nu^{225,232}_{690} &= -25840(4930) \text{MHz} \nonumber\\ \delta\nu^{226,232}_{690} &= -22113(4219) \text{MHz} \nonumber\\ \delta\nu^{227,232}_{690} &= -18631(6238) \text{MHz} \nonumber\\ \delta\nu^{228,232}_{690} &= -14970(6181) \text{MHz} \nonumber\\ \delta\nu^{231,232}_{690} &= -3742(715) \nonumber\text{MHz} \end{align} Our results for the isotopes of $^{227,228}$Th use experimental $\delta\langle r^2\rangle$ values, whereas the remaining ones use theoretical values of $\delta\langle r^2\rangle$ calculated from quadrupole deformation coefficients. These results provide essential spectroscopic data for future precision measurements with thorium isotopes in various ionized states. This work is supported by BNL award #460913, a Phi Kappa Phi Fellowship, and generous support from Prof. R. P. Redwine and MIT LNS.

Targeting of the CD161 inhibitory receptor enhances T-cell–mediated immunity against hematological malignancies

2026-02-26T03:07:30Z

Targeting of the CD161 inhibitory receptor enhances T-cell–mediated immunity against hematological malignancies Alvarez Calderon, Francesca; Kang, Byong H; Kyrysyuk, Oleksandr; Zheng, Shiwei; Wang, Hao; Mathewson, Nathan D; Luoma, Adrienne M; Ning, Xiaohan; Pyrdol, Jason; Cao, Xuan; Suvà, Mario L; Yuan, Guo-Cheng; Wittrup, K Dane; Wucherpfennig, Kai W The CD161 inhibitory receptor is highly upregulated by tumor-infiltrating T cells in multiple human solid tumor types, and its ligand, CLEC2D, is expressed by both tumor cells and infiltrating myeloid cells. Here, we assessed the role of the CD161 receptor in hematological malignancies. Systematic analysis of CLEC2D expression using the Cancer Cell Line Encyclopedia revealed that CLEC2D messenger RNA was most abundant in hematological malignancies, including B-cell and T-cell lymphomas as well as lymphocytic and myelogenous leukemias. CLEC2D protein was detected by flow cytometry on a panel of cell lines representing a diverse set of hematological malignancies. We, therefore, used yeast display to generate a panel of high-affinity, fully human CD161 monoclonal antibodies (mAbs) that blocked CLEC2D binding. These mAbs were specific for CD161 and had a similar affinity for human and nonhuman primate CD161, a property relevant for clinical translation. A high-affinity CD161 mAb enhanced key aspects of T-cell function, including cytotoxicity, cytokine production, and proliferation, against B-cell lines originating from patients with acute lymphoblastic leukemia, diffuse large B-cell lymphoma, and Burkitt lymphoma. In humanized mouse models, this CD161 mAb enhanced T-cell–mediated immunity, resulting in a significant survival benefit. Single cell RNA-seq data demonstrated that CD161 mAb treatment enhanced expression of cytotoxicity genes by CD4 T cells as well as a tissue-residency program by CD4 and CD8 T cells that is associated with favorable survival outcomes in multiple human cancer types. These fully human mAbs, thus, represent potential immunotherapy agents for hematological malignancies.

Upcycling spent medium-Ni cathodes via novel liquified salts sourcing

2026-02-26T03:07:25Z

Upcycling spent medium-Ni cathodes via novel liquified salts sourcing Yoon, Moonsu; Park, Jin-Sung; Chen, Weiyin; Huang, Yimeng; Dai, Tao; Lee, Yumin; Shin, Jungmin; Lee, Seungmi; Kim, Yongil; Lee, Dongsoo; Shin, Daiha; Cho, Jaephil; Dong, Yanhao; Li, Ju The rapid growth in lithium-ion battery technology underscores the urgent need for sustainable recycling to address the environmental and economic challenges of battery waste. This study introduces a liquified-salts-assisted upcycling approach to transform spent medium-Ni cathodes into high-performance single-crystalline Ni-rich cathodes. Utilizing the LiOH–LiNO3–Ni(NO3)2·6H2O eutectic, this method leverages planetary centrifugal mixing to create a liquid-like environment for accelerated elemental diffusion and microstructural refinement. The in situ liquefaction of these salts ensures seamless precursor integration, achieving compositional uniformity and minimizing impurity formation. Compared to conventional solid-state methods, our method significantly suppresses rock-salt phase formation, and improves electrochemical performance with superior cycling stability and rate capability. The environmental and economic advantages of our approach highlight its potential to reduce greenhouse gas emissions and energy consumption. This scalable, energy-efficient strategy provides a transformative solution for battery waste management, paving the way for the sustainable production of next-generation cathode materials.

CLN-617 Retains IL2 and IL12 in Injected Tumors to Drive Robust and Systemic Immune-Mediated Antitumor Activity

2026-02-26T03:07:34Z

CLN-617 Retains IL2 and IL12 in Injected Tumors to Drive Robust and Systemic Immune-Mediated Antitumor Activity Mehta, Naveen K; Rakhra, Kavya; Meetze, Kristan A; Li, Bochong; Momin, Noor; Chang, Jason YH; Wittrup, K Dane; Baeuerle, Patrick A; Michaelson, Jennifer S Despite clinical evidence of antitumor activity, the development of cytokine therapies has been hampered by a narrow therapeutic window and limited response rates. Two cytokines of high interest for clinical development are interleukin 2 (IL2) and interleukin 12 (IL12), which potently synergize to promote the activation and proliferation of T cells and NK cells. However, the only approved human IL2 therapy, Proleukin, is rarely used in the clinic due to systemic toxicities, and no IL12 product has been approved to date due to severe dose-limiting toxicities. Here, we describe CLN-617, a first-in-class therapeutic for intratumoral (IT) injection that co-delivers IL2 and IL12 on a single molecule in a safe and effective manner. CLN-617 is a single-chain fusion protein comprised of IL2, leukocyte-associated immunoglobulin-like receptor 2 (LAIR2), human serum albumin (HSA), and IL12. LAIR2 and HSA function to retain CLN-617 in the treated tumor by binding collagen and increasing molecular weight, respectively. We found that IT administration of a murine surrogate of CLN-617, mCLN-617, eradicated established treated and untreated tumors in syngeneic models, significantly improved response to anti-PD1 checkpoint therapy, and generated a robust abscopal response dependent on cellular immunity and antigen cross-presentation. CLN-617 is being evaluated in a clinical trial in patients with advanced solid tumors (NCT06035744).

Porous Organic Materials-Based Atomically Dispersed Metal Electrocatalysts

2026-02-26T03:07:23Z

Porous Organic Materials-Based Atomically Dispersed Metal Electrocatalysts Zhang, Hao; Wang, Suwen; Lv, Enmin; Qi, Menghui; He, Chengchao; Dong, Xinglong; Qiu, Jieshan; Wang, Yong; Wen, Zhenhai The transition to renewable energy sources and the need for efficient energy conversion technologies have led to the development of various types of catalysts, among which atomically dispersed metal catalysts (ADMCs) supported by porous organic materials (POMs) have attracted attention for their high catalytic efficiency and stability. This review focuses on the development and application of ADMCs supported by POMs, such as MOFs, COFs, and HOFs, which offer catalytic performance due to their high atomic utilization, stability, and selectivity. This paper systematically explores various strategies for synthesizing ADMCs, including the use of organic linkers, metal nodes, and pore spaces within POMs to stabilize metal atoms and prevent aggregation. Key applications highlighted include energy conversion and storage technologies, such as fuel cells, water splitting, CO2 reduction and nitrogen reduction, where ADMCs demonstrate the potential to replace noble metals. Despite the progress, challenges remain in achieving high metal loading, long-term stability, and cost-effective large-scale production. This study underscores the importance of advanced characterization techniques and computational models to deepen the understanding of ADMCs’ catalytic mechanisms and guide future material design, paving the way for their broader application in sustainable energy technologies.

Quantifying Mining Requirement and Waste for Energy Sustainability

2026-02-26T03:07:18Z

Quantifying Mining Requirement and Waste for Energy Sustainability Ermakova, Dinara; Sen, Drishti; Wainwright, Haruko; Bae, Jin Whan; Chene, Lisha; Vujic, Jasmina This study demonstrates the life-cycle assessment of different energy sources-coal, natural gas, solar, wind, nuclear, and hydro-particularly focused on mining activities and waste per given electricity capacity and generation. It also includes carbon dioxide emissions generated during the transportation of raw materials to build and operate electricity generating systems and their environmental impacts in the US from 2023 to 2050. We identify the raw material and metal requirements for the U.S.-based typical systems in each energy type and synthesize datasets on typical ore fraction and material recycling factors, while taking into account the capacity factor of the power plants. We then compute the total mass and volume of material requirements and waste mass and volume for the front-end (i.e., mining, material needed for construction), operation (i.e., fuel, maintenance), and back-end (i.e., decommissioning) activities. The key findings are that (1) the energy transition from fossil fuel to low-carbon energy sources would reduce mining waste as well as the shipping carbon footprint; (2) the difference in capacity and actual electricity generation is significant for the life-cycle assessment due to low capacity factors of solar and wind energy; (3) several key metals with low abundance or high requirements dominate mining waste, which highlights the need for recycling and establishing a circular economy; (4) mining of critical minerals becomes important during the clean energy transition and (5) nuclear energy generates least waste and contributes least to shipping emissions among the low-carbon sources due to the high energy density and capacity factor and the small mass of materials it requires. Although the waste mass may not necessarily be equal to the environmental impact due to different waste isolation technologies, we aim to highlight the importance of considering mining and decommissioning waste, which are often ignored but important for accounting for the environmental impacts and addressing energy justice issues.

The Scholarly Knowledge Ecosystem: Challenges and Opportunities for the Field of Information

2026-02-26T03:07:41Z

The Scholarly Knowledge Ecosystem: Challenges and Opportunities for the Field of Information Altman, Micah; Cohen, Philip N The scholarly knowledge ecosystem presents an outstanding exemplar of the challenges of understanding, improving, and governing information ecosystems at scale. This article draws upon significant reports on aspects of the ecosystem to characterize the most important research challenges and promising potential approaches. The focus of this review article is the fundamental scientific research challenges related to developing a better understanding of the scholarly knowledge ecosystem. Across a range of disciplines, we identify reports that are conceived broadly, published recently, and written collectively. We extract the critical research questions, summarize these using quantitative text analysis, and use this quantitative analysis to inform a qualitative synthesis. Three broad themes emerge from this analysis: the need for multi-sectoral cooperation and coordination, for mixed methods analysis at multiple levels, and interdisciplinary collaboration. Further, we draw attention to an emerging consensus that scientific research in this area should by a set of core human values.

Investigation of L2/Ln Pragmatic Competence: Its Core and Route Map

2026-02-26T03:07:44Z

Investigation of L2/Ln Pragmatic Competence: Its Core and Route Map Mao, Tiaoyuan How to use language properly and acquire the capacity for language use has become the focus of linguists and philosophers for centuries. Therefore, pragmatic competence underlying language use arouses enormous interests of language acquisition practitioners. This study reveals the core properties of various models or theories of pragmatic competence, such as the communicative componential models, the form-function mapping proposal of the functionalist, the tripartite cognitive model, and the current integrated model of pragmatic competence. The common core includes (but not limited to) integration of thought and communication, one uniform pragmatic mechanism, dynamic form-function mapping, and complementarity between grammatical and pragmatic competences. With the findings as a departure, a brief outline for further investigation of pragmatic competence is proposed finally, including pathological and neurobiological examination of pragmatic competence.

Investigating Structural Biophysical Features for Antigen-Binding Fragment Crystallization via Machine Learning

2026-02-26T03:07:12Z

Investigating Structural Biophysical Features for Antigen-Binding Fragment Crystallization via Machine Learning Chattaraj, Krishna Gopal; Ferreira, Joana; Myerson, Allan S.; Trout, Bernhardt L. Antibody-based therapeutics continue to be an important pharmaceutical development modality. Crystallization of antibodies is important for structural characterization, but in addition has the potential for use as a separation method and for use as a dosage form. Nevertheless, bringing about controlled crystallization of an antibody remains a challenging task due to its large size, high degree of segmental flexibility, and the intricacy of all the occurring interactions (e.g., protein–protein interactions, protein–solvent interactions, etc.). Methods to predict important contact sites could help to develop such crystallization methods. However, limited data and understanding have hitherto not allowed the development of such robust methods. This study employs machine learning combined with in silico modelling of crystal structures using available experimental structures to identify the crucial physicochemical features necessary for successful antibody crystallization in an attempt to remedy that gap. The developed method can with good accuracy distinguish crystal-site residues from non-crystal-site residues. A set of 510 descriptors is utilized to characterize each residue, which is treated as a distinct data point. Moreover, new algorithms have been developed to design novel descriptors that improve the model's predictive capabilities. Fragment antigen-binding (Fab) regions are investigated due to the scarcity of full-length monoclonal antibodies (mAbs) crystal structures. The current findings show that the extreme gradient boosting (XGBoost) algorithm effectively identifies crystal site residues, as evidenced by an AUPRC value that is more than 3-fold higher than that of the baseline model. The top-ranked descriptors indicate that crystal-site residues are primarily characterized by solvent-exposed residues with high spatial aggregation propensity (SAP), signifying hydrophobic patches, and their immediate surface-exposed neighbors. Moreover, these high SAP residues are often surrounded by other solvent-exposed residues that are either polar, charged, or both. In contrast, residues not involved in crystal interfaces generally lack these essential features, though some might be excluded due to specific crystal lattice arrangements. Additionally, reducing the feature set from 510 to the top 15% in the XGBoost model yields similar performance while significantly simplifying the model.

Modeling the Role of Supramolecular Clustering in Multivalent Assembly

2026-02-26T03:07:14Z

Modeling the Role of Supramolecular Clustering in Multivalent Assembly Sbalbi, Nicholas; Petrov, Artem; Sass, Jacob; Ye, Matthew; Alexander-Katz, Alfredo; Macfarlane, Robert J. In self-assembled systems, a combination of multiple weak supramolecular interactions is often utilized to enable strong yet reversible binding. When modeling the behavior of these multivalent interfaces, it is commonly assumed that binding pairs are independent, i.e., the probability of a pair being bound is unaffected by the bound state of neighboring pairs. Inspired by recent experimental work, we report that for a variety of systems this assumption may not hold, leading to the formation of clusters at the binding interface. Through a series of analytical and numerical models of end-functionalized brushes, we reveal the role of cluster size on binding thermodynamics, detail how entropic contributions from polymer chains provide tunable control of cluster size, and provide predictions for cluster size as a function of system architecture. Investigation of these models yields surprising results: within the melting window, the enthalpy of binding of multivalent interfaces is predicted to depend only on cluster size and not on the overall valency of the multivalent system. Moreover, clustering is predicted to be significant even in systems with only weak dipole and dispersion interactions between neighboring groups. Combined, this work brings to light the potential impacts of clustering on multivalent self-assembly, providing theoretical justification for previous experimental observations and paving the way for future work in this area.

Elucidating the effect of Fe substitution on structural and redox stability of Na2Mn3O7

2026-02-26T03:07:21Z

Elucidating the effect of Fe substitution on structural and redox stability of Na2Mn3O7 Smith, Hugh B.; Lee, Gi-Hyeok; Kumar, Bachu Sravan; Penn, Aubrey N.; Venturi, Victor; Gao, Yifan; Davis, Ryan C.; Stone, Kevin Hunter; Hunt, Adrian; Waluyo, Iradwikanari; Stavitski, Eli; Yangb, Wanli; Abate, Iwnetim I. Sodium-ion batteries have the potential to meet the growing demand for energy storage due to their low costs stemming from natural resource abundances, but their cathode energy densities must be improved to be comparable to those of lithium-ion batteries. One strategy is accessing high voltage capacity through high-valent redox reactions. Such reactions usually cause instability in cathode materials, but Na2Mn3O7 (NMO) has demonstrated excellent performance and reversibility in the high-valent regime due to its unique lattice structure with ordered Mn vacancies. This work expands the universality of the ordered vacancy as a design principle and increases the material candidates with such exceptional electrochemical behavior. Our approach involves synergizing cationic ordered vacancies with tunable metal–ligand hybridization through partial metal substitution. In particular, we successfully incorporated Fe3+ for Mn4+ in NMO to make Na2.25Mn2.75Fe0.25O7 and achieved improved high-valent redox behavior. Fe substitution leads to larger specific capacities (171 vs. 159 mA h g−1 first cycle), enhanced cycle stability (97 vs. 60 mA h g−1 after 50 cycles), and superior rate performance. This study lays the foundation for developing new cathode materials with stable high-valent redox through substitution of redox-active transition metals by employing cationic ordered vacancies and partial transition metal substitution as design principles in tandem.

Progress in Computational Methods and Mechanistic Insights on the Growth of Carbon Nanotubes

2026-02-26T03:07:16Z

Progress in Computational Methods and Mechanistic Insights on the Growth of Carbon Nanotubes Wang, Linzheng; Tricard, Nicolas; Chen, Zituo; Deng, Sili Carbon nanotubes (CNTs), as a promising nanomaterial with broad applications across various fields, are continuously attracting significant research attention. Despite substantial progress in understanding their growth mechanisms, synthesis methods, and post-processing techniques, two major goals remain challenging: achieving property-targeted growth and efficient mass production. Recent advancements in computational methods driven by increased computational resources, the development of platforms, and the refinement of theoretical models, have significantly deepened our understanding of the mechanisms underlying CNT growth. This review aims to comprehensively examine the latest computational techniques that shed light on various aspects of CNT synthesis. The first part of this review focuses on progress in computational methods. Beginning with atomistic simulation approaches, we introduce the fundamentals and advancements in density functional theory (DFT), molecular dynamics (MD) simulations, and kinetic Monte Carlo (kMC) simulations. We discuss the applicability and limitations of each method in studying mechanisms of CNT growth. Then, the focus shifts to multiscale modeling approaches, where we demonstrate the coupling of atomic-scale simulations with reactor-scale multiphase flow models. Given that CNT growth inherently spans multiple temporal and spatial scales, the development and application of multiscale modeling techniques are poised to become a central focus of future computational research in this field. Furthermore, this review emphasizes the growing role played by machine learning in CNT growth research. Compared with traditional physics-based simulation methods, data-driven machine learning approaches have rapidly emerged in recent years, revolutionizing research paradigms from molecular simulation to experimental design. In the second part of this review, we highlight the latest advancements in CNT growth mechanisms and synthesis methods achieved through computational techniques. These include novel findings across fundamental growth stages, i.e., from nucleation to elongation and ultimately termination. We also examine the dynamic behaviors of catalyst nanoparticles and chirality-controlled growth processes, emphasizing how these insights contribute to advancing the field. Finally, in the concluding section, we propose future directions for advancements of computational approaches toward deeper understanding of CNT growth mechanisms and better support of CNT manufacturing.

Superthermal Solar Interfacial Evaporation is not due to Reduced Latent Heat of Water

2026-02-26T03:07:09Z

Superthermal Solar Interfacial Evaporation is not due to Reduced Latent Heat of Water Zhang, James H.; Mittapally, Rohith; Lva, Guangxin; Chen, Gang To explain reported solar interfacial-evaporation rates from porous materials beyond an apparent 100% efficiency using the thermal evaporation mechanism, many publications hypothesize that intermediate water inside porous materials has a reduced latent heat. Key supporting evidence is that water-only surfaces have lower natural evaporation rates than porous evaporators, with the ratio of the two rates taken as the latent heat reduction. Through simulations and experiments, we study natural evaporation of water and show that reported differences in evaporation rates between porous materials and water are likely due to experimental error from recessed evaporating surfaces. A few millimeter recession of the water surface relative to the container lip can drop evaporation rates by over 50% due to a stagnant air layer, suggesting that the comparative experiments are prone to error. Furthermore, in the reduced latent heat picture, interfacial cooling must occur at the porous sample–water interface due to the enthalpy difference between bulk water and intermediate water. Our transport modeling shows that reduced latent heat cannot explain superthermal evaporation and that new mechanistic directions need to be pursued.

Generative design and molecular mechanics characterization of silk proteins based on unfolding behavior

2026-02-26T03:07:20Z

Generative design and molecular mechanics characterization of silk proteins based on unfolding behavior Lu, Wei; Buehler, Markus J. Spider silk exhibits exceptional mechanical properties, biocompatibility, and biodegradability, making it a promising material for bioengineered applications. However, the complexity and diversity of silk proteins, coupled with limited experimental data, have hindered the rational design of silk-based biomaterials. Furthermore, the mechanobiology of these proteins and their impact on silk fiber properties remain underexplored. In this study, we introduce a series of novel silk protein sequences and characterize their nonlinear unfolding behavior and mechanical properties through molecular dynamics (MD) simulations. Focusing on major ampullate spidroin (MaSp) silk proteins, we curate a dataset that integrates experimentally acquired sequences with synthetic sequences generated by SilkomeGPT, a generative model for silk-inspired proteins. Structural predictions are performed using OmegaFold, from which high-fidelity regions are extracted and analyzed. Their unfolding responses are assessed via implicit all-atom MD simulations, enabling characterization of their mechanical behavior. This computationally efficient framework facilitates the rational design of spider silk proteins by linking atomistic and sequence features to larger-scale properties. The developed dataset systematically captures structural uncertainties, while simulations provide atomic-level insights into how protein mechanics contribute to fiber properties, advancing the mechanobiological understanding of spider silk and supporting diverse applications in biomaterials design.

Tumor-Localized Interleukin-2 and Interleukin-12 Combine with Radiation Therapy to Safely Potentiate Regression of Advanced Malignant Melanoma in Pet Dogs

2026-02-25T07:11:58Z

Tumor-Localized Interleukin-2 and Interleukin-12 Combine with Radiation Therapy to Safely Potentiate Regression of Advanced Malignant Melanoma in Pet Dogs Stinson, Jordan A; Barbosa, Matheus Moreno P; Sheen, Allison; Momin, Noor; Fink, Elizabeth; Hampel, Jordan; Selting, Kim A; Kamerer, Rebecca L; Bailey, Keith L; Wittrup, Karl D; Fan, Timothy M Purpose: Cytokines IL2 and IL12 exhibit potent anticancer activity but suffer a narrow therapeutic window due to off-tumor immune cell activation. Engineering cytokines with the ability to bind and associate with tumor collagen after intratumoral injection potentiated response without toxicity in mice and was previously safe in pet dogs with sarcoma. Here, we sought to test the efficacy of this approach in dogs with advanced melanoma. Patients and Methods: This study examined 15 client-owned dogs with histologically or cytologically confirmed malignant melanoma that received a single 9-Gy fraction of radiotherapy, followed by six cycles of combined collagen-anchored IL2 and IL12 therapy every 2 weeks. Cytokine dosing followed a 3 + 3 dose escalation design, with the initial cytokine dose chosen from prior evaluation in canine sarcomas. No exclusion criteria for tumor stage or metastatic burden, age, weight, or neuter status were applied for this trial. Results: Median survival regardless of the tumor stage or dose level was 256 days, and 10/13 (76.9%) dogs that completed treatment had CT-measured tumor regression at the treated lesion. In dogs with metastatic disease, 8/13 (61.5%) had partial responses across their combined lesions, which is evidence of locoregional response. Profiling by NanoString of treatment-resistant dogs revealed that B2m loss was predictive of poor response to this therapy. Conclusions: Collectively, these results confirm the ability of locally administered tumor-anchored cytokines to potentiate responses at regional disease sites when combined with radiation. This evidence supports the clinical translation of this approach and highlights the utility of comparative investigation in canine cancers.

Bivalent target-binding bioPROTACs induce potent degradation of oncogenic SHP2

2026-02-25T07:11:56Z

Bivalent target-binding bioPROTACs induce potent degradation of oncogenic SHP2 Hoffman, Megan; Krum, David; Wittrup, K Dane Targeted protein degradation is an emergent and rapidly evolving therapeutic strategy. In particular, biologics-based targeted degradation modalities (bioPROTACs) are relatively under explored compared to small molecules. Here, we investigate how target affinity, cellular localization, and valency of bioPROTACs impact efficacy of targeted degradation of the oncogenic phosphatase src-homology 2 containing protein tyrosine phosphatase-2 (SHP2). We identify bivalent recruitment of SHP2 by bioPROTACs as a broadly applicable strategy to improve potency. Moreover, we demonstrate that SHP2-targeted bioPROTACs can effectively counteract gain-of-function SHP2 mutants present in cancer, which are otherwise challenging to selectively target with small molecule constructs. Overall, this study demonstrates the utility of bioPROTACs for challenging targets, and further explicates design principles for therapeutic bioPROTACs.

Local delivery of cell surface-targeted immunocytokines programs systemic antitumor immunity

2026-02-25T07:11:50Z

Local delivery of cell surface-targeted immunocytokines programs systemic antitumor immunity Santollani, Luciano; Maiorino, Laura; Zhang, Yiming J; Palmeri, Joseph R; Stinson, Jordan A; Duhamel, Lauren R; Qureshi, Kashif; Suggs, Jack R; Porth, Owen T; Pinney, William; Msari, Riyam Al; Walsh, Agnes A; Wittrup, K Dane; Irvine, Darrell J Systemically administered cytokines are potent immunotherapeutics but can cause severe dose-limiting toxicities. To overcome this challenge, cytokines have been engineered for intratumoral retention after local delivery. However, despite inducing regression of treated lesions, tumor-localized cytokines often elicit only modest responses at distal untreated tumors. In the present study, we report a localized cytokine therapy that safely elicits systemic antitumor immunity by targeting the ubiquitous leukocyte receptor CD45. CD45-targeted immunocytokines have lower internalization rates relative to wild-type counterparts, leading to sustained downstream cis and trans signaling between lymphocytes. A single intratumoral dose of αCD45-interleukin (IL)-12 followed by a single dose of αCD45-IL-15 eradicated treated tumors and untreated distal lesions in multiple syngeneic mouse tumor models without toxicity. Mechanistically, CD45-targeted cytokines reprogrammed tumor-specific CD8+ T cells in the tumor-draining lymph nodes to have an antiviral transcriptional signature. CD45 anchoring represents a broad platform for protein retention by host immune cells for use in immunotherapy.

Tumor Integrin-Targeted Glucose Oxidase Enzyme Promotes ROS-Mediated Cell Death that Combines with Interferon Alpha Therapy for Tumor Control

2026-02-25T07:12:00Z

Tumor Integrin-Targeted Glucose Oxidase Enzyme Promotes ROS-Mediated Cell Death that Combines with Interferon Alpha Therapy for Tumor Control Stinson, Jordan A; Sheen, Allison; Lax, Brianna M; Yang, Grace N; Duhamel, Lauren; Santollani, Luciano; Fink, Elizabeth; Palmeri, Joseph R; Wittrup, Karl Dane Although heightened intratumoral levels of reactive oxygen species (ROS) are typically associated with a suppressive tumor microenvironment, under certain conditions ROS contribute to tumor elimination. Treatment approaches, including some chemotherapy and radiation protocols, increase cancer cell ROS levels that influence their mechanism of cell death and subsequent recognition by the immune system. Furthermore, activated myeloid cells rapidly generate ROS upon encounter with pathogens or infected cells to eliminate disease, and recently, this effector function has been noted in cancer contexts as well. Collectively, ROS-induced cancer cell death may help initiate adaptive antitumor immune responses that could synergize with current approved immunotherapies, for improved control of solid tumors. In this work, we explore the use of glucose oxidase, an enzyme which produces hydrogen peroxide, a type of ROS, to therapeutically mimic the endogenous oxidative burst from myeloid cells to promote antigen generation within the tumor microenvironment. We engineer the enzyme to target pan-tumor-expressed integrins both as a tumor-agnostic therapeutic approach and as a strategy to prolong local enzyme activity following intratumoral administration. We found the targeted enzyme potently induced cancer cell death and enhanced cross-presentation by dendritic cells in vitro and further combined with interferon alpha for long-term tumor control in murine MC38 tumors in vivo. Optimizing the single-dose administration of this enzyme overcomes limitations with immunogenicity noted for other prooxidant enzyme approaches. Overall, our results suggest ROS-induced cell death can be harnessed for tumor control and highlight the potential use of designed enzyme therapies alongside immunotherapy against cancer.

Yeast as a tool for exploring disulfide-rich peptides

2026-02-25T07:12:01Z

Yeast as a tool for exploring disulfide-rich peptides Yap, Kuok; Porth, Owen T; Xie, Jing; Wang, Conan K; Durek, Thomas; Wittrup, K Dane; Craik, David J Cyclic disulfide-rich peptides have become increasingly popular in drug development because their structures enhance molecular stability and allow for mutagenesis to introduce non-native functions. This review focuses on yeast-based platform technologies and their utility in advancing cyclic disulfide-rich peptides as drug modalities and for large-scale biomanufacturing. These technologies include yeast surface display which facilitates the screening of large libraries to develop peptide binders with strong affinity and selectivity for protein targets, while maintaining the innate high stability of the peptide scaffold via protease-based selection pressure. We also describe a recently developed platform that leverages yeast’s ability to secrete correctly folded disulfide-rich peptides while simultaneously displaying peptide or protein tags on their surfaces. In combination with microfluidics technology, the platform creates single-cell yeast-in-droplets reactors, enabling the screening of large libraries based on functional output rather than solely on binding affinity. After identifying cyclic peptide candidates through library-based discovery, these candidates can be produced using a versatile yeast-based bioproduction platform. Traditionally, cyclic disulfide-rich peptides are produced through solid-phase synthesis, a method that generates significant amounts of toxic waste. In contrast, yeast-based bioproduction offers an environmentally sustainable alternative. It has the capability to produce structurally distinct peptides with minimal adjustments and is easily scalable using microbial fermenters, making it an ideal choice for large-scale production.

Aligning supply chain design for boosting resilience

2026-02-25T07:12:02Z

Aligning supply chain design for boosting resilience Sáenz, María Jesús; Revilla, Elena; Acero, Beatriz Many researchers have analyzed the effect of disruptive events, such as natural disasters and economic and market forces, on global supply chains. However, there is a lack of consensus on delineating a universal collection of supply chain risk management practices that will help companies operate in a global market with large-scale disruptions. In this article, we present an analysis, in conjunction with a worldwide online survey, based on successful global brands and their supply chains. We propose a framework that deploys the dynamics of building supply chain resilience, first linking the design of the supply chain portfolio (local versus global scope, as well as strategic responsiveness versus cost reduction) with supply chain vulnerabilities (external versus internal). We describe the transition between different supply chain structures as a way of coping with disruptions and thus proactively developing resilience. In this article, we introduce both a supply chain risk management approach and the reactive-by-deployment mode, as illustrated by successful global company examples.

Pennies, Penny Pools

2026-02-25T03:01:14Z

Pennies, Penny Pools Bucciarelli, Louis On Feburay 25, 2025, our president ordered the US Treasury to cease minting pennies. In this essay, I recount how members in Congress have tried to legislate the same - calling for rounding to the nearest nickel - but have not succeeded, some fearing that prevailing price structures (e.g., $xx.89) would leave customers on the short end of the stick over time with rounding up (customer loses) being more prevalent than rounding down (customer wins). I analyze the situation, then suggest how a “penny pool” might be used by retailers to ease the transition to a just and fair penny-less society.

Directed evolution-based discovery of ligands for in vivo restimulation of chimeric antigen receptor T cells

2026-02-24T03:07:59Z

Directed evolution-based discovery of ligands for in vivo restimulation of chimeric antigen receptor T cells Grzywa, Tomasz M; Neeser, Alexandra; Ramasubramanian, Ranjani; Romanov, Anna; Tannir, Ryan; Mehta, Naveen K; Cossette, Benjamin; Morgan, Duncan M; Goncalves, Beatriz; Sukaj, Ina; Bergaggio, Elisa; Kadauke, Stephan; Myers, Regina M; Paruzzo, Luca; Ghilardi, Guido; Cozzone, Austin; Schuster, Stephen J; Frey, Noelle; Zhang, Libin; Yousefpour, Parisa; Abraham, Wuhbet; Suh, Heikyung; Ruella, Marco; Grupp, Stephan A; Chiarle, Roberto; Wittrup, K Dane; Ma, Leyuan; Irvine, Darrell J Chimeric antigen receptor (CAR) T cell therapy targeting CD19 elicits remarkable clinical efficacy in B cell malignancies, but many patients relapse owing to failed expansion and/or progressive loss of CAR-T cells. We recently reported a strategy to potently restimulate CAR-T cells in vivo, enhancing their functionality by administration of a vaccine-like stimulus comprised of surrogate peptide ligands for a CAR linked to a lymph node-targeting amphiphilic PEG-lipid (amph-vax). Here we demonstrate a general strategy to discover and optimize peptide mimotopes enabling amph-vax generation for any CAR. We use yeast surface display to identify peptide binders to FMC63 (the scFv used in clinical CD19 CARs), which are then subsequently affinity matured by directed evolution. CAR-T vaccines using these optimized mimotopes triggered marked expansion and memory development of CD19 CAR-T cells in both syngeneic and humanized mouse models of B-acute lymphoblastic leukaemia/lymphoma, and enhanced control of disease progression compared with CD19 CAR-T-only-treated mice. This approach enables amph-vax boosting to be applied to any clinically relevant CAR-T cell product.

Machine learning prediction of antibody aggregation and viscosity for high concentration formulation development of protein therapeutics

2026-03-08T03:40:22Z

Machine learning prediction of antibody aggregation and viscosity for high concentration formulation development of protein therapeutics Lai, Pin-Kuang; Gallegos, Austin; Mody, Neil; Sathish, Hasige A; Trout, Bernhardt L Machine learning has been recently used to predict therapeutic antibody aggregation rates and viscosity at high concentrations (150 mg/ml). These works focused on commercially available antibodies, which may have been optimized for stability. In this study, we measured accelerated aggregation rates at 45°C and viscosity at 150 mg/ml for 20 preclinical and clinical-stage antibodies. Features obtained from molecular dynamics simulations of the full-length antibody and sequences were used for machine learning model construction. We found a k-nearest neighbors regression model with two features, spatial positive charge map on the CDRH2 and solvent-accessible surface area of hydrophobic residues on the variable fragment, gives the best performance for predicting antibody aggregation rates (r = 0.89). For the viscosity classification model, the model with the highest accuracy is a logistic regression model with two features, spatial negative charge map on the heavy chain variable region and spatial negative charge map on the light chain variable region. The accuracy and the area under precision recall curve of the classification model from validation tests are 0.86 and 0.70, respectively. In addition, we combined data from another 27 commercial mAbs to develop a viscosity predictive model. The best model is a logistic regression model with two features, number of hydrophobic residues on the light chain variable region and net charges on the light chain variable region. The accuracy and the area under precision recall curve of the classification model are 0.85 and 0.6, respectively. The aggregation rates and viscosity models can be used to predict antibody stability to facilitate pharmaceutical development.

Enhanced O-glycosylation site prediction using explainable machine learning technique with spatial local environment

2026-03-08T03:40:22Z

Enhanced O-glycosylation site prediction using explainable machine learning technique with spatial local environment Hong, Seokyoung; Chattaraj, Krishna Gopal; Guo, Jing; Trout, Bernhardt L; Braatz, Richard D Motivation: The accurate prediction of O-GlcNAcylation sites is crucial for understanding disease mechanisms and developing effective treatments. Previous machine learning (ML) models primarily relied on primary or secondary protein structural and related properties, which have limitations in capturing the spatial interactions of neighboring amino acids. This study introduces local environmental features as a novel approach that incorporates three-dimensional spatial information, significantly improving model performance by considering the spatial context around the target site. Additionally, we utilize sparse recurrent neural networks to effectively capture sequential nature of the proteins and to identify key factors influencing O-GlcNAcylation as an explainable ML model. Results: Our findings demonstrate the effectiveness of our proposed features with the model achieving an F1 score of 28.3%, as well as feature selection capability with the model using only the top 20% of features achieving the highest F1 score of 32.02%, a 1.4-fold improvement over existing PTM models. Statistical analysis of the top 20 features confirmed their consistency with literature. This method not only boosts prediction accuracy but also paves the way for further research in understanding and targeting O-GlcNAcylation. Availability and implementation: The entire code, data, features used in this study are available in the GitHub repository: https://github.com/ pseokyoung/o-glcnac-

Leveraging microtopography to pattern multi-oriented muscle actuators

2026-02-20T03:08:08Z

Leveraging microtopography to pattern multi-oriented muscle actuators Rossy, Tamara; Schwendeman, Laura; Kohli, Sonika; Bawa, Maheera; Umashankar, Pavankumar; Habba, Roi; Tchaicheeyan, Oren; Lesmanbc, Ayelet; Raman, Ritu Engineering skeletal muscle tissue with precisely defined alignment is of significant importance for applications ranging from drug screening to biohybrid robotics. Aligning 2D contractile muscle monolayers, which are compatible with high-content imaging and can be deployed in planar soft robots, typically requires micropatterned cues. However, current protocols for integrating microscale topographical features in extracellular matrix hydrogels require expensive microfabrication equipment and multi-step procedures involving error-prone manual handling steps. To address this challenge, we present STAMP (simple templating of actuators via micro-topographical patterning), an easily accessible and cost-effective one-step method to pattern microtopography of various sizes and configurations on the surface of hydrogels using reusable 3D printed stamps. We demonstrate that STAMP enables precisely controlling the alignment of mouse and human skeletal muscle fibers without negatively impacting their maturation or function. To showcase the versatility of our technique, we designed a planar soft robot inspired by the iris, which leverages spatially segregated regions of concentric and radial muscle fibers to control pupil dilation. Optogenetic skeletal muscle fibers grown on a STAMPed iris substrates formed a multi-oriented actuator, and selective light stimulation of the radial and concentric fibers was used to control the function of the iris, including pupil constriction. Computational modeling of the biohybrid robot as an active bilayer matched experimental outcomes, showcasing the robustness of our STAMP method for designing, fabricating, and testing planar biohybrid robots capable of complex multi-DOF motion.

Enhancing Spray Retention Using Cloaked Droplets to Reduce Pesticide Pollution

2026-02-20T03:08:05Z

Enhancing Spray Retention Using Cloaked Droplets to Reduce Pesticide Pollution Jayaprakash, Vishnu; Rufer, Simon; Panata, Sreedath; Varanasi, Kripa K. Enhancing agrochemical spray retention on plant surfaces would have tremendous benefits to global health and the environment. The bouncing of sprayed pesticide droplets from hydrophobic leaves is a major source of water and soil pollution, and the resultant overuse of pesticides is a human health hazard and a financial burden for farmers. Here we report on the development of sustainable agricultural sprays consisting of cloaked droplets that significantly enhance droplet retention on plant surfaces. By leveraging wetting dynamics, we create cloaked droplets that consist of an ultra-thin food and environmentally safe oil layer (<1% by volume) that encapsulates water droplets. We develop a fundamental understanding of the dynamics of cloaked droplet impact and retention on superhydrophobic surfaces. Using high-speed imaging, we capture how the oil cloak transforms into a wetting ridge that pins the droplets and suppresses their rebound. We span a wide range of impact conditions, oils, oil viscosities, and oil volume fractions to demonstrate the robustness of the approach. By considering a balance of kinetic energy, the work of adhesion, and viscous dissipation in this four-phase system, we develop a physical model that allows us to establish a regime map for rebound suppression. Finally, these findings are implemented into a prototype sprayer which leads to a ∼5-fold reduction in spray waste on crop leaves. We believe that our spray approach can greatly reduce agrochemical pollution as well as pesticide and surfactant usage.

Total Synthesis and 13C NMR Revision of Nagelamide C

2026-02-20T03:08:04Z

Total Synthesis and 13C NMR Revision of Nagelamide C Tong, Guanghu; Nguyen, Long V.; Jamison, Timothy F. Nagelamide C (1), a dimeric pyrrole–imidazole alkaloid, exhibits antimicrobial and antibacterial activities. We demonstrate herein the first total synthesis of nagelamide C. This concise work was enabled by a series of significant transformations featuring: an imidazole benzylic Wittig olefination, a site selective bromination, and a regioselective trans-hydrostannylation/Stille coupling to construct a unique trisubstituted olefin. In addition, we show the original 13C NMR data of nagelamide C to be in error and revise the data.

Symmetry-Constrained Generation of Diverse Low-Bandgap Molecules with Monte Carlo Tree Search

2026-02-20T03:08:03Z

Symmetry-Constrained Generation of Diverse Low-Bandgap Molecules with Monte Carlo Tree Search Subramanian, Akshay; Damewood, James; Nam, Juno; Greenman, Kevin P.; Singhal, Avni P.; Gómez-Bombarelli, Rafael Organic optoelectronic materials are a promising avenue for next-generation electronic devices due to their solution processability, mechanical flexibility, and tunable electronic properties. In particular, near-infrared (NIR) sensitive molecules have unique applications in night-vision equipment and biomedical imaging. Molecular engineering has played a crucial role in developing non-fullerene acceptors (NFAs) such as the Y-series molecules, which feature a rigid fused-ring electron donor core flanked by electron-deficient end groups, leading to strong intramolecular charge-transfer and extended absorption into the NIR region. However, systematically designing molecules with targeted optoelectronic properties while ensuring synthetic accessibility remains a challenge. To address this, we leverage structural priors from domain-focused, patent-mined datasets of organic electronic molecules using a symmetry-aware fragment decomposition algorithm and a fragment-constrained Monte Carlo Tree Search (MCTS) generator. Our approach generates candidates that retain symmetry constraints from the patent dataset, while also exhibiting red-shifted absorption, as validated by TD-DFT calculations.

Molecular analysis and design using generative artificial intelligence via multi-agent modeling

2026-02-20T03:08:09Z

Molecular analysis and design using generative artificial intelligence via multi-agent modeling Stewart, Isabella; Buehler, Markus J. We report the use of a multiagent generative artificial intelligence framework, the X-LoRA-Gemma large language model (LLM), to analyze, design and test molecular design. The X-LoRA-Gemma model, inspired by biological principles and featuring 7 billion parameters, dynamically reconfigures its structure through a dual-pass inference strategy to enhance its problem-solving abilities across diverse scientific domains. The model is used to first identify molecular engineering targets through a systematic human–AI and AI–AI self-driving multi-agent approach to elucidate key targets for molecular optimization to improve interactions between molecules. Next, a multi-agent generative design process is used that includes rational steps, reasoning and autonomous knowledge extraction. Target properties of the molecule are identified either using a principal component analysis (PCA) of key molecular properties or sampling from the distribution of known molecular properties. The model is then used to generate a large set of candidate molecules, which are analyzed via their molecular structure, charge distribution, and other features. We validate that as predicted, increased dipole moment and polarizability is indeed achieved in the designed molecules. We anticipate an increasing integration of these techniques into the molecular engineering workflow, ultimately enabling the development of innovative solutions to address a wide range of societal challenges. We conclude with a critical discussion of challenges and opportunities of the use of multi-agent generative AI for molecular engineering, analysis and design.

Automated fast-flow synthesis of the immune checkpoint receptors PD-1 and PD-L1

2026-02-20T03:08:10Z

Automated fast-flow synthesis of the immune checkpoint receptors PD-1 and PD-L1 Fittolani, Giulio; Callahan, Alex J.; Loas, Andrei; Pentelute, Bradley L. Programmed cell death protein 1 (PD-1) and programmed cell death ligand 1 (PD-L1) are key targets for cancer therapy. Here, we use automated fast-flow peptide synthesis (AFPS) to rapidly produce these challenging β-sheet-rich proteins in their active forms following oxidative refolding protocols. The methods presented here provide rapid access to synthetic, air-stable mutants of PD-1 and PD-L1 in which L-methionine residues are substituted with L-norleucine, potentially enabling investigation of post-translational modifications and mirror-image analogs for drug discovery.

Geology of Deception Gulch and the Verde Central mine

2026-02-20T03:04:15Z

Geology of Deception Gulch and the Verde Central mine Benedict, P. C. (Platt Carrico), 1900-1969. Thesis: M.S., Massachusetts Institute of Technology, Department of Geology and Geophysics, 1923

Structural geology of Eastern Massachusetts

2026-02-20T03:02:14Z

Structural geology of Eastern Massachusetts Ilsley, Ralph, 1896- Thesis: Sc. D., Massachusetts Institute of Technology, Department of Geology, 1934; Vita.

The design of a hydraulic draft gear for railway passenger cars

2026-02-20T03:05:00Z

The design of a hydraulic draft gear for railway passenger cars Pearson, Harry L.; McGrady, Charles T. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1922; Includes bibliographical references.

A study of "Chu-Ma" as a textile fiber

2026-02-20T03:04:57Z

A study of "Chu-Ma" as a textile fiber Chou, Cheng Yu, 1901-; Hsueh, Tsu Kang. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1924

The transportation decision making process in metropolitan Boston

2026-02-20T03:04:52Z

The transportation decision making process in metropolitan Boston Zinner, Richard Mark. Thesis: B.S., Massachusetts Institute of Technology, Department of Political Science, 1967; One unnumbered page inserted.; Bibliography: leaf 74.

Oscillographic presentation of impedances on the reflection-coefficient plane

2026-02-20T03:04:47Z

Oscillographic presentation of impedances on the reflection-coefficient plane Eckhart, Myron.; Fowler, Earl Bealle. Thesis: B.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1949

Radiation transfer in massive binary x-ray systems

2026-02-20T03:02:25Z

Radiation transfer in massive binary x-ray systems Lewis, Wayne Lloyd. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Physics, 1991; Includes bibliographical references (leaves 167-173).

Laser induced photoionization of helium

2026-02-20T03:04:45Z

Laser induced photoionization of helium Lewis, Wayne Lloyd. Thesis: B.S., Massachusetts Institute of Technology, Department of Physics, 1980; Includes bibliographical references.

Nonlinear elastic analysis of reinforced concrete structures by the finite element method

2026-02-20T03:04:07Z

Nonlinear elastic analysis of reinforced concrete structures by the finite element method Tulga, Said Şahin. Thesis: M.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1979; Includes bibliographical references.

Subcontractor bidding strategy

2026-02-20T03:04:10Z

Subcontractor bidding strategy Gilbane, Thomas Freeman. Thesis: M.S., Massachusetts Institute of Technology, Department of Civil Engineering, 1975; Bibliography: leaves 104-105.

Petrography and geology of the Shoshone mining region in northwestern Wyoming

2026-02-20T03:04:35Z

Petrography and geology of the Shoshone mining region in northwestern Wyoming Benedict, P. C. (Platt Carrico), 1900-1969. Thesis: B.S., Massachusetts Institute of Technology, Department of Geology, 1922

GeoXCP: uncertainty quantification of spatial explanations in explainable AI

2026-02-20T03:08:01Z

GeoXCP: uncertainty quantification of spatial explanations in explainable AI Lou, Xiayin; Luo, Peng; Li, Ziqi; Gao, Song; Meng, Liqiu Understanding and explaining complex geographic phenomena—ranging from climate change to socioeconomic disparities—is a central focus in both geography and the broader scientific community. Various methods have been developed to elucidate relationships between variables, from coefficient estimates in linear regression models to the increasingly dominant use of feature attribution scores in Explainable AI (XAI) techniques. However, explanations generated by XAI methods often carry uncertainty, stemming from the model itself and the data used to train the model. Despite the critical importance of accounting for such uncertainty, this issue remains largely overlooked in the geospatial domain. In this study, we developed an uncertainty quantification framework for XAI explanations based on conformal prediction, termed Geospatial eXplanation Conformal Prediction (GeoXCP). By incorporating spatial dependence into the modeling process, GeoXCP produced spatially adaptive explanations with calibrated uncertainty estimates. We validated the effectiveness of GeoXCP through extensive simulation experiments and real-world datasets. The results demonstrated that GeoXCP provided reliable explanations while effectively quantifying uncertainty across diverse geospatial scenarios. Our approach represented a significant advancement in explainable geospatial machine learning, enabling decision-makers to better assess the trustworthiness of model-driven insights. The proposed framework was implemented in a python package, named GeoXCP.

Open-source device for high sensitivity magnetic particle spectroscopy, relaxometry, and hysteresis loop tracing

2026-02-19T04:36:04Z

Open-source device for high sensitivity magnetic particle spectroscopy, relaxometry, and hysteresis loop tracing Mattingly, E.; Barksdale, A. C.; Śliwiak, M.; Chacon-Caldera, J.; Mason, E. E.; Wald, L. L. Magnetic nanoparticles (MNPs) are used extensively across numerous disciples, with applications including Magnetic Particle Imaging (MPI), targeted hyperthermia, deep brain stimulation, immunoassays, and thermometry. The assessment of MNPs, especially those being designed for MPI, is performed with magnetic particle spectrometers, relaxometers, loop tracers, or similar devices. Despite the many applications and the need for particle assessment, there are few consolidated resources for designing or building such a MNP assessment system. Here, we describe the design and performance of an open-source device capable of spectroscopy, relaxometry, and loop tracing. We show example measurements from the device and quantify the detection sensitivity by measuring a dilution series of Synomag-D 70 nm (from 0.5 mg Fe/ml to 7 ng Fe/ml) with a 10 mT drive field at 23.8 kHz. The device measures 260 pg Fe with SNR = 1 and 1.3 ng at SNR = 5 in spectroscopy mode in under one second of measurement time. The system has a dynamic range of 60 μg to 260 pg Fe without changing the hardware configuration. As an example application, we characterize Synomag-D’s relaxation time constant for drive fields 2–18 mT and compare the magnetization responses of two commonly used MNPs.

Precise Fermi level engineering in a topological Weyl semimetal via fast ion implantation

2026-02-19T04:36:09Z

Precise Fermi level engineering in a topological Weyl semimetal via fast ion implantation Mandal, Manasi; Chotrattanapituk, Abhijatmedhi; Woller, Kevin; Wu, Lijun; Xu, Haowei; Hung, Nguyen Tuan; Mao, Nannan; Okabe, Ryotaro; Boonkird, Artittaya; Nguyen, Thanh; Drucker, Nathan C; Chen, Xiaoqian M; Momiki, Takashi; Li, Ju; Kong, Jing; Zhu, Yimei; Li, Mingda The precise controllability of the Fermi level is a critical aspect of quantum materials. For topological Weyl semimetals, there is a pressing need to fine-tune the Fermi level to the Weyl nodes and unlock exotic electronic and optoelectronic effects associated with the divergent Berry curvature. However, in contrast to two-dimensional materials, where the Fermi level can be controlled through various techniques, the situation for bulk crystals beyond laborious chemical doping poses significant challenges. Here, we report the milli-electron-volt (meV) level ultra-fine-tuning of the Fermi level of bulk topological Weyl semimetal tantalum phosphide using accelerator-based high-energy hydrogen implantation and theory-driven planning. By calculating the desired carrier density and controlling the accelerator profiles, the Fermi level can be experimentally fine-tuned from 5 meV below, to 3.8 meV below, to 3.2 meV above the Weyl nodes. High-resolution transmission electron microscopy reveals the crystalline structure is largely maintained under irradiation, while electrical transport indicates that Weyl nodes are preserved and carrier mobility is also largely retained. Our work demonstrates the viability of this generic approach to tune the Fermi level in semimetal systems and could serve to achieve property fine-tuning for other bulk quantum materials with ultrahigh precision.

A facility for cryogenic ion irradiation and in situ characterization of rare-earth barium copper oxide superconducting tapes

2026-02-19T04:36:11Z

A facility for cryogenic ion irradiation and in situ characterization of rare-earth barium copper oxide superconducting tapes Devitre, AR; Fischer, DX; Woller, KB; Clark, BC; Short, MP; Whyte, DG; Hartwig, ZS Superconducting magnets based on Rare Earth Barium Copper Oxides (REBCO) offer transformative capabilities in the fields of fusion energy, high energy physics, and space exploration. A challenge shared by these applications is the limited lifetime of REBCO due to radiation damage sustained during operation. Here we present a new ion-beam facility that enables simultaneous cryogenic irradiation and in situ characterization of commercial REBCO tapes. The ion source provides spatially uniform fluxes up to 1018 protons/m2s with kinetic energies up to 3.4 MeV, in addition to helium and higher-Z species. Using this facility, we can induce uniform damage profiles in the first 10–20 µm of REBCO tapes with less than 0.25 appm of hydrogen implanted in REBCO after a dose of 1020 protons/m2. The tape can be held between 20 and 300 K with an accuracy of ±0.1 K and is connected to a four-point probe measuring the critical current, Ic, and critical temperature, Tc, before, during, and after irradiation with transport current ranging from 100 nA to 100 A, and a typical voltage noise less than 0.1 μV. These capabilities are presently used to study the effect of irradiation temperature on REBCO performance change during and after proton bombardment, to assess the possibility of Ic and Tc recovery after irradiation through thermal annealing, and to explore the instantaneous and recoverable suppression of Ic and Tc observed during irradiation.

High temperature stability of regrown and alloyed Ohmic contacts to AlGaN/GaN heterostructure up to 500 °C

2026-02-19T04:36:06Z

High temperature stability of regrown and alloyed Ohmic contacts to AlGaN/GaN heterostructure up to 500 °C Niroula, John; Xie, Qingyun; Rajput, Nitul S; Darmawi-Iskandar, Patrick K; Rahman, Sheikh Ifatur; Luo, Shisong; Palash, Rafid Hassan; Sikder, Bejoy; Yuan, Mengyang; Yadav, Pradyot; Micale, Gillian K; Chowdhury, Nadim; Zhao, Yuji; Rajan, Siddharth; Palacios, Tomás This Letter reports the stability of regrown and alloyed Ohmic contacts to AlGaN/GaN-on-Si high electron mobility transistors (HEMTs) for high temperature applications up to 500 °C. Transfer length method (TLM) measurements from 25 to 500 °C in air show that the regrown contacts appear to be stable up to 500 °C during short term (approximately 1 h) testing, while alloyed contacts appear to decrease in contact resistance from 300 to 500 °C though increases in the error bounds due to increase sheet resistance make it difficult to conclude definitely. Additionally, longer term testing shows both technologies remain stable at least up to 48 h at 500 °C, after which the large increase in sheet resistance makes the measurement uncertainty too large to conclude definitively. Advanced microscopy images indicate both the regrown and alloyed contact regions remain structurally intact after prolonged high temperature exposure with no visible degradation in crystallinity or metal composition.

Optical-pump–terahertz-probe spectroscopy in high magnetic fields with kHz single-shot detection

2026-02-19T04:36:14Z

Optical-pump–terahertz-probe spectroscopy in high magnetic fields with kHz single-shot detection Dastrup, Blake S; Miedaner, Peter R; Zhang, Zhuquan; Nelson, Keith A We demonstrate optical pump–THz probe (OPTP) spectroscopy with a variable external magnetic field (0–9 T), in which the time-dependent THz signal is measured by echelon-based single-shot detection at a repetition rate of 1 kHz. The method reduces data acquisition times by more than an order of magnitude compared to conventional electro-optic sampling using a scanning delay stage. The approach illustrates the wide applicability of the single-shot measurement approach to non-equilibrium systems that are studied through OPTP spectroscopy, especially in cases where parameters such as magnetic field strength (B) or other experimental parameters are varied. We demonstrate the capabilities of our measurement by performing cyclotron resonance experiments in bulk silicon, where we observe B-field-dependent carrier relaxation and distinct relaxation rates for different carrier types. We use a pair of economical linear array detectors to measure 500 time points on each shot, offering an equivalent performance to camera-based detection with possibilities for higher repetition rates.

Validation of the OpenMC Code for Fusion Applications: The FNG-Streaming Benchmark Case

2026-02-19T04:36:08Z

Validation of the OpenMC Code for Fusion Applications: The FNG-Streaming Benchmark Case Segantin, Stefano; Ebiwonjumi, Bamidele; Peterson, Ethan In this work, we benchmark OpenMC against the FNG-ITER streaming experiment. FNG-ITER streaming, a high-quality experiment carried out at the ENEA laboratories in Frascati, Italy, was initially included in SINBAD (Shielding Integral Benchmark Archive and Database). More recently, the benchmark was included in the Compilation of Nuclear Data Experiments for Radiation Characterization as well. It consists of a neutron shielding experiment with a rather complex geometry that constitutes an appropriate validation study for the use of weight windows within OpenMC. Measurements include flux detection via four different types of activation foils divided into three batches and a set of thermoluminescent detectors for nuclear heating. The OpenMC results are in very good agreement with those of MCNP and the experimental measurements, with the majority of the discrepancies within the combined statistical error and experimental uncertainty (less than 10% computed measured discrepancy).

Use of Bayesian decision analysis to maximize value in patient-centered randomized clinical trials in Parkinson’s disease

2026-02-19T04:36:13Z

Use of Bayesian decision analysis to maximize value in patient-centered randomized clinical trials in Parkinson’s disease Chaudhuri, Shomesh E; Ben Chaouch, Zied; Hauber, Brett; Mange, Brennan; Zhou, Mo; Christopher, Stephanie; Bardot, Dawn; Sheehan, Margaret; Donnelly, Anne; McLaughlin, Lauren; Caldwell, Brittany; Benz, Heather L; Ho, Martin; Saha, Anindita; Gwinn, Katrina; Sheldon, Murray; Lo, Andrew W A fixed one-sided significance level of 5% is commonly used to interpret the statistical significance of randomized clinical trial (RCT) outcomes. While it is necessary to reduce the false positive rate, the threshold used could be chosen quantitatively and transparently to specifically reflect patient preferences regarding benefit–risk tradeoffs as well as other considerations. How can patient preferences be explicitly incorporated into RCTs in Parkinson’s disease (PD), and what is the impact on statistical thresholds for device approval? In this analysis, we apply Bayesian decision analysis (BDA) to PD patient preference scores elicited from survey data. BDA allows us to choose a sample size (𝑛) and significance level (𝛼) that maximizes the overall expected value to patients of a balanced two-arm fixed-sample RCT, where the expected value is computed under both null and alternative hypotheses. For PD patients who had previously received deep brain stimulation (DBS) treatment, the BDA-optimal significance levels fell between 4.0% and 10.0%, similar to or greater than the traditional value of 5%. Conversely, for patients who had never received DBS, the optimal significance level ranged from 0.2% to 4.4%. In both of these populations, the optimal significance level increased with the severity of the patients’ cognitive and motor function symptoms. By explicitly incorporating patient preferences into clinical trial designs and the regulatory decision-making process, BDA provides a quantitative and transparent approach to combine clinical and statistical significance. For PD patients who have never received DBS treatment, a 5% significance threshold may not be conservative enough to reflect their risk-aversion level. However, this study shows that patients who previously received DBS treatment present a higher tolerance to accept therapeutic risks in exchange for improved efficacy which is reflected in a higher statistical threshold.

Choice denied: impact of income and credit-based tenant screening on the Housing Choice Voucher program

2026-02-19T04:35:56Z

Choice denied: impact of income and credit-based tenant screening on the Housing Choice Voucher program So, Wonyoung; Gade, Anisha; Hangen, Forrest The Housing Choice Voucher program supports over 2.5 million households by subsidizing rent payments within the private housing market. However, challenges arise due to exclusionary practices, undermining the program’s goal of ‘choice.’ Tenant screening practices have been critical in exacerbating these challenges, yet their impact remains understudied. Drawing on tenant screening criteria documents from property management websites and the Survey of Consumer Finances, this study finds that while voucher holders generally meet rent-to-income thresholds due to the subsidies—keeping their rent burden relative to their income, they still face barriers related to credit scores, bankruptcy history, and debt. These criteria, which apply to both voucher and non-voucher renters, may exclude approximately one in ten voucher holders, despite the guaranteed portion of rent covered by public assistance. These findings show an urgent need for policy interventions to the potential exclusionary impacts of tenant screening services.

AgentNexus: Accelerating AI Agent Development and Enhancing Interoperability with MCP

2026-02-18T03:09:03Z

AgentNexus: Accelerating AI Agent Development and Enhancing Interoperability with MCP Yae, Jung; Hamilton, Lei The DoD faces significant challenges in its pursuit of AI superiority, as disparate data and development platforms create redundant efforts and limit interoperability. Additionally, existing DoD systems are ill-equipped to handle the recent paradigm shift toward agentic AI, which requires modern standards and tools. To address these gaps, this paper introduces AgentNexus, an application designed to streamline the development, deployment, and servicing of AI agents. AgentNexus provides an application featuring an advanced agents processing backend, a scalable service layer, and an intuitive user interface. It provides pre-built toolkits, sophisticated RAG pipeline, and MCP for enhanced interoperability. The successful development of an Education Assistant agent validates the application’s capacity to support the rapid implementation of multi-agent workflows. By fostering a collaborative and standardized environment, AgentNexus mitigates critical barriers of interoperability and duplicated effort, accelerating the delivery of multi-agent AI to warfighters.

Intelligent C-17 Load Planning for Flight Optimization

2026-02-18T03:08:58Z

Intelligent C-17 Load Planning for Flight Optimization McAlister, Catherine; Jones, Mathew; McConville, Sean C-17 Globemaster III cargo capacity is significantly underutilized, with many sorties transporting only a few pallets despite the aircraft’s 170,900-pound payload capability. Historical flight data analysis reveals inefficient scheduling practices that increase operational costs, crew workload, and overall negatively effect mission capability. This paper details the development of an AI-powered optimization model to improve C-17 cargo utilization and reduce required flight operations. We analyzed historical C-17 transportation data and created both traditional optimization algorithms and predictive AI models to determine optimal flight scheduling for 3-week operational periods. The AI model achieved 97.9% accuracy in predicting optimal flight count requirements and 89.3% accuracy in predicting optimal flight assignment for specific cargo, representing a 23% reduction in total flights and a 15% increase in average cargo utilization. These results demonstrate that data-driven flight scheduling can significantly improve C-17 operational efficiency, reduce costs across the airlift community, and enabling additional time towards advanced training, contingency support, and critical warfighter operations, ultimately increasing the lethality and readiness of the Department of Defense.

Securing Intelligence: The Strategic Necessity of Air-Gapped AI Systems in the Age of Cloud-Based LLMs

2026-02-18T03:09:05Z

Securing Intelligence: The Strategic Necessity of Air-Gapped AI Systems in the Age of Cloud-Based LLMs Viggh, Herbert; Tsagaratos, Jennifer The increasing use of large language models (LLMs) in applications, from military strategy to customer service, raises concerns about data sovereignty, security, and privacy. Cloudbased API models, created by companies such as OpenAI, pose significant risks due to training data exposure and prompt injection attacks, which can compromise sensitive information and hidden biases that could influence reporting or executive decision-making processes. Real-world incidents, such as the leakage of Samsung’s proprietary source code through ChatGPT, highlight the dangers of relying on cloud providers with complete visibility into client queries. Furthermore, data localization laws and regulations, such as the General Data Protection Regulation (GDPR), underscore the risks associated with outsourcing intelligence and decision support systems to foreign entities. Airgapped AI solutions, which run on isolated networks disconnected from the outside world, offer a secure alternative for sensitive environments such as national defense, research laboratories, and critical infrastructure. By maintaining control over AI processes, organizations can ensure information safety, comply with regulations, and mitigate risks associated with cloud-based AI infrastructure, ultimately safeguarding their data integrity, privacy, and operational independence.

RAIMOND Requirements AI for Military Operational Needs Development

2026-02-18T03:09:07Z

RAIMOND Requirements AI for Military Operational Needs Development Garcia, Fabio; Steilberg, Jackson The Joint Capabilities Integration and Development System (JCIDS) was created as a means to overhaul military procurement processes. Ideally, the requirements development process is meant to take a total of 2-4 years from concept to manufacturing. However the actual length of concept development is much longer. As a result, technologies that are conceptualized through the analytical process often enter the acquisition too late to need for the warrior. To reduce the lengthy timeline in requirements development, we used Large Language Models (LLMs) to conduct the necessary research and synthesize documents that abide by strict JCIDS guidelines. Prompt engineering can achieve these results as a proof of concept. However, the output responses lack the content length and depth necessary to pass through the requirements validation process. Therefore, a combination of agentic workflows, prompt engineering, and sufficient context is needed to achieve the desired outcomes. This project utilizes a novel framework to derive Capabilities Based Assessments (CBAs) at an approximate 80 percent readiness level requiring the final steps of validation and verification by subject matter experts.

Machine Learning for the Enhancement of Adaptive Optics

2026-02-18T03:09:05Z

Machine Learning for the Enhancement of Adaptive Optics Hall, Robert; Chen, Justin Optical systems (telescopes, lasers, microscopes, etc.) have degraded performance over long distances due to scintillation caused by Earth’s atmosphere, where adaptive optics (AO) is often used to enhance its signal-to-noise (SNR) ratio or image quality. Astronomers have found success in laser-based adaptive optics where they survey the atmosphere with a laser and subtract its effects on the resultant image. Although effective in most cases, these systems can be extremely costly, are computationally intensive in real time, and fall short in some edge cases. We propose an autoencoder/ decoder and a generalized sequence to sequence model (LSTM) as a cost-effective method to off-load computational complexity from real time and enhance performance in edge cases. This study utilizes four simulated datasets of wavefront sensor frames for a variety of atmospheric conditions, done in collaboration with MIT Lincoln Laboratory [1]–found auto-encoding performance just shy of traditional methodology and found LSTM performance that predicts well the general shape on the WFS, but suffers from scaling issues.

From Hype to Reality: Real-World Lessons and Recommendations for AI in Military Applications

2026-02-18T03:09:06Z

From Hype to Reality: Real-World Lessons and Recommendations for AI in Military Applications Lynch, Joshua; Niss, Laura The current use cases, limitations, and future capacity of large language models (LLMs) as assistants to military personnel remain an open question. This paper presents a case study of an Airman’s interaction with and trust calibration of LLMs over three months, both as an everyday assistant and for development of ROMAD-AI, a tactical military application. Through intuitive, AI-generated software development, an approach relying on iterative code generation through natural language prompting of LLMs from a technical novice rather than human generated programming from a technical expert, the research reveals significant gaps between industry curated AI capability demonstrations and operational reality, requiring systematic trust calibration and realistic scope management. Outcomes are analyzed through operational and technical expertise perspectives to provide practical guidance for both military service members seeking effective AI integration and researchers developing military-focused AI systems.

Large Language Models and Defense Strategy: Escalation Risks and National Security Challenges

2026-02-18T03:09:04Z

Large Language Models and Defense Strategy: Escalation Risks and National Security Challenges Hou, Jonathan; Lax, Edwin This literature review examines the strategic vulnerabilities posed by Large Language Models (LLMs) in military and national security contexts. It synthesizes recent research on their propensity for escalatory reasoning, cultural misalignment, semantic manipulation, and dual-use ambiguity. Findings from conflict s imulations a nd c oalition p lanning m odels reveal how LLMs may default to aggressive or biased outputs under ambiguity. These tendencies threaten alliance cohesion, distort decision-making, and undermine trust in AI-enabled operations. The review concludes by advocating for safeguards such as culturally calibrated training, rigorous output verification, a nd the integration of human-AI intermediaries to prevent destabilizing outcomes.

Synchronization-Aware Diffusion Models for Intra-Family RF Signal Classification

2026-02-18T03:09:02Z

Synchronization-Aware Diffusion Models for Intra-Family RF Signal Classification Hayden, Hunter; Botero, Joey Classification of radio frequency (RF) signals in the presence of channel-induced synchronization errors remains a critical challenge in spectrum awareness systems. Traditional classification pipelines generally rely on fixed synchronization algorithms or assume aligned signals, which limits robustness under real world timing, phase, and frequency distortions. We introduce SyncDiff, a novel encoder-only diffusion model architecture that predicts synchronization parameters through iterative denoising steps prior to classification. By replacing conventional synchronization algorithms with a learned datadriven correction mechanism, our approach enables adaptive signal alignment based on current channel distortions in unsynchronized input data. SyncDiff employs a UNet based encoder to refine synchronization parameters across multiple inference steps, dynamically reducing channel-induced alignment errors while preserving the inherit modulation specific characteristics that allow these signals to be discriminable. Evaluations of the RadioML2018 RF standard benchmark data set [1] demonstrates improved classification accuracy across varying SNRs, modulation schemes and synchronization impairments. Our findings highlight the potential of diffusion-based synchronization learning to improve downstream RF classification without reliance on expert-engineered synchronization routines.

Three-Dimensional Full-Core BEAVRS Using OpenMOC with Transport Equivalence

2026-02-18T03:07:45Z

Three-Dimensional Full-Core BEAVRS Using OpenMOC with Transport Equivalence Giudicelli, G; Forget, B; Smith, K Using an optimized implementation of the three-dimensional (3D) method of characteristics for neutron transport, along with a novel equivalence method for transport calculations that was designed to correct self-shielding errors from neglecting the angular dependence of resonant group absorption, a 3D full-core light water reactor hybrid stochastic-deterministic eigenvalue calculation was achieved. This paper presents the optimizations developed and compares the transport solutions obtained. For the statepoint, run times near 10 000 CPU hours are achieved—improving on previous works by an order of magnitude—with near 1% error on pin fission to 238U capture ratios and a few dozen pcms on the eigenvalue.

Considering a US-Supported Self-Defense Option for Taiwan

2026-02-18T03:07:50Z

Considering a US-Supported Self-Defense Option for Taiwan Glaser, Charles L. There is wide agreement that Taiwan is the most dangerous issue dividing the United States and China. China believes Taiwan is part of its homeland, views unification with Taiwan as a core interest, and is determined to gain full control of the island. China continues to prefer peaceful unification, but explicitly retains the option of using military forces to achieve unification and seeks to use the threat of military force to strengthen its negotiating hand. Current US policy includes an ambiguous commitment to defend Taiwan if attacked or severely coerced by China—it leaves open whether and how the United States would respond. In addition, the United States provides Taiwan with weapons to improve its ability to defend itself. The United States is pressing Taiwan to deploy smaller mobile weapons that would increase the survivability and lethality of its forces; these forces would support a “porcupine strategy” that makes Taiwan harder to invade and conquer and would, at a minimum, provide time for US forces to arrive to aid Taiwan’s defense.

Needs, Wants . . . and Excuses: What Executives Can Learn from Zig Ziglar About Working with Universities

2026-02-18T03:07:49Z

Needs, Wants . . . and Excuses: What Executives Can Learn from Zig Ziglar About Working with Universities Wright, Randall S. Zig Ziglar was a famous sales trainer, motivational speaker, and author on salesmanship. When he died on November 28, 2012, Kevin Kruse (Citation2024)—best-selling author of Emotional Intelligence: 52 Strategies, coach to Fortune 500 CEOs, Marine Corps generals, and Silicon Valley entrepreneurs—wrote this in Forbes: “Zig Ziglar died today at age 86. A World War II veteran, Zig Ziglar became the top salesperson in several organizations before striking out on his own as a motivational speaker and trainer. With a Southern charm and lessons grounded in Christianity, Ziglar wrote over two dozen books and amassed a following of millions who were encouraged by his lessons for success.”

Computational studies of electric field effects in CO2 methanation on Ni metal surfaces

2026-02-18T03:07:48Z

Computational studies of electric field effects in CO2 methanation on Ni metal surfaces Wakamatsu, Katsuhiro; Yasuda, Takaaki; Aratani, Masato; Ogura, Teppei Non-Faradaic electrochemical modification of catalytic activity (NEMCA) with an electric field (EF) has attracted attention as one of the methods to improve catalyst performance. However, this activation mechanism is not still clear. In this study, we focused on the NEMCA mechanism in CO2 methanation on Ni metal catalyst with solid oxide electrolysis cell (SOEC) and calculated two possible effects of the NEMCA mechanism; direct EF applications and oxygen atom co-adsorptions, using the density functional theory calculations and detailed kinetic simulations. Compared with these effects in terms of kinetic energy changes in the rate-determining steps, it has been revealed that the spillover effect of lattice oxygen toward the catalyst surface is dominant in the NEMCA mechanism. Also, we have found that overall CO2 methanation is promoted in SOEC mode with oxygen atom co-adsorptions in both cases of Ni flat and step sites.

Equilibrium configurations of line arrays with respect to the deviatoric mean drift forces

2026-02-18T03:07:47Z

Equilibrium configurations of line arrays with respect to the deviatoric mean drift forces Tokić, Grgur; Yue, Dick KP Monochromatic waves incident on an array of structures give rise to nonlinear, time-constant mean drift forces (MDFs). These forces depend on the array's spatial configuration; their magnitude and the direction is, in general, different for every structure in the array. If the spatial configuration of an array is not fixed, as is the case in arrays of individually anchor-moored structures, the time-constant differences in MDF on individual bodies can lead to a change in spatial configuration, which could, in turn, significantly affect both the first-order, time-harmonic response of the array, as well as the downwave component of the MDF. Here, we explore the dependency of these deviatoric forces on array configurations and on the frequency of the incident monochromatic waves. We consider configurations of line arrays (consisting of 2–5 vertical circular cylinders) that are described by 1 or 2 parameters, and we focus on the along-array component of deviatoric forces. Using multiple scattering computational simulations, we identify the array configurations in which the deviatoric drift forces are zero, and we discuss the stability of these equilibrium configurations with respect to class-preserving configuration perturbations. Both stable and unstable equilibria exist, but the relative number of unstable equilibria grows as the number of degrees of freedom of the configuration perturbations increases. Interestingly, the stable configurations experience a generally lower downwave mean drift force on the entire array than the unstable ones. Overall, the variations in the deviatoric and the downwave MDFs between equilibria are significant (on the order of the isolated body MDF).

1000-MW CSP with 100-gigawatt-hour crushed-rock heat storage to replace dispatchable fossil-fuel electricity

2026-02-18T03:07:41Z

1000-MW CSP with 100-gigawatt-hour crushed-rock heat storage to replace dispatchable fossil-fuel electricity Forsberg, Charles We are developing 100-GWh heat-storage systems for use with 1000-MW Concentrated Solar Power (CSP) and nuclear reactor systems with capital cost goals of several dollars per kWh of heat storage—a factor of 50 under lithium ion batteries per unit of electricity. The capabilities of a 100-GWh heat storage system are similar to the Tennessee Valley Authority Raccoon Mountain pumped hydro facility that can provide 1652 MW(e) for 22 hours to address daily to weekly storage. The low capital cost of the Crushed Rock Ultra-large Stored Heat (CRUSH) system is only possible in large-capacity systems; thus, the CSP system average 24/7 heat inputs may exceed 1000 MW to match the heat storage capacity. Hot oil or nitrate salt is pumped from multiple solar farms or towers to the central CRUSH system and associated power block. The peak power block output may be 2 to 4 times average output with large economics of scale relative to the smaller power blocks associated with existing CSP systems. The cost savings from the large storage and the power block exceed the cost of hot oil or hot nitrate salt insulated pipelines over 10+ kilometers. The heat is stored in crushed rock in piles 20 m high and up to 250 m by 250 m on a side within an insulated floor and building structure. The sides of the rock pile are sloped rock that allow rock expansion and contraction with temperature without generating mechanical forces against walls. Heat is transferred from CSP to the crushed rock and then to the power cycle using (1) heat transfer oils for lower-temperature power systems to 400°C or (2) nitrate salts for higher-temperature power systems to 600°C. In charging mode, hot heat transfer fluid is sprayed over crushed rock and drains through the rock to the collection pans at the bottom to be reheated. Sections of rock are heated sequentially. In discharge mode cold heat transfer fluid is sprayed over crushed rock and drains through the rock to the collection pans below to deliver hot fluid to the power cycle. Heat storage costs are minimized by three features. Crushed rock is the lowest-cost storage material. The large building size minimizes the surface-to-volume ratio and thus building, insulation and foundation costs. The inventory and thus cost of oil and nitrate salt is minimized by using these fluids to transfer heat from CSP collectors to storage and then to the power block—but not for heat storage.

The way forward: The path to monolithic additive manufacture of lower hybrid current drive launchers

2026-02-18T03:07:38Z

The way forward: The path to monolithic additive manufacture of lower hybrid current drive launchers Seltzman, AH; Wukitch, SJ Additive Manufacturing (AM) is a key enabling technology for the rapid production of complex radio-frequency (RF) structures used in lower hybrid current drive (LHCD) launchers. Glenn Research Copper 84 (GRCop-84), a Niobium Chromide (Cr2Nb) 8 at. % Cr, 4 at. % Nb precipitation hardened alloy, is suitable for AM with Laser Powder Bed Fusion (L-PBF), achieving 99.5% density, Ra=3-4 µm surface roughness, yield strength of 470 MPa and an ultimate tensile strength (UTS) of 710 MPa in as-printed condition. AM of a high field side (HFS) lower LHCD launcher from GRCop-84 alloy demonstrated several critical advancements in AM of RF launchers. Waveguides with a pentagonal cross-section were designed to support the top internal waveguide surface with 45-degree chamfers from the sidewall, eliminating collapse of the ceiling, while maintaining RF properties near identical to a rectangular cross section. Hot Isostatic Pressing (HIPing) consolidated residual voids within the material, increasing density from 99.5% to 100%. Chemical-Mechanical Polishing (CMP) reduced residual surface roughness from the L-PBF process to Ra=0.1 µm / Rq=0.4 µm to lower RF losses. Advancements in L-PBF for the AM of copper alloys have increased the maximum build volume from 250x250x300mm on the Concept Laser M2 printer to 400x400x400mm on the EOS M400 printer. This increased build volume now enables monolithic AM of complete LHCD launchers with integrated cooling channels that eliminate the time-consuming laser welding assembly of launcher segments previously required by the smaller build volume.

Particle-in-cell simulations of parasitic electrostatic wave excitation in the ion cyclotron range of frequencies and high harmonic fast wave regimes

2026-02-18T03:07:31Z

Particle-in-cell simulations of parasitic electrostatic wave excitation in the ion cyclotron range of frequencies and high harmonic fast wave regimes Diab, Raymond; Baek, Seung-Gyou; Bonoli, Paul; Jenkins, Thomas G; Ono, Masayuki; Smithe, David Using the open-source code SMILEI [J. Derouillat et al., Comput. Phys. Commun. 222, 351-373 (2018)], we perform one-dimensional full-f particle-in-cell (PIC) simulations of parasitic electrostatic wave excitation in the Ion Cyclotron Range of Frequencies (ICRF) and High Harmonic Fast Wave (HHFW) regimes in an inhomogeneous plasma. We first study direct coupling from the fast wave to electrostatic waves at the lower hybrid (LH) resonance (S=0). In the ICRF regime, we show that the fast wave can couple to the Ion Bernstein Wave (IBW), which propagates beyond the LH resonance layer. On the other hand, in the HHFW regime, no direct coupling to the IBW is observed, but electrostatic waves, likely to be Hot Ion Plasma Waves (HIPW or HPW), are seen on the low-density side of the LH resonance layer. The coupling efficiency to electrostatic waves is seen to increase with ion temperature. Parametric decay instabilities (PDIs) are then investigated in both regimes. In the ICRF regime, both resonant and non-resonant decay channels are observed and compared with theory. In the HHFW regime, we observe multiple sidebands separated by the ion cyclotron frequency, as measured experimentally on NSTX [J. R. Wilson et al., AIP Conf. Proc. 787, 66 (2005)]. The nature of these waves is discussed. Perpendicular ion heating is also found in the region where PDIs occur, consistent with experimental observations.

Towards fast, accurate predictions of RF simulations via data-driven modeling: Forward and lateral models

2026-02-18T03:07:43Z

Towards fast, accurate predictions of RF simulations via data-driven modeling: Forward and lateral models Wallace, GM; Bai, Z; Bertelli, N; Bethel, EW; Perciano, T; Shiraiwa, S; Wright, JC Three machine learning techniques (multilayer perceptron, random forest, and Gaussian process) provide fast surrogate models for lower hybrid current drive (LHCD) simulations. A single GENRAY/CQL3D simulation without radial diffusion of fast electrons requires several minutes of wall-clock time to complete, which is acceptable for many purposes, but too slow for integrated modeling and real-time control applications. More accurate simulations with fast electron diffusion are even slower, requiring multiple hours of run time with parallel processing. The machine learning models use a database of 16,000+ GEN-RAY/CQL3D simulations for training, validation, and testing. Latin hypercube sampling methods implemented in πScope ensure that the database covers the range of 9 input parameters (ne0, Te0, Ip, Bt, R0, n∥︀, Ze f f, Vloop, PLHCD) with sufficient density in all regions of parameter space. The surrogate models reduce the computation time from minutes-hours to ms with high accuracy across the input parameter space. Data-driven surrogate models also allow for solving inverse and “lateral” problems. A surrogate model for the inverse problem maps from a desired current drive or power deposition profile to a set of input parameters that would result in such a profile, while a surrogate model for the lateral problem maps from a measured experimental quantity such as hard x-ray emission to a current drive or power deposition profile. The πScope database creation workflow is flexible and applicable to other RF simulation codes such as TORIC.

Decrypting the mechanisms of wicking and evaporation heat transfer on micro-pillars during the pool boiling of water using high-resolution infrared thermometry

2026-02-18T03:07:40Z

Decrypting the mechanisms of wicking and evaporation heat transfer on micro-pillars during the pool boiling of water using high-resolution infrared thermometry Wang, Chi; Rahman, Md Mahamudur; Bucci, Matteo Surfaces with micrometer-scale pillars have shown great potential in delaying the boiling crisis and enhancing the critical heat flux (CHF). However, physical mechanisms enabling this enhancement remain unclear. This knowledge gap is due to a lack of diagnostics that allow elucidating how micro-pillars affect thermal transport phenomena on the engineered surface. In this study, for the first time, we are able to measure time-dependent temperature and heat flux distributions on a boiling surface with engineered micro-pillars using infrared thermometry. Using these data, we reveal the presence of an intra-pillar liquid layer, created by the nucleation of bubbles and partially refilled by capillary effects. However, contrarily to conventional wisdom, the energy removed by the evaporation of this liquid cannot explain the observed CHF enhancement. Yet, predicting its dry out is the key to delaying the boiling crisis. We achieve this goal using simple analytic models and demonstrate that this process is driven by conduction effects in the boiling substrates and, importantly, in the intra-pillar liquid layer itself. Importantly, these effects also control the wicking flow rate and its penetration length. The boiling crisis occurs when, by coalescing, the size of the intra-pillar liquid layer becomes too large for the wicking flow to reach its innermost region. Our study reveals and quantifies unidentified physical aspects, key to the performance optimization of boiling surfaces for cooling applications.

APP2 Status Summary: Proposed New VLBI Capabilities for Cycle 8

2026-03-05T18:44:02Z

APP2 Status Summary: Proposed New VLBI Capabilities for Cycle 8 Matthews, Lynn D.; Crew, Geoffrey B. This document contains material regarding the new capabilities proposed for ALMA Cycle 8. These included a passive phasing mode (for weaker sources), an ALMA Band 3 pulsar observing mode, a prototype spectral line mode, also for Band 3, and minor improvements to the code for operational reasons. In the event, the first two were accepted for a Cycle 8 that was delayed by the COVID-19 pandemic. The spectral line capability was deferred until Cycle 9. In response to reviewer questions, a modified figure was prepared and presented as an addendum. Both documents are presented here as a single PDF. This report was prepared for the formal acceptance of the software required for Cycle 8. Notionally it is ALMA Technical Note #21, but not (yet) published as such.

Final Report: Enabling New VLBI Science with the ALMA Phasing System - Phase 3 (APP3): An ALMA North America Development Project

2026-03-04T18:34:05Z

Final Report: Enabling New VLBI Science with the ALMA Phasing System - Phase 3 (APP3): An ALMA North America Development Project Matthews, Lynn D.; Crew, Geoffrey B. Executive Summary: This document provides a summary of activities undertaken as part of the ALMA North America Development Project “Enabling New VLBI Science with the ALMA Phasing System - Phase 3 (APP3)”, whose period of performance ex- tended from January 17, 2022 to July 16, 2024. The successful completion of this project has resulted in the introduction of flexible tuning for ALMA very long baseline interfer- ometry (VLBI) operations, a fully flexible spectral line VLBI observing mode, and the enabling of panchromatic VLBI, allowing ALMA in principle to operate as a phased array in any available receiver band. The project also carried out a series of activities aimed at maintaining and optimizing existing VLBI infrastructure and provided training to staff at the Joint ALMA Observatory (JAO) to enable a transition to autonomous VLBI observing. A video feature and accompanying news article were produced near the conclusion of this project to make the results accessible to a broader audience. This report was prepared as a final report on the activities undertaken under the NA Development program mentioned in the abstract.

Open Access Task Force - Implementation Team Progress Report

2026-02-14T03:08:05Z

Open Access Task Force - Implementation Team Progress Report Bebergal, Peter; Bourg, Chris; Dunn, Katharine H.; Nurnberger, Amy; Pierce, Marianna; Shirer, Karen; Weeramuni, Lindsey; Wilcoxson, Jaren D. This report outlines the progress to date of the Open Access Task Force Implementation Team (OATF-IT), first convened in December 2019 to prioritize, shepherd, and support the final recommendations (October 2019) of the MIT-wide Ad Hoc Task Force on Open Access to MIT’s Research (OATF). The OATF launched in 2017 to update and revise the Institute’s policies and practices around open publications, data, educational materials, and software.

Recommendations of the MIT Ad Hoc Task Force on Open Access to MIT's Research

2026-02-14T03:08:04Z

Recommendations of the MIT Ad Hoc Task Force on Open Access to MIT's Research Abelson, Harold; Bourg, Chris; Bebergal, Peter; Bond, Robert A.; Cheng, Herng Yi; Chuang, Isaac L.; Cummins, Christopher C; Fitzgerald, Deborah K; Jarzombek, Mark; Lindsay, Nick; Pollard, Tom Joseph; Reid, Jack; Shirer, Karen; Trout, Bernhardt L.; Vander Heiden, Matthew G.; von Hippel, Eric A; Wilcoxson, Jaren D.

Regulating Wait-Driven Requests in Queues

2026-03-08T03:22:53Z

Regulating Wait-Driven Requests in Queues Freund, Daniel; Hausman, David; Weng, Wentao The study of rational queueing has a long and distinguished history focused on individuals' preference to avoid waiting. Surprisingly, there are settings in which some potential arrivals (which we also refer to as requests) derive utility from waiting and disutility from service. Our primary example is the U.S. affirmative asylum process. In this context, applicants obtain a work permit while waiting for an asylum interview; hence, if the (expected) wait is long enough, then even an applicant who knows that their application will be denied and lead to deportation proceedings, may find it in their interest to apply and thus benefit from legally working during the wait. Similar dynamics could occur in other settings like content moderation in social networks. The common thread of these examples is the potentially self-exciting queue: when wait times are long, many arrivals are incentivized to join, and wait times become even longer. However, the system designer usually wants to avoid a large backlog. Indeed, the US Citizenship and Immigration Services (USCIS) mostly schedules asylum interviews in a Last-In-First-Out (LIFO) manner with the explicit goal of dissuading applicants with non-meritorious cases trying to exploit the long backlog. Despite this interesting scheduling choice in practice, and the potential prevalence of similar settings in other applications, the existing literature on rational queueing lacks frameworks to study the impact of wait-driven requests. Motivated by this gap in the literature, we formalize a dynamical system where in each round, a given scheduling policy and a realized request rate determine the wait time distribution in a fluid queueing system. Observing the expected benefit from waiting in one round, requests update their decisions, setting the request rate for the next round. Assuming a concave benefit function from waiting, alongside general conditions, we prove that, for minimizing the backlog, LIFO is most effective while First-In-First-Out (FIFO) is least effective among all work-conserving policies. Moreover, we show that the dynamical system exhibits metastability: for either FIFO or LIFO, the system converges to either a zero-wait or a congested equilibrium. Although some asylum practitioners support the use of LIFO, critics often admonish the real-world use of LIFO for its failure to maintain FIFO's order fairness: earlier requests should get earlier service. Our results demonstrate this trade-off between LIFO and FIFO. But we also show limitations of hybrid policies, which probabilistically follow either LIFO or FIFO, in navigating the trade-off between LIFO's efficiency and FIFO's fairness. Our work formalizes the concept of order fairness in queueing systems with abandonment and demonstrates that hybrid policies can be Pareto-dominated by LIFO: they may have both longer backlog and worse order fairness. Finally, we use real-world data on the scheduling of affirmative asylum applications to evaluate the change in fairness over the past 20 years under different policies. EC ’25, July 7–10, 2025, Stanford, CA, USA

DeepSeek Inside: Origins, Technology, and Impact

2026-03-08T03:23:14Z

DeepSeek Inside: Origins, Technology, and Impact Cusumano, Michael The release of DeepSeek V3 and R1 in January 2025 caused steep declines in the stock prices of companies that provide generative artificial intelligence (GenAI) infrastructure technology and datacenter services. These two large language models (LLMs) came from a little-known Chinese startup with approximately 200 employees compared to at least 3,500 employees for industry-leader OpenAI. DeepSeek seemed to have developed this powerful technology much more cheaply than previously thought possible. If true, DeepSeek had the potential to disrupt the economics of the entire GenAI ecosystem and the dominance of U.S. companies ranging from OpenAI to Nvidia.

Density-Dependent Graph Orientation and Coloring in Scalable MPC

2026-03-08T03:23:16Z

Density-Dependent Graph Orientation and Coloring in Scalable MPC Ghaffari, Mohsen; Grunau, Christoph This paper presents massively parallel computation (MPC) algorithms in the strongly sublinear memory regime (aka, scalable MPC) for orienting and coloring graphs as a function of its subgraph density. Our algorithms run in poly(log log n) rounds and compute an orientation of the edges with maximum outdegree O (α log log n) as well as a coloring of the vertices with O (α log log n) colors. Here, α denotes the density of the densest subgraph. Our algorithm's round complexity is notable because it breaks the [EQUATION] barrier, which applied to the previously best known density-dependent orientation algorithm [Ghaffari, Lattanzi, and Mitrovic ICML'19] and is common to many other scalable MPC algorithms. PODC ’25, Huatulco, Mexico

Privacy-Preserving Mechanisms for Coordinating Airspace Usage in Advanced Air Mobility

2026-03-08T03:22:45Z

Privacy-Preserving Mechanisms for Coordinating Airspace Usage in Advanced Air Mobility Maheshwari, Chinmay; Mendoza, Maria; Tuck, Victoria; Su, Pan-Yang; Qin, Victor; Seshia, Sanjit; Balakrishnan, Hamsa; Sastry, Shankar Advanced Air Mobility (AAM) operations are expected to transform air transportation while challenging current air traffic management practices. By introducing a novel market-based mechanism, we address the problem of on-demand allocation of capacity-constrained airspace to AAM vehicles with heterogeneous and private valuations. We model airspace and air infrastructure as a collection of contiguous regions (or sectors) with constraints on the number of vehicles that simultaneously enter, stay, or exit each region. Vehicles request access to airspace with trajectories spanning multiple regions at different times. We use the graph structure of our airspace model to formulate the allocation problem as a path allocation problem on a time-extended graph. To ensure that the cost information of AAM vehicles remains private, we introduce a novel mechanism that allocates each vehicle a budget of "air-credits" (an artificial currency) and anonymously charges prices for traversing the edges of the time-extended graph. We seek to compute a competitive equilibrium that ensures that: (i) capacity constraints are satisfied, (ii) a strictly positive resource price implies that the sector capacity is fully utilized, and (iii) the allocation is integral and optimal for each AAM vehicle given current prices, without requiring access to individual vehicle utilities. However, a competitive equilibrium with integral allocations may not always exist. We provide sufficient conditions for the existence and computation of a fractional-competitive equilibrium, where allocations can be fractional. Building on these theoretical insights, we propose a distributed, iterative, two-step algorithm that: 1) computes a fractional competitive equilibrium, and 2) derives an integral allocation from this equilibrium. We validate the effectiveness of our approach in allocating trajectories for the emerging urban air mobility service of drone delivery.

Meschers: Geometry Processing of Impossible Objects

2026-03-08T03:23:14Z

Meschers: Geometry Processing of Impossible Objects Dodik, Ana; Yu, Isabella; Chandra, Kartik; Ragan-Kelley, Jonathan; Tenenbaum, Joshua; Sitzmann, Vincent; Solomon, Justin Impossible objects, geometric constructions that humans can perceive but that cannot exist in real life, have been a topic of intrigue in visual arts, perception, and graphics, yet no satisfying computer representation of such objects exists. Previous work embeds impossible objects in 3D, cutting them or twisting/bending them in the depth axis. Cutting an impossible object changes its local geometry at the cut, which can hamper downstream graphics applications, such as smoothing, while bending makes it difficult to relight the object. Both of these can invalidate geometry operations, such as distance computation. As an alternative, we introduce Meschers, meshes capable of representing impossible constructions akin to those found in M.C. Escher's woodcuts. Our representation has a theoretical foundation in discrete exterior calculus and supports the use-cases above, as we demonstrate in a number of example applications. Moreover, because we can do discrete geometry processing on our representation, we can inverse-render impossible objects. We also compare our representation to cut and bend representations of impossible objects.

Learning-Augmented Competitive Algorithms for Spatiotemporal Online Allocation with Deadline Constraints

2026-03-08T03:23:15Z

Learning-Augmented Competitive Algorithms for Spatiotemporal Online Allocation with Deadline Constraints Lechowicz, Adam; Christianson, Nicolas; Sun, Bo; Bashir, Noman; Hajiesmaili, Mohammad; Wierman, Adam; Shenoy, Prashant We introduce and study spatiotemporal online allocation with deadline constraints (SOAD), a new online problem motivated by emerging challenges in sustainability and energy. In SOAD, an online player completes a workload by allocating and scheduling it on the points of a metric space (X, d) while subject to a deadline T. At each time step, a service cost function is revealed that represents the cost of servicing the workload at each point, and the player must irrevocably decide the current allocation of work to points. Whenever the player moves this allocation, they incur a movement cost defined by the distance metric d(•, •) that captures, e.g., an overhead cost. SOAD formalizes the open problem of combining general metrics and deadline constraints in the online algorithms literature, unifying problems such as metrical task systems and online search. We propose a competitive algorithm for SOAD along with a matching lower bound establishing its optimality. Our main algorithm, ST-CLIP, is a learning-augmented algorithm that takes advantage of predictions (e.g., forecasts of relevant costs) and achieves an optimal consistency-robustness trade-off. We evaluate our proposed algorithms in a simulated case study of carbon-aware spatiotemporal workload management, an application in sustainable computing that schedules a delay-tolerant batch compute job on a distributed network of data centers. In these experiments, we show that ST-CLIP substantially improves on heuristic baseline methods. SIGMETRICS Abstracts ’25, Stony Brook, NY, USA

Faraday Cage Estimation of Normals for Point Clouds and Ribbon Sketches

2026-03-08T03:23:22Z

Faraday Cage Estimation of Normals for Point Clouds and Ribbon Sketches Scrivener, Daniel; Cui, Daniel; Coldren, Ellis; Abulnaga, Mazdak; Bessmeltsev, Mikhail; Chien, Edward We propose a novel method (FaCE) for normal estimation of unoriented point clouds and VR ribbon sketches that leverages a modeling of the Faraday cage effect. Input points, or a sampling of the ribbons, form a conductive cage and shield the interior from external fields. The gradient of the maximum field strength over external field scenarios is used to estimate a normal at each input point or ribbon. The electrostatic effect is modeled with a simple Poisson system, accommodating intuitive user-driven sculpting via the specification of point charges and Faraday cage points. On inputs sampled from clean, watertight meshes, our method achieves comparable normal quality to existing methods tailored for this scenario. On inputs containing interior structures and artifacts, our method produces superior surfacing output when combined with Poisson Surface Reconstruction. In the case of ribbon sketches, our method accommodates sparser ribbon input while maintaining an accurate geometry, allowing for greater flexibility in the artistic process. We demonstrate superior performance to an existing approach for surfacing ribbon sketches in this sparse setting.

Draft Recommendations of the MIT Ad Hoc Faculty Task Force on Open Access to MIT's Research

2026-02-14T03:08:03Z

Draft Recommendations of the MIT Ad Hoc Faculty Task Force on Open Access to MIT's Research Abelson, Harold; Bourg, Chris; Bebergal, Peter; Bond, Robert A.; Cheng, Herng Yi; Chuang, Isaac L.; Cummins, Christopher C; Fitzgerald, Deborah K; Jarzombek, Mark; Lindsay, Nick; Pollard, Tom Joseph; Reid, Jack; Shirer, Karen; Trout, Bernhardt L.; Vander Heiden, Matthew G.; von Hippel, Eric A; Wilcoxson, Jaren D.; Finnie, Ellen; Dunn, Katharine H.

Networking Systems for Video Anomaly Detection: A Tutorial and Survey

2026-03-08T03:22:52Z

Networking Systems for Video Anomaly Detection: A Tutorial and Survey Liu, Jing; Liu, Yang; Lin, Jieyu; Li, Jielin; Cao, Liang; Sun, Peng; Hu, Bo; Song, Liang; Boukerche, Azzedine; Leung, Victor The increasing utilization of surveillance cameras in smart cities, coupled with the surge of online video applications, has heightened concerns regarding public security and privacy protection, which propelled automated Video Anomaly Detection (VAD) into a fundamental research task within the Artificial Intelligence (AI) community. With the advancements in deep learning and edge computing, VAD has made significant progress and advances synergized with emerging applications in smart cities and video internet, which has moved beyond the conventional research scope of algorithm engineering to deployable Networking Systems for VAD (NSVAD), a practical hotspot for intersection exploration in the AI, IoVT, and computing fields. In this article, we delineate the foundational assumptions, learning frameworks, and applicable scenarios of various deep learning-driven VAD routes, offering an exhaustive tutorial for novices in NSVAD. In addition, this article elucidates core concepts by reviewing recent advances and typical solutions and aggregating available research resources accessible at https://github.com/fdjingliu/NSVAD. Lastly, this article projects future development trends and discusses how the integration of AI and computing technologies can address existing research challenges and promote open opportunities, serving as an insightful guide for prospective researchers and engineers.

Analysis and Performance Evaluation of Blockchain Consensus Mechanisms for Network Sharing

2026-03-08T03:23:06Z

Analysis and Performance Evaluation of Blockchain Consensus Mechanisms for Network Sharing Zeydan, Engin; MANGUES-BAFALLUY, JOSEP; Arslan, Suayb; Turk, Yekta; Antevski, Kiril The growing demand for mobile data services has made it necessary to find efficient and cost-effective ways to share networks. Blockchain technology is a promising solution to the challenges of network sharing, such as interoperability, trust, and accountability. This paper presents a comprehensive classification and categorization of blockchain-based network sharing scenarios, highlighting their advantages and limitations. Seven network sharing scenarios are identified, ranging from centralized network sharing to fully decentralized spectrum sharing. The suitability of some selected blockchain consensus algorithms (namely Proof-of-Work (PoW) with Ethereum, Proof-of-Authority (PoA) with Ethereum, Practical Byzantine Fault Tolerance (PBFT) with Tendermint and Proof-of-Stake (PoS) with Cosmo) is assessed for selected scenarios through extensive evaluations. This paper also identifies gaps and opportunities in blockchain-based network sharing solutions, and presents future research directions.

MapTune: Versatile ASIC Technology Mapping via Reinforcement Learning Guided Library Tuning

2026-03-08T03:22:54Z

MapTune: Versatile ASIC Technology Mapping via Reinforcement Learning Guided Library Tuning Liu, Mingju; Robinson, Daniel; Li, Yingjie; Maximilian Kuehn, Johannes; Liang, Rongjian; Ren, Haoxing; Yu, Cunxi Technology mapping involves mapping logical circuits to a library of cells. Traditionally, the full technology library is used, leading to a large search space and potential overhead. Motivated by randomly sampled technology mapping case studies, we propose a MapTune framework that addresses this challenge by utilizing reinforcement learning to make design-specific choices during cell selection. By learning from the environment, MapTune refines the cell selection process, resulting in a reduced search space and potentially improved mapping quality. The effectiveness of MapTune is evaluated on a wide range of benchmarks, different technology libraries, and various technology mappers. The experimental results demonstrate that MapTune achieves higher mapping accuracy and reduces delay/area across diverse circuit designs, technology libraries, and mappers. The paper also discusses the Pareto-Optimal exploration and confirms the perpetual delay-area trade-off. Conducted on benchmark suites ISCAS 85/89, ITC/ISCAS 99, VTR8.0, and EPFL benchmarks, the post-technology mapping and post-sizing quality-of-results (QoR) have been significantly improved, with average Area-Delay Product (ADP) improvement of 16.56\% among all different exploration settings in MapTune. The improvements consistently remained for four different technologies (7nm, 45nm, 130nm, and 180 nm) with various mappers from both state-of-the-art open-source and commercial synthesis tools.

A Two-Stage Approach to Improve Poverty Mapping Spatial Resolution

2026-03-08T03:39:51Z

A Two-Stage Approach to Improve Poverty Mapping Spatial Resolution Salas, Joaquín; Zea-Ortiz, Marivel; Vera, Pablo; Wood, Danielle Global extreme poverty has fallen dramatically over the past two centuries, yet hundreds of millions remain impoverished, underscoring the need for scalable monitoring tools. In Mexico, poverty metrics are available only sporadically in terms of time and space (e.g., every 5 years at the municipal level), making it difficult for decision-makers to access reliable, up-to-date, and sufficiently detailed information, highlighting the need for higher-resolution, timely methods. To address this problem, we propose a two-stage approach that combines socioeconomic and Earth Observations-based data. Initially, a machine learning model maps census variables to official poverty indicators belonging to a multidimensional model, yielding fine-scale poverty estimates. A census-based model trained with eXtreme Gradient Boosting (XGBoost) achieved a determination coefficient (𝑅2) of approximately 0.842, indicating strong agreement with official poverty figures and providing high-resolution proxies. Afterward, we use features based on remote observations to predict these poverty estimates at a 469 m grid scale. In this case, advanced foundation models outperformed other machine learning (ML) approaches, achieving an 𝑅2 of 0.683. While foundation models enable more accurate, fine-scale poverty mapping and could accelerate poverty assessments, their use comes at a heavy price in terms of carbon emissions.

Effectiveness of a Participatory Voice Intervention on Psychological Well-Being Among Warehouse Workers: Results From the Fulfillment Center Intervention Study, United States, 2021‒2023

2026-03-08T03:39:54Z

Effectiveness of a Participatory Voice Intervention on Psychological Well-Being Among Warehouse Workers: Results From the Fulfillment Center Intervention Study, United States, 2021‒2023 Siebach, Kirsten F.; Diaz-Linhart, Yaminette; Kubzansky, Laura D.; Berkman, Lisa F.; Wang, Molin; Ge, Lin; Kowalski, Alexander M.; Rahmandad, Hazhir; Kelly, Erin L. Objectives. To examine whether a novel workplace intervention designed to increase worker voice can reduce psychological distress and improve emotional vitality at 6- and 12-months follow-up. Methods. We conducted a cluster-randomized trial in 16 fulfillment centers throughout the United States between 2021-2023. Data were collected at three time points; 2813 workers participated in at least one survey. Treated fulfillment centers established a new, participatory committee called the Health and Well-Being Committee (HaWC). We compared differences in psychological distress and emotional vitality and explored differential treatment effects by gender. Results. At baseline, moderate or severe psychological distress was 51%. Intervention sites had lower average psychological distress at the 6-month follow-up compared to control sites, with no significant differences at 12-month follow-up. Gender moderation analyses suggest the HaWC was particularly effective in reducing psychological distress among men at 6-month follow-up. Conclusions. Our findings suggest that opportunities for workers to share concerns to a committee of their peers tasked with identifying solutions can support mental health. Our study contributes important experimental evidence on workplace interventions that improve the well-being of low-wage U.S. populations.

ML Prediction Models to Identify Novel Beyond Visual Range Tactics and Error Analysis for DARPA AIR Agents

2026-02-13T05:07:40Z

ML Prediction Models to Identify Novel Beyond Visual Range Tactics and Error Analysis for DARPA AIR Agents Li, William; Castor, Jeremy This paper investigates the utility of using machine learning models to predict the outcome of simulated 2 vs. 2 Tactical Intercept engagements flown by autonomous agents in support of the DARPA Artificial Intelligence Reinforcements (AIR) program. We investigated the performance of four models: Feed Forward Neural Network, Random Forest, Extreme Gradient Boost, and Long Short Term Memory (LSTM). We examined their ability to successfully predict the outcomes of simulated engagements, tactical errors, and the execution of novel game plans by autonomous agents. The models were trained on 53 features pertaining to the agents including distance between aircraft, altitude, speed, missile availability, and other eventbased features from simulated runs. The LSTM model had the best performance towards the beginning of a run and was able to predict the correct winner with 87.8% accuracy only one minute into a run while the XGBoost model achieved the best overall performance with a 91.7% classification accuracy and an R² of 0.712. The XGB model was also able to correctly predict the winner of 84.7% of the runs after only seven minutes into the simulated engagement. These results demonstrate the utility and need for further investigation into other ML models potential to identify unique attributes and predictive analysis of more complex multi-agent scenarios that include additional criteria such as varying rules of engagement, incorporating acceptable levels of risk as well as other requirements fighter pilots must take into account during offensive and defensive operations needed to gain air superiority and support the objectives of the Joint Forces Commander.

Open Access at MIT and Beyond: A White Paper of the MIT Ad Hoc Task Force on Open Access to MIT's Research

2026-02-13T04:26:28Z

Open Access at MIT and Beyond: A White Paper of the MIT Ad Hoc Task Force on Open Access to MIT's Research Dunn, Katharine H.; Abelson, Harold; Bourg, Chris; Finnie, Ellen

Enabling New Science with the ALMA Phasing System - Phase 2 (APP2): An ALMA North America Development Project

2026-03-04T18:32:50Z

Enabling New Science with the ALMA Phasing System - Phase 2 (APP2): An ALMA North America Development Project Matthews, L. D.; Crew, G. B. This document provides a summary of activities undertaken as part of the Cycle 5 ALMA North America Development Project “Enabling New Science with the ALMA Phasing System - Phase 2 (APP2)”, whose period of performance extended from January 1, 2018 to August 31, 2024. APP2 provided a series of enhancements to ALMA’s very long baseline interferometry (VLBI) and phased array capabilities, leading to the introduction of submillimeter (Band 7) phasing and VLBI capabilities, a passive phasing mode, a Phased Array (pulsar) observing mode, a prototype spectral line VLBI capability, an improved method of handling baseband delays, and a number of other minor system enhancements. This report was prepared as a final report on the activities undertaken under the NA Development program mentioned in the abstract.

Mechanisms of Interaction Between Hydraulic and Natural Fractures in Shale Rocks

2026-02-13T03:10:59Z

Mechanisms of Interaction Between Hydraulic and Natural Fractures in Shale Rocks Arzuaga García, Ignacio Martín Understanding the interaction between hydraulically induced fractures and pre-existing natural fractures in geologic formations is key for optimizing subsurface energy systems that rely on fluid injection into fractured rocks. These include Enhanced Geothermal Systems (EGS), CO₂ sequestration, hydrogen storage in depleted reservoirs, unconventional oil and gas development in shale formations, and nuclear waste disposal, among others. In all these applications, controlling fracture propagation and interaction is essential for ensuring operational efficiency, safety, and long-term integrity of the system. This thesis presents a comprehensive experimental and theoretical investigation of hydraulic fracture (HF) interactions with natural fractures (NFs), using Opalinus Clayshale as a representative anisotropic material. The experimental work involved a series of hydraulic fracturing tests on Opalinus Clayshale specimens under controlled quasi-true-triaxial stress conditions, comparing normal and dried states. Novel monitoring techniques, including high-resolution imaging, high-speed video, acoustic emissions (AE), and pressure tracking, were employed to capture the fracturing process in real-time. Three dominant interaction modes (Crossing, Arrest, and Opening) were systematically characterized and linked to key parameters, including stress ratio, fracture geometry, and injection rates. A critical stress ratio (σ₁/σ₃) of approximately 20 was identified as the threshold for achieving fracture crossing under our experimental conditions: cohesionless, “open” natural fractures, with a low viscosity injection fluid, in a toughness-dominated regime. In dried specimens, high flaw pressurization rates were necessary to overcome matrix fluid loss and achieve crossing. To complement and interpret the experimental results, existing theoretical models were reviewed and implemented. Furthermore, a simplified version of the OpenT model (Chuprakov et al., 2014) was developed and applied for Opalinus Clayshale, incorporating stress, energy, friction, and permeability effects. By integrating laboratory results with theoretical frameworks, this thesis offers an integral approach to predictive understanding of fracture propagation in naturally fractured rocks, stating that not only the characteristics of the discontinuity or the far-field stresses involved in the process are important in determining the mechanism of interaction, but also the dynamic energy balance at the fracture tip, which is influenced by injection rate, fluid viscosity, and discontinuity properties. Overall, this thesis bridges the gap between laboratory experiments and theoretical models, advancing a more comprehensive understanding of fracture propagation in naturally fractured media. The findings highlight the importance of considering both mechanical and hydraulic parameters, particularly in low-viscosity, toughness-dominated regimes, for accurately predicting fracture behavior.

Satellite Drag and Sustainable Space Operations in a Dynamic Thermosphere

2026-02-13T03:11:02Z

Satellite Drag and Sustainable Space Operations in a Dynamic Thermosphere Parker, William E. Earth’s orbit has become increasingly congested and contested in recent years. The surge in launched payloads, combined with satellite failures, explosions, and collisions, has contributed to a large and growing population of orbital debris objects that can remain in orbit for decades, centuries, or longer. Meanwhile, decreasing launch costs and maturing satellite technology have created conditions favorable for rapid commercialization across orbital regimes, especially in low Earth orbit (LEO). Today, a small number of commercial entities operate the large majority of the world’s active satellites as part of proliferated LEO constellations. Sustaining productive activity in an increasingly crowded orbital environment has made satellite conjunction assessment and collision avoidance essential for safe operations. These efforts require not just accurate trajectory predictions, but also credible estimates of uncertainty. In LEO, variability in atmospheric drag is by far the dominant source of propagation error, often leading to deviations of several kilometers per day due to unpredictable solar and geomagnetic activity. Even over short timescales, trajectory prediction is challenging because existing forecasts exhibit limited predictive skill. Although forecast errors are often non-Gaussian and heteroscedastic, operational products are generally presented as deterministic, and atmospheric models rarely provide rigorous uncertainty characterization. This work introduces a new approach for probabilistic satellite drag modeling based on historical correlations between space weather drivers and satellite dynamics. Unlike traditional methods, it models satellite behavior directly without reconstructing thermospheric mass density or requiring detailed knowledge of satellite properties such as the ballistic coefficient. This end-to-end strategy offers substantial computational and operational advantages for many space domain awareness tasks. Capturing both trajectory predictions and their associated uncertainty is critical for enabling informed collision avoidance decisions, particularly during geomagnetic storms when current infrastructure frequently fails. Because the orbital lifetime of debris objects can exceed hundreds of years, population dynamics in space critically depend on long-term variability in the composition of Earth’s thermosphere. Rising concentrations of carbon dioxide and other greenhouse gases have caused warming in the troposphere but cooling and contraction in the upper atmosphere. This contraction decreases atmospheric density in LEO, reducing drag and extending the orbital lifetime of debris objects. Longer-lived debris populations pose a persistent collision hazard for all active satellites as long as they remain in orbit. Even natural events, such as a prolonged grand solar minimum, could further reduce thermospheric density and contribute to longer debris lifetime in LEO. With little ability to predict such an event, it is necessary to understand the potential consequences and to identify strategies that enable the continued safe and productive use of LEO. This work models the impact of such long-term environmental changes on limits for sustainable satellite deployments. LEO is a finite respource increasingly at risk of overexploitation. Conserving it and sharing it fairly requires that we first understand its fundamental capacity and our current occupation of that capacity. Some metrics have been proposed to measure the satellite carrying capacity of Earth’s orbit, but none have previously accounted for the potential influence of a changing space climate. This work develops new methods for defining carrying capacity as a common currency, enabling clear constraint-driven thresholds on activity and a better understanding of how existing and proposed missions consume available capacity. These new metrics provide insight into how environmental variability may affect the long-term sustainability of operations in LEO. Respecting and understanding this influence that the natural environment has on our collective ability to operate spacecraft in LEO is critical to preventing the overexploitation of this regime and protecting it for future generations.

Evaluating Large Language Models as Circuit Design Assistants

2026-02-13T03:49:28Z

Evaluating Large Language Models as Circuit Design Assistants Cox, Matthew J. Large language models (LLMs) have exploded in capability in recent years. Previous attempts at AI systems for circuit design have had limited proficiency and been restricted in problem scope. LLMs, with their breadth of knowledge and reasoning ability, are a promising technology for a much more general-purpose circuit design assistant. We developed a dataset of electrical engineering problems and solutions with which to test an LLM-based system, since no such publicly available dataset exists to our knowledge; unmodified GPT-4 was able to solve 42% of the problems. We did a preliminary comparison of several knowledge bases to use for RAG knowledge injection, finding that a small, curated set of resources performed better than a larger, less-focused set of resources, though there were confounding factors which may have skewed the result. While this work is a start, significant future work is needed to continue developing an LLM-based circuit design assistant.

Increasing Program Code Coverage Using Fuzzing and Targeted Branch Exploration

2026-02-13T03:49:33Z

Increasing Program Code Coverage Using Fuzzing and Targeted Branch Exploration Nguyen, Gary Code coverage is a longstanding metric for evaluating how thoroughly a program has been tested. Achieving high coverage remains a priority goal for quality assurance and software stability. Exhaustive enumeration of possible input paths to every code region is desirable in theory but computationally infeasible in practice, especially in large-scale codebases. Fuzzing is a widely used technique for input generation and is effective at exploring smaller programs but often struggles with more complex conditional logic and nested modules. Concolic execution, which exhaustively explores paths using constraint solving, can work effectively with complex conditional logic but suffers from path explosion. Targeted branch exploration is a similar approach for input generation but sidesteps the path explosion problem by focusing more on specific constraint paths of interest. In this thesis, I introduce a hybrid system that combines fuzzing and targeted branch exploration with the goal of improving code coverage by leveraging the complementary strengths of each. The system uses fuzzing to quickly generate a broad input corpus and follows up with targeted branch exploration to explore paths that fuzzing struggles to reach. Findings from experiments on two C projects of different complexities show that the system did not outperform the individual techniques in terms of raw coverage, revealing limitations of the approach and opportunities for future improvement.

Essays in Financial Economics and Econometrics

2026-02-13T03:10:50Z

Essays in Financial Economics and Econometrics Orestes, Victor M. This thesis comprises three essays in finance and econometrics. The first two essays focus on the role of credit access and liquidity in shaping real firm outcomes. The first essay examines the transmission of modern monetary policy through corporate asset markets. Exploiting quasi-experimental variation in the Central Bank of Brazil’s collateral framework and implementing a novel dynamic regression discontinuity design, it shows that monetary policy can ease expected future borrowing constraints, reduce firms’ precautionary cash holdings, and stimulate employment. The second essay analyzes how receivables financing through factoring helps firms smooth cash flows. Using a shift-share instrument and matched administrative data, it finds that cheaper liquidity leads firms to rely more on permanent labor. The third essay develops a new method for distributional inference—nonparametric quantile mixture models. This framework can be applied to financial settings such as tail risk estimation and density forecasting, as well as to causal inference when the objective is to estimate the distributional effects of interventions. It is used here to quantify the heterogeneous wage effects of a major environmental disaster. The first chapter (joint with Luis Alvarez and Thiago Christiano Silva) studies how modern monetary policy tools, which increasingly operate through corporate asset markets, affect real firm outcomes. We exploit quasi-experimental variation from the inclusion of specific corporate debt instruments in the Central Bank of Brazil’s collateral framework and implement a novel dynamic regression discontinuity design. We find that eligibility increases firms’ debt issuance, modestly lowers spreads, and reduces cash holdings, reflecting a decline in precautionary savings. These effects translate into higher employment and greater supply chain liquidity. We interpret the mechanism through the lens of segmented financial markets: by relaxing firms’ expected future borrowing constraints, the policy acts as a persistent borrowing subsidy and liquidity injection. This encourages firms to reduce cash hoarding and expand production. Using a semi-structural framework calibrated to our reduced-form estimates, we find that an induced 0.8% borrowing subsidy leads to a 1% increase in debt issuance, a 0.2% reduction in cash holdings, and a 0.4% increase in the wage bill. The second chapter (joint with Thiago Christiano Silva and Henry Zhang) shows that firms experience large increases in sales and purchases after receiving cheaper liquidity. We focus on factoring, defined as the supplier-initiated sale of receivables. In Brazil, receivables funds (FIDCs) securitize receivables for institutional investors. By assembling a novel transaction-level dataset of factoring with other credit operations for all registered firms and FIDCs, we construct a shift-share instrument for factoring financing supply based on FIDC flows. We then use a novel combination of electronic payments, trade credit, and employer-employee matched data to estimate the impacts. A flow-induced increase in receivables demand reduces firms’ factoring interest rate. In response, firms demand more permanent labor and less temporary labor. In our model, these effects arise from factoring’s purpose of reducing cash inflow volatility, helping firms match inflows to outflows, which firms otherwise achieve at an efficiency cost through substitution across labor types. The third chapter (joint with Luis Alvarez) introduces nonparametric quantile mixture models as a computationally convenient and flexible alternative to standard nonparametric density mixtures, which are widely used in Statistics and Econometrics but face significant computational and inferential challenges. We propose a sieve estimator based on a generalized method of L-moments and develop a full inferential theory. In doing so, we contribute to the statistical literature by extending a numerical bootstrap method to high-dimensional settings. As a direct application of our theory, we provide the first inference method for the distributional synthetic controls of Gunsilius (2023), a novel tool for counterfactual analysis that previously lacked formal inference procedures. We apply this method to evaluate the effects of the Brumadinho dam collapse—a large-scale environmental disaster—on the local wage distribution. The results reveal substantial heterogeneity across the distribution, with evidence of displacement effects in which median-paying jobs are replaced by lower-wage contracts. JEL Codes: C1, E4, E5, G2, G3

Machine-Learned Representations of Basis Sets and Their Application in Quantum Computational Chemistry

2026-02-13T03:49:30Z

Machine-Learned Representations of Basis Sets and Their Application in Quantum Computational Chemistry He, Wenhao Quantum simulations of electronic structure promise to deliver significant speedups over classical methods, but remain limited by the number of qubits on near-term devices. A key strategy to reduce quantum resource requirements is to truncate the molecular Hilbert space via compact and efficient basis sets. However, most optimized basis sets either rely on predefined heuristics or require expensive classical computations, such as CASSCF orbital optimization or ℓ1-norm minimization of the Hamiltonian. In this work, we introduce a general machine learning framework for fast basis set prediction in quantum computational chemistry. Our method employs an equivariant graph neural network that outputs a Hermitian matrix encoding optimized molecular orbitals. The eigenvectors of this matrix define a transferable and efficient basis set, trained on orbitals obtained via CASSCF and Hamiltonian ℓ1 norm optimization. We evaluate our model on hydrogen chains and demonstrate that the predicted bases achieve energy accuracy and Hamiltonian sparsity comparable to orbital-optimized methods, while reducing classical preprocessing time. In addition, the predicted orbitals can be directly used as high-quality initial guesses for CASSCF calculations, further accelerating their convergence.

Signaling at the Tumor-Immune Interface in Glioblastoma

2026-02-13T03:10:43Z

Signaling at the Tumor-Immune Interface in Glioblastoma D'Souza, Alicia D. Glioblastoma (GBM) is a devastating brain cancer, and the standard of care has not changed in over 20 years. GBM tumors are composed of a milieu of cancer cells and innate immune cells, which are co-opted by the cancer cells to promote an anti-inflammatory environment. Despite tremendous success in immunotherapy in several cancers over the past 10 years, immunotherapies have failed to show efficacy in GBM. A systems biology approach to characterizing temporal changes in tumor-immune interface of glioblastoma could illuminate new strategies to activate an anti-tumor immune response by examining changes in cell signaling and antigen presentation. In the first part of my thesis, I investigated how macrophages alter their phenotype in response to tumor co-culture and how these changes are reflected at the level of the phosphoproteome. To characterize signaling changes in distinct cell populations during co-culture, I developed a method to preserve and analyze cell-type-specific signaling using fixation. This approach enables phosphoproteomic profiling of two interacting cell types, capturing dynamic signaling events with cell-type resolution. I applied this method to study co-cultures of glioblastoma (GBM) cells and primary human macrophages. When cultured together, GBM cells induced an anti-inflammatory, immunosuppressive phenotype in macrophages, mirroring features observed in the glioblastoma tumor microenvironment. Phosphoproteomic analysis revealed that this phenotypic shift was accompanied by distinct signaling alterations in macrophages, including the upregulation of ABL kinase activity. To test this finding, I treated macrophages with an ABL kinase inhibitor and observed a reduction in the anti-inflammatory phenotype, suggesting that ABL signaling plays a role in supporting immunosuppressive macrophage polarization. Furthermore, in a mouse model of GBM, treatment with an ABL kinase inhibitor led to a reduction in the abundance of anti-inflammatory macrophages within the tumor and was associated with a modest extension of survival. In the second part, I examined changes in antigen presentation and signaling in glioblastoma tumors in response to treatment with an oncolytic virus (OV). In patient derived tumor (PDX) models in mice, mice treated with OV have increased antigen presentation, pointing to the use of OV therapy to reshape the tumor micro-environment to a more inflammatory state. Finally, tissue obtained from serial biopsies of GBM patients treated with OV shows an increase in antigen presentation and both Class I and Class II MHC protein expression. We also observed an increase in interferon alpha and interferon gamma signaling pathways as well as early induction of apoptotic pathways. These findings highlight the role of therapeutics in altering the tumor microenvironment and potentially priming it for combination immunotherapies. This thesis explores the dynamic nature of the tumor and immune compartments in glioblastoma and underscores how therapies can act on the immune compartment to promote anti-tumor activity.

SmellNet: A Large-scale Dataset for Real-world Smell Recognition

2026-02-13T03:49:31Z

SmellNet: A Large-scale Dataset for Real-world Smell Recognition Feng, Dewei The ability of AI to sense and identify various substances based on their smell alone can have profound impacts on allergen detection (e.g., smelling gluten or peanuts in a cake), monitoring the manufacturing process, and sensing hormones that indicate emotional states, stress levels, and diseases. Despite these broad impacts, there are virtually no large-scale benchmarks, and therefore little progress, for training and evaluating AI systems’ ability to smell in the real world. In this paper, we use portable gas and chemical sensors to create SmellNet, the first large-scale database that digitizes a diverse range of smells in the natural world. SmellNet contains about 180,000 time steps of 50 substances (spanning nuts, spices, herbs, fruits, and vegetables) with 50 hours of data. Using SmellNet, we trained AI models for real-time classification of substances based on their smell alone. Our best methods leverage sequence models, contrastive learning to integrate high-resolution Gas Chromatography–Mass Spectrometry molecular data, and a new temporal difference method that identifies sharp changes in sensor readings. Our best models achieve up to 65.35% accuracy on pre-recorded data, and generalize to real-world conditions with 10.71% accuracy on nuts and 25.38% on spices in the challenging 50-way online classification task. Despite these promising results, SmellNet highlights many technical challenges in building AI for smell, including richer feature learning, on-edge smell models, and robustness to environmental changes.

Towards Zero-Shot Pretrained Models for Efficient Black-Box Optimization

2026-02-13T03:49:16Z

Towards Zero-Shot Pretrained Models for Efficient Black-Box Optimization Meindl, Jamison Chivvis Global optimization of expensive, derivative-free black-box functions requires extreme sample efficiency. While Bayesian optimization (BO) is the current state-of-the-art, its performance hinges on surrogate and acquisition function hyperparameters that are often hand-tuned and fail to generalize across problem landscapes. We present ZeroShotOpt, the first general-purpose, pretrained model for continuous black-box optimization tasks ranging from 2 D to 20 D. Our approach leverages offline reinforcement learning on large-scale optimization trajectories collected from 12 BO variants. To scale pretraining, we generate millions of synthetic Gaussian process-based functions with diverse landscapes, enabling the model to learn transferable optimization policies. As a result, ZeroShotOpt achieves robust zero-shot generalization on a wide array of unseen synthetic and real-world benchmarks, matching or surpassing the sample efficiency of leading global optimizers, including BO, while also offering a reusable foundation for future extensions.

Temperature Characterization of Colloidal Quantum Dot Light Emitting Diodes

2026-02-13T03:49:26Z

Temperature Characterization of Colloidal Quantum Dot Light Emitting Diodes Nguyen, Thienan D. Colloidal quantum dot light emitting diodes reveal to be promising candidates for the next generation of display technologies. Their brighter emissions, greater color purity, and higher efficiency make them highly desirable in consumer electronics. As such, research into the performance and stability of these novel LEDs are crucial for their operation in displays. These investigations are ongoing, with focused efforts on improving the operating stability through different quantum dot materials and passivation methods. However, less attention is paid in confidently understanding the fundamental relationships between current, voltage, and luminance by which these devices operate. These electrical characteristics reveal insights into the operation of these devices and the behavior of charge carriers. Additionally, temperature-dependent electrical measurements can showcase different behavior at different temperatures and deviations from the expected performance at set temperatures. Temperature dependent processes are revealed and from such, a better understanding of how the device operates is gained. In this thesis, an investigation into the temperature-dependent electrical characteristics of quantum dot light emitting diodes was conducted by measuring the current-voltage-luminance, JVL, relationships at various cryogenic temperatures. These temperatures ranged from 78K, liquid nitrogen boiling point, to 293K, room temperature. This investigation revealed the temperature dependent nature and origin of turn-on voltage, current, EQE, EQE roll-off, and hysteresis.

Uncertainty-Aware Knowledge Graph Retrieval Methods and Their Use in LLM Question-Answering

2026-02-13T03:49:13Z

Uncertainty-Aware Knowledge Graph Retrieval Methods and Their Use in LLM Question-Answering Rich, Benjamin R. Knowledge Graph Question Answering (KGQA) encompasses a set of techniques aimed at generating accurate, interpretable responses to natural language queries posed over structured, graph-based datasets. Recent approaches to KGQA involve reducing the knowledge graph (KG) to a relevant subgraph, which is then encoded in natural language as a series of triples (subject, predicate, object) and passed to a large language model (LLM) for interpretation and answer generation. These methods have shown state-of-the-art accuracy. However, this paradigm is undermined by a critical vulnerability: the retrieval of irrelevant or erroneous facts can amplify LLM hallucinations and degrade system trustworthiness, while the reasoning process remains opaque. This thesis addresses this challenge by extending an existing stateof-the-art KGQA architecture with uncertainty-aware subgraph retrieval methods. To achieve this, we modify the retrieval component to learn the epistemic uncertainty of each candidate triple’s relevance to a given query. We implement these modifications using Bayesian methods and learn a well-calibrated approximation of the posterior distribution over triple relevance. By explicitly modeling this uncertainty, the retriever model is shown to provide a fine-grained confidence score for each piece of evidence. We expose these metrics downstream to the LLM during reasoning and evaluate whether LLMs can reason over uncertainty-related metrics to improve KGQA. We find that LLMs cannot reason effectively over uncertainties in most cases, but that agentic workflows that provide selective access to uncertainty metrics may enhance performance. We evaluate our approach against established benchmarks using HIT-rate and set-comparison accuracy metrics. Additionally, we introduce reasoning-path and statistical trust metrics derived from calibrated uncertainty scores. Our analysis reveals a significant positive correlation between path-based uncertainty metrics and the veracity of the Large Language Model’s (LLM) answers. These findings establish a robust foundation for developing uncertainty-grounded trust mechanisms in LLM-agnostic KGQA systems. As a proof of concept, a lightweight classifier trained exclusively on the LLM’s inputs and outputs demonstrates substantial predictive power in identifying correct responses. Finally, we briefly explore using uncertainty to identify out-of-distribution (OOD) queries.

Applied Compiler Optimizations for Proving Code

2026-02-13T03:49:29Z

Applied Compiler Optimizations for Proving Code Ruiz, Ricardo The recent popularity of massively distributed, trustless systems has created a demand for cryptographic proofs: systems to prove that a piece of data is a valid output for a given program. These systems exist, but face very high runtimes for the generation of proofs. Significant effort has been invested in optimizing the prover systems, but relatively less has been focused on optimizing the code that gets read as an input. This paper proposes a new approach to optimizing prover systems by modifying the compiler to produce proof-ready code. It proposes a benchmarking framework for comparing the relative proof costs of RISC-V instructions; the resulting analyis find that shift instructions do not offer heavy savings over multiplication. The finding suggests that strength reduction, a fundamental optimization in modern compilers, can sabotage end-to-end performance. The paper proposes methods for applying this knowledge to better optimize code, leaving the door open for future researchers to continue to make code proofs more performant and accessible.

Reconstructing Cross-Species Ancestral Adeno-Associated Viruses for Enhanced Gene Therapy Delivery

2026-02-13T03:49:32Z

Reconstructing Cross-Species Ancestral Adeno-Associated Viruses for Enhanced Gene Therapy Delivery Xie, Yuxin Adeno-associated viruses (AAV) are one of the most promising vectors for gene therapy because of their established safety, low immunogenicity, and capability to achieve sustained gene expression. However, many naturally occurring AAV variants have limitations in their potency, particularly in penetrating biological barriers like the blood-brain barrier (BBB). Additionally, their broad and nonspecific tropism can translate into suboptimal cross-species transduction efficiency and potential toxicity, complicating the clinical transition from animal model to humans. These challenges impede the use of naturally occurring AAVs for therapeutic gene delivery in many neurological disorders-such as autism spectrum disorders (ASD), Parkinson’s disease (PD), Huntington’s disease (HD)—as well as other systemic conditions like cystic fibrosis (CF). To overcome these barriers, we developed a computational framework based on ancestral sequence reconstruction (ASR) to engineer synthetic ancestral AAV capsids with the goal of enhanced targeting specificity and potency. We first validated this computational framework by replicating the previously engineered Anc80L65 capsid. Then, with 75 naturally occurring functional AAV sequences and additional experimentally screened variants exhibiting brain-targeting potency, we built an evolutionary framework. We applied multiple computational methods such as enhanced multiple sequence alignment, maximum-likelihood-based phylogenetic tree inference, and ancestral sequence reconstruction with Bayesian inference. With this methodology, we predicted several novel ancestral AAV capsid sequences at critical evolutionary nodes, particularly those representing functional transitions with potential improved blood-brain barrier penetration and CNS tropism. Our computational framework thus streamlines and accelerates the process of designing ancestral AAV variants with targeted gene therapy applications.

Intercellular flow-mediated force relaxation measurement on the three-dimensional multicellular tissue

2026-02-13T03:10:55Z

Intercellular flow-mediated force relaxation measurement on the three-dimensional multicellular tissue Liu, Fan Three-dimensional (3D) multicellular tissues are prevailing over 2D monolayer or single cells; their mechanical properties like stiffness, surface tension, and viscosity have been shown to relate to diseases like fibrosis or tumor metastasis. Multicellular tissues have been traditionally modeled as a viscoelastic material due to their apparent shape rearrangement, which hardly considers the internal structure, including the extracellular matrix (ECM) and resulting intercellular water flow. These intercellular communications usually provide significant information on diseases such as tumor invasion, but immediate supporting evidence of this behavior is lacking. In this work, we investigate the bulk response of 3D multicellular tissues due to such intercellular flows and explore the related mechanism through a tailored micro-mechanics platform. Firstly, we design and establish a micro-mechanics platform based on the parallel plate compression (PPC) method. We adopt a precise micro-balance as the sensor to detect the force variation of the sample during compression. A piezo linear stage is incorporated to exert such tiny vertical displacement. Besides, a lateral microscope is designed to monitor the compression process instantaneously. This platform has proved to be applicable to various samples, including hydrogels, cell spheroids, and natural tissues or organs. Then, we propose the critical criterion, the size dependency of force relaxation time, to distinguish a material's properties, i.e., viscoelasticity and poroelasticity. For poroelastic material, the force relaxation is due to water redistribution; hence, the speed highly depends on the sample sizes. In contrast, for viscoelastic material, it is determined by the bulk material properties, thus independent of the size. We theoretically verify this criterion via Abaqus simulation and experimentally on classic poro-/visco-elastic materials with various dimensions. Next, we apply the size-dependency criterion on the 3D multicellular tissues to distinguish the poro-/visco-elasticity in this biomaterial. We take the PPC on multiple cell spheroids with different sizes through the platform. It is observed that the force relaxation times are linearly proportional to the size of all tested cell lines, demonstrating poroelasticity in our experimental time range. Intriguingly, we take tests on the natural organs of the mouse islets and find such linear correlation as well. Hence, both cultured spheroids and natural tissues are poroelastic. Finally, we explore the mechanism determining the poroelasticity inside the 3D multicellular tissues. By inhibiting the cell-cell junctions, we demonstrate the intercellular water flow through the extracellular gaps dominates this poroelastic force relaxation in the biomaterial. Further experiments show that the stiffness of the structure and the extracellular gaps inside the 3D multicellular tissues couple to contribute to the intercellular water flow, i.e., the stiffer the structure and/or the larger the gaps, the faster the water flows, thus quicker the force decays after compression. These findings highlight the fundamental role of intercellular water flow in the mechanical properties of 3D multicellular tissues. The designed micro-mechanics platform is also beneficial to research at the tissue level with micro-newton forces owing to the development of artificial organoids for early disease diagnosis and treatment.

Generating Unprecedented Extreme Scenarios with Limited Data

2026-02-13T03:49:19Z

Generating Unprecedented Extreme Scenarios with Limited Data Chang, Kai Quantifying and predicting rare and extreme events persists as a crucial yet challenging task in understanding complex dynamical systems, ubiquitous in science and engineering. Many practical challenges arise from the infrequency and severity of these events, including the considerable variance of simple sampling methods and the substantial computational cost of high-fidelity numerical simulations. Numerous data-driven methods have recently been developed to tackle these challenges. However, a typical assumption for the success of these methods is the occurrence of multiple extreme events, either within the training dataset or during the sampling process. This leads to accurate models in regions of quiescent events but with high epistemic uncertainty in regions associated with extremes. To overcome this limitation, we introduce the framework of Extreme Event Aware (e2a or eta) or η-learning which does not assume the existence of extreme events in the available data. η-learning reduces the uncertainty even in ‘unchartered’ extreme event regions, by enforcing the extreme event statistics of a few observables during training, which can be available or assumed through qualitative arguments or other forms of analysis. This type of statistical regularization results in models that fit the observed data, but also enforces consistency with the prescribed statistics of some observables, enabling the generation of unprecedented extreme events even when the training data lack extremes therein. Theoretical results based on optimal transport offer a rigorous justification and highlight the optimality of the introduced method. Additionally, extensive numerical experiments illustrate the favorable properties of the ηlearning framework on several prototype problems and real-world precipitation downscaling problems.

A*-Decoding: Token-Efficient Inference Scaling

2026-02-13T03:49:18Z

A*-Decoding: Token-Efficient Inference Scaling Chatziveroglou, Ioannis Inference-time scaling has emerged as a powerful alternative to parameter scaling for improving language model performance on complex reasoning tasks. While existing methods have shown strong performance gains under fixed compute budgets, there has been little focus on optimally utilizing that budget during inference. In this work, we introduce A*-decoding, a search-based inference-time strategy that builds on the A* search algorithm to optimally utilize a fixed compute budget by prioritizing high-quality reasoning paths during generation. We frame language model decoding as a structured search in a state space of partial solutions, applying the A* transition model to identify promising continuations guided by an external process supervision signal. In our experiments, A*-decoding reaches the performance levels of strong inference scaling baselines like best-of-N and particle filtering while using up to 3x fewer tokens and 30% fewer PRM passes under equivalent compute budgets. On the MATH500 and AIME 2024 benchmarks, A*-decoding enables Llama-3.2-1B-Instruct to match the performance of the 70x larger Llama-3.1-70B-Instruct, and allows Qwen3-1.7B to reach o1-like reasoning accuracy. These results highlight the power of structured search in decoding, offering an alternative to brute-force sampling or scale-driven gains. Our work demonstrates how thoughtful inference-time strategies can enhance reasoning in SLMs, pointing toward future advances in more efficient and scalable language model deployment.

U-Net Network Enhancements to Facilitate Rapid Electron Microscopy Imaging for Connectomics

2026-02-13T03:49:27Z

U-Net Network Enhancements to Facilitate Rapid Electron Microscopy Imaging for Connectomics Varma, Vikram Imaging the structural and functional connections between cells in the brain allows neuroscientists to understand the brain by studying neuronal wiring diagrams. To automatically segment and classify images that are used in the construction of these neuronal wiring diagrams, or connectomes today, machine learning segmentation techniques require an image scanned with an electron microscope at either a slow dwell time or with small pixel sizes. However, a scalable and more rapid implementation of connectome construction has not yet been realized because of the significant cost of multi-beam electron microscopes and the relatively slow time in which connectomes can be constructed using a single-beam electron microscope. Segmented connectomes include sections that can be segmented properly with a fast scanned image as well as sections that require slow scanning for proper segmentation. Due to this fact, a potential way to enhance the time in which connectomes can be produced and segmented is to first scan samples at fast resolution and perform segmentation using a convolutional neural network, identify those areas of interest that require more detailed imaging through a learning-based error detection network, and then rescan only those identified high interest areas to produce a fused image for segmentation. The proposed thesis will analyze various machine learning methods for segmentation using the U-Net network and review proposed enhancements to the U-Net network that can better utilize electron microscopy images for construction of segmented connectomes. The successful use of fused electron microscopy images will potentially enable higher speed and lower cost electron microscopy imaging for connectomics.

Low-Temperature Germanium Waveguides for Mid-Infrared Sensing Applications

2026-02-13T03:49:21Z

Low-Temperature Germanium Waveguides for Mid-Infrared Sensing Applications Zhang, Erin Wei Waveguide integrated devices that operate in the mid-infrared (mid-IR) wavelength range (2.5-12 µm) are used for sensing the fundamental absorption bands in a variety of molecules. Germanium (Ge) is commonly used for photodetection in the nearinfrared (near-IR) wavelength range of 1.2-1.6 µm due to its strong absorption from a 0.8 eV direct band gap. At longer wavelengths in the mid-IR range, Ge exhibits transparency that makes it a desirable waveguide material for sensing applications. Its epitaxial growth compatibility with silicon (Si) substrates makes Ge-on-Si an effective platform for mid-IR waveguides. For back-end-of-line (BEOL) integration of waveguides in sensing applications, the thermal budget limits the temperature to below 450°C. In this work, we investigated the use of h-line exposure as a commercially viable, low-cost option for patterning low temperature (LT) Ge-on-Si waveguides using direct write lithography. Waveguide dimensions for optimal confinement in single-mode transverse electric (TE) polarization at wavelengths of 3 µm and 10.4- 11.3 µm were modeled and the direct lithography process was refined. Through dose testing and adjustments to the raster direction and pixel resolution, it was found that direct write lithography lacked the resolution required for low-loss waveguides. Scanning electron microscopy (SEM) revealed inconsistent waveguide widths and sidewall roughness, and e-beam lithography was identified as the preferred lithography process. For future integration of LT-Ge in a foundry process design kit (PDK), a universal thickness of 1.7 µm was found to support single-mode waveguide operation from 3-11.3 µm wavelength.

Assessing Log-Based Coordination Systems for Managed Cloud Environments

2026-02-13T03:49:14Z

Assessing Log-Based Coordination Systems for Managed Cloud Environments Jimenez, Gabriel The distributed systems landscape is undergoing a significant shift toward managed cloud environments, reducing the prevalence of self-hosted coordination services such as ZooKeeper. While ZooKeeper remains a proven and feature-rich solution for coordination tasks, its deployment in cloud environments can introduce component redundancy. This is because the underlying cloud platform already provides internal mechanisms to ensure coordination guarantees. This thesis investigates the design and evaluates the performance of a log-based coordination service library tailored for managed cloud environments. The proposed library removes the ensemble management overhead inherent in ZooKeeper by delegating durability and consistency responsibilities to the cloud provider’s data layer. This architectural simplification enables a modular design, allowing for tailored implementations that exploit the strengths and mitigate the limitations of a system's specified data layer. The library demonstrated feature parity with ZooKeeper for a targeted subset of coordination features, including leader election, membership tracking, and ephemeral state management. The same is noted for migration from an existing ZooKeeper-based application to this work's library, requiring minimal design changes while preserving coordination guarantees. While the results show that this design does not yet match mature coordination services in raw performance, they highlight potential avenues for further research, particularly in optimizing log-based coordination systems for the unique characteristics of cloud-managed data layers. Given the industry’s steady movement toward cloud-native infrastructure, these findings provide a foundation for future exploration into lightweight, platform-integrated coordination solutions.

Language Comprehension, Production, and Reasoning in Humans and Neural Language Models

2026-02-13T03:10:48Z

Language Comprehension, Production, and Reasoning in Humans and Neural Language Models Eisape, Tiwalayo How closely do neural language models mirror human language processing, and what can this alignment teach us about cognition? This dissertation presents convergent evidence in comprehension, production, and reasoning that neural language models (LMs) can serve as productive instruments for understanding naturalistic human language use at scale. Studies 1-2 examine comprehension with complementary methods. First, Cloze Distillation—a novel method for aligning models with human next-word predictions—improves both language modeling and reading time prediction, demonstrating that LMs and humans make distinct, complementary predictions. Second, new methods for identifying syntactic information in LM hidden states demonstrate that models learn to implicitly represent incremental syntactic state. These probes also enable targeted interventions, allowing us to manipulate representations to resolve (or induce) temporary misinterpretations, confirming mechanistic understanding. While these studies demonstrate prediction’s role in comprehension, a complete account requires examining whether these mechanisms also shape how humans produce language in real-time. Study 3 analyzes a massive corpus of 2.3 million competitive typing events from TypeRacer.com, uncovering the first evidence of in-context predictability effects in this domain of production. Finally, Study 4 compares human and LM reasoning systematically—LMs achieve higher syllogistic reasoning accuracy than humans while still replicating several fine-grained human-like error patterns that are orthogonal to logical accuracy, including premise ordering effects. These converging findings reveal prediction as a fundamental mechanism in comprehension, production, and reasoning in both humans and LMs. While models achieve this through statistical learning rather than specialized cognitive architecture—often outperforming humans yet replicating their systematic biases—this alignment supports predictive processing theories of cognition. This work establishes LMs as scalable cognitive laboratories that can complement traditional experiments, and contributes psycholinguistically principled methods for understanding and controlling LMs.

Proof-of-Work Based Mitigation of Real Time Video DDoS Attacks

2026-02-13T03:49:20Z

Proof-of-Work Based Mitigation of Real Time Video DDoS Attacks Echezona, Chukwuemekalum As the Internet continues to grow in size and complexity, Distributed Denial of Service (DDoS) attacks grow in size and complexity alongside it. One particularly common form of DDoS attack is the TCP SYN flood, which exploits the TCP handshake process to exhaust server resources. This thesis investigates the use of a novel proof-of-work (PoW) based mitigation method to respond to such attacks, specifically in the context of WebRTC video conferencing applications. PoW aims to shift the computational burden from the server to the client, by utilizing a hard to solve puzzle that is easily verifiable. Guided by the same evaluation framework used by the original contributors, we conducted controlled experiments using SPHERE, a national research testbed, and the open-source Jitsi Meet video conference application to simulate DDoS attacks and measure their impact on video quality metrics such as upload/download bitrate and video framerate. Our experiments involved multiple scenarios with and without active attacks and PoW mitigation activate. Results demonstrate that PoW imposes minimal overhead on legitimate clients while maintaining high efficacy when faced with the threat of a SYN Flood attack, regardless of whether the attackers do the proof-of-work before sending traffic. These findings highlight PoW as a promising low overhead mitigation method for WebRTC conference systems under the threat of DDoS attacks.

Optimizing Priority-Based Search for Lifelong Multi-Agent Path Finding

2026-02-13T03:49:13Z

Optimizing Priority-Based Search for Lifelong Multi-Agent Path Finding Huang, Natalie The lifelong Multi-Agent Path Finding (MAPF) problem requires planning collision-free trajectories for agents operating continuously in dynamic environments. Traditional solvers such as Priority-Based Search (PBS) use fixed branching heuristics, which can be inefficient in high-congestion scenarios. This work explores how learning-based methods can improve PBS decision-making. We develop supervised learning (SL) policies trained from high-quality beam search trajectories and reinforcement learning (RL) policies learned directly through simulation, enabling adaptive branching strategies. Evaluations on warehouse-style and Kiva-style maps with varying agent densities show that learned policies can significantly boost throughput in congested warehouse layouts, while identifying scenarios where classical heuristics remain competitive. Our findings provide guidance on solver selection based on environment layout and congestion characteristics.

Learning to Interpret Language Model Diffs

2026-02-13T03:49:23Z

Learning to Interpret Language Model Diffs Goel, Avichal Finetuning-induced changes to a model’s weights (a “model diff”) are semantically meaningful but often difficult to interpret. This makes us wonder: can we describe the content of an unknown model diff using natural language? We introduce diff interpretation training, a method that teaches a model describe its own finetuning-induced modifications. Our approach uses synthetic model diffs to train a lightweight adapter, which in turn can be applied to a compatible finetuned model to make it self-describing. Using two simple task settings, we demonstrate that our method can successfully decode model diffs into accurate natural language descriptions.

Product architectures for solar-powered drip irrigation (SPDI) systems in the Middle East and North Africa

2026-02-13T03:10:44Z

Product architectures for solar-powered drip irrigation (SPDI) systems in the Middle East and North Africa Grant, Fiona R. To feed the growing global population, agriculture production must be intensified using existing land and resources. Sustainable agriculture intensification is particularly important in the Middle East and North Africa (MENA), the most water-stressed region in the world. Solar-powered drip irrigation (SPDI) has the potential to increase water use efficiency and reduce fossil fuel use for irrigation. Despite these benefits, SPDI adoption is limited by its high investment cost and the misalignment between farmers' risk tolerance and broader sustainability goals. Past work has explored three areas of SPDI innovation: low-pressure drip emitters, system cost optimization, and precision irrigation control. This thesis integrates previous innovations in an end-to-end design process to generate SPDI architectures that are accessible to resource-constrained farmers. A market study was conducted to understand farmers' priorities and constraints and articulate SPDI value propositions for the target users. Stakeholder surveys were conducted in Jordan and Morocco for farms ranging from 1–130 hectares. Three market segments were identified, grouping farmers who face similar economic and knowledge barriers. While farmers generally prioritized irrigation reliability and low system costs, the observed variety in farm size, production volume, and technical expertise suggested that SPDI architectures must be tailored to each market segment. This thesis proposes an energetic framework that captures system parametric relationships to identify feasible SPDI design trade-offs. The optimized solar power systems were 14%–80% less expensive than conventionally-sized designs. Despite significant changes to the hydraulic operating parameters, the proposed SPDI architectures were as reliable as existing systems. For farms with long irrigation times, it was optimal to pair low-pressure drip emitters with an irrigation schedule that tracks the daily solar profile, termed “solar profile matching” (SPM), to maximize direct solar power use. The SPM schedule reduced system cost by minimizing the battery capacity. An economic analysis demonstrated that the optimal SPDI designs could be made cost-competitive with grid power through SPDI retrofit subsidies, which some local governments already support. Researchers and industry professionals could use the energetic framework and techno-economic analysis presented in this thesis to inform system design and policy decisions and promote SPDI adoption. Finally, this work created guidelines for designing a precision irrigation controller in resource-constrained markets. A controller was conceptualized to implement the SPDI-SPM architecture. The controller functional requirements and design specifications were iteratively defined with stakeholders, and a prototype was tested on two farms in the MENA region. The controller reduced water and energy use by up to 44% and 43%, respectively, while maintaining crop yield. However, the controller relied on battery power to execute the irrigation schedule. A yield loss sensitivity analysis found that using 72%–79% of the available solar energy on average, an increase of about 40% from the experiment SPM schedules, would have been sufficient to reliably irrigate with solar alone. The results suggest that, with software modifications, the proposed controller could eliminate the need for a battery and enable low-cost SPDI systems. If adopted, the proposed controller could make sustainable irrigation practices more accessible to farmers.

Approximate L² Error Control by Solution Post-Processing for Finite Element Solutions of PDEs with Higher-Order Adaptive Methods

2026-02-13T03:49:17Z

Approximate L² Error Control by Solution Post-Processing for Finite Element Solutions of PDEs with Higher-Order Adaptive Methods Botto Tornielli, Marcos Julian With the substantial computing resources available today, computational fluid dynamics simulations allow scientists and engineers to simulate physical problems very accurately. However, achieving this accuracy requires a sufficiently refined computational mesh, which is a primary driver for the high cost of complex simulations. Mesh adaptation methods provide an automated way to determine the regions where a mesh needs the most refinement and generate a new mesh that efficiently targets these regions. In this thesis, we build on previous work in a posteriori error estimation and mesh adaptation for finite element methods to propose a new mesh adaptation method based on L² error control by solution post-processing. A key feature of our method is its natural extension to higher-order discretizations while providing a problem-independent adaptation methodology. Problem-independent adaptation methods do not depend on specific information about the partial differential equation (PDE) problem being solved, and can therefore be applied to a wide range of problems without modification. We present numerical results applying the approximate L² error control method to a two-dimensional advection-diffusion problem with anisotropic features. These results demonstrate the proposed method’s ability to generate well-adapted anisotropic meshes for solutions with polynomial orders 1, 2, and 3. We also apply the approximate L² error control method to a more complex two-dimensional Reynolds-Averaged Navier-Stokes problem with turbulent flow over a flat plate. We compare the convergence of the drag coefficient and the characteristics of adapted meshes obtained with the proposed method and with an output-based adaptation approach. As expected, the approximate L² error control method is not as effective as the output-based approach in reaching a converged drag coefficient value, but it nevertheless demonstrates the ability to effectively control the approximate L² error in the Mach field.

Advancing Ubiquitous Tactile Sensing through Comprehensive Tooling for Resistive Matrix-Based Sensors

2026-02-13T03:49:25Z

Advancing Ubiquitous Tactile Sensing through Comprehensive Tooling for Resistive Matrix-Based Sensors Murphy, Devin Resistive matrix-based tactile sensors offer a scalable and intuitive approach to capturing human-environment interactions, yet deploying them in real-world systems remains challenging because they must remain portable, adaptive, and long-lasting. This thesis presents the WiReSens Toolkit, an open-source hardware and software platform for developing resistive tactile sensing systems that meet the demands of real world applications. The toolkit features adaptive hardware for interfacing with resistive sensors and a web-based GUI that mediates access to otherwise complex functionality, including 1) multi-device programming and wireless visualization across three distinct communication protocols 2) autocalibration methods for adaptive sensitivity and 3) intermittent data transmission for low-power operation. As a use case for the toolkit, the thesis then introduces a method for the automatic design and fabrication of custom tactile sensing gloves using flexible printed circuit boards (FPCBs), enabling rapid, scalable production. Together, these contributions lower barriers to adoption and support broader exploration of tactile sensing in HCI, robotics, and ubiquitous computing.

Topics in Geometric Machine Learning

2026-02-13T03:10:46Z

Topics in Geometric Machine Learning Tahmasebi, Behrooz Recent advances and the widespread adoption of neural networks have revolutionized machine learning and artificial intelligence. These developments demand learning paradigms capable of processing data from diverse applications and sources. In structured domains such as molecules, graphs, sets, and 3D objects, as well as fields such as drug discovery, materials science, and astronomy, models must account for data structures. The emerging field of geometric machine learning has gained attention for enabling neural networks to handle geometric structures, unlocking novel solutions across scientific disciplines. Despite recent advances, theoretical gaps remain. This thesis aims to address these gaps by studying the benefits and limitations of leveraging geometric structures and symmetries in data. We explore sample complexity, generalization bounds, hypothesis testing for the presence of symmetries in data, time complexity of learning under symmetries, and regularization and optimization in symmetric settings. The goal is to build a robust theoretical framework that validates recent successes and sheds light on unexplored aspects, fostering future progress in geometric machine learning.

Quantization Methods for Matrix Multiplication and Efficient Transformers

2026-02-13T03:49:20Z

Quantization Methods for Matrix Multiplication and Efficient Transformers Savkin, Semyon We study quantization in Machine Learning. First, we introduce NestQuant — a technique for quantization of matrix products and post-training quantization of LLMs. Beyond reducing the memory footprint, quantization accelerates inference, as the primary bottleneck during autoregressive generation is often the memory bandwidth. NestQuant leverages two nested lattices to construct an efficient vector codebook for quantization, along with practical encoding and decoding algorithms. The approach is grounded in recent theoretical work that characterizes the optimal rate–distortion trade-off for matrix products. Empirically, on Llama-3-8B, it reduces the perplexity gap between full-precision and quantized models by more than 55% relative to the current state-of-the-art technique (SpinQuant). Second, we investigate data-domain quantization for RF signals. We propose a tokenized transformer for source separation that discretizes RF waveforms into learned tokens and operates directly on the resulting sequences, outperforming strong convolutional baselines. Together, these contributions connect information-theoretic limits with deployable systems: structured vector quantizers accelerate LLM inference and enable competitive discrete representations for RF tasks.

Improving Data-Driven Contact Localization and Force Estimation for Barometric Tactile Sensors

2026-02-13T03:49:15Z

Improving Data-Driven Contact Localization and Force Estimation for Barometric Tactile Sensors Chun, Ethan Barometric tactile sensors offer a cheap, robust, and customizable means for robots to perceive the world. Central to their operation are models that extract useful information from the sensors’ raw pressure readings. In this work, I focus on improving data-driven methods for single-point contact localization and force estimation using a previously presented three-quarter sphere barometric tactile sensor. To allow modeling of time-dependent effects in the sensor material, I introduce a multi-threaded data collection system that captures ground truth contact and sensor data at exactly 100 Hz. I construct both feed-forward and recurrent networks using this data, finding that a recurrent network achieves a 15% lower mean absolute error for angular contact localization on the sphere compared to prior methods. The recurrent architecture’s computational efficiency ensures that the architecture can still run within the constraints of the sensors’ microcontroller. Despite this improvement, I find that more expressive models such as LSTMs tend to overfit on the collected data and physical phenomena observed during deployment were not well represented by the training metrics. To better understand the extent that these data-driven methods alone can improve sensor performance, I shift focus away from the modeling and analyze the physical sensor instead. I find that viscous effects in the sensor can render the prediction task unlearnable without historical data and that thermal effects introduce a train-test distribution shift. Finally, I discuss design criteria for a theoretical future barometric tactile sensor that may mitigate the effects found during my modeling and analysis.

Probabilistic Programming over Heterogeneous Language and Hardware Targets

2026-02-13T03:49:09Z

Probabilistic Programming over Heterogeneous Language and Hardware Targets Rojas Collins, Elias G. Modern probabilistic programming applications, from large-scale Bayesian inference to real-time decision making, require both the expressiveness of CPU-oriented languages such as Gen.jl and the massive parallelism of GPU-backed array languages such as GenJAX, yet existing platforms force users to trade modeling flexibility for performance. This thesis introduces GenUflect, a metalanguage that embeds multiple Gen-compatible dialects inside a single program, allowing each sub-component to run on the most appropriate language and hardware target while preserving Gen’s programmable-inference interface. GenUflect extends Gen’s dynamic-modeling language with the @union, @vmap, @amortize, @amortize≤, and @runtime_union combinators; these macros compile at build-time (or justin-time) to autonomous generative functions written in the target dialect, link them through a lightweight FFI layer, and manage cross-device data via zero-copy MirrorArrays and lazily materialized traces. The resulting programs remain sound by construction because each foreign subtrace is itself a valid Gen generative function. Empirical studies demonstrate that this hybrid approach yields large practical gains. On a split linear-vs-sinusoidal regression task, GenUflect matches pure GenJAX throughput while running higher-order control logic on the CPU, and is up to two orders of magnitude faster than a pure Gen implementation for datasets of 105 points. In a collapsed-Gibbs sampler for a Dirichlet-process mixture model, GenUflect’s elastic allocation (@amortize≤) lets vectorized GPU kernels adapt to a growing number of clusters; the same inference that takes over an hour in Gen executes in seconds with GenUflect. A probabilistic inverse-graphics pipeline further showcases how heterogeneous submodels can cooperate seamlessly within unified inference code. By coupling language interoperability with automated data movement and compile-time code generation, GenUflect bridges the gap between flexibility and speed, enabling scalable, expressive probabilistic programs that natively exploit both CPUs and accelerators.

Under-Coverage of Double Machine Learning Due to Implementation Choices

2026-02-13T03:49:08Z

Under-Coverage of Double Machine Learning Due to Implementation Choices Siegmann, Charlotte B. Double ML estimators can estimate coefficients of interest with far fewer functional form assumptions than linear econometric methods. However, DML requires researchers to make a range of implementation choices, including the selection of the function class, the random seed, and hyperparameter configurations. While asymptotic theory suggests these choices should not affect final estimates, we show that for 10 economic analyses (8 of them published and peer-reviewed), implementation choices affect the results. In half of the datasets, different implementation choices even change the interpretation of findings between negative, null, or positive effects. We link these results to a framework for empirically assessing the performance of machine-learning-based estimators, focusing on precision, coverage, and susceptibility to manipulation. This is meant to complement asymptotic theory. We demonstrate that the coverage of DML confidence intervals is too low—placing an upper bound of 48% on the expected coverage of conventional 95% confidence intervals for published DML economics papers. We show that in the status quo, the susceptibility of DML to manipulation by researchers is high, but propose ways to mitigate this susceptibility.

Optimizing Irreversible Perturbations of the Unadjusted Langevin Algorithm

2026-02-13T03:49:05Z

Optimizing Irreversible Perturbations of the Unadjusted Langevin Algorithm Zhu, Qianyu Julie A central task in Bayesian inference and scientific computing is to compute expectations with respect to probability distributions that are only known up to a normalizing constant. Markov chain Monte Carlo (MCMC) methods, and in particular Langevin dynamics, provide a powerful framework for this task by constructing stochastic processes that converge to the target distribution. However, practical implementations face two challenges: slow mixing when the target distribution is anisotropic or multimodal, and persistent discretization bias introduced by numerical schemes. This thesis investigates irreversible perturbations of overdamped Langevin dynamics, aiming to accelerate mixing while controlling discretization error. Irreversible perturbations introduce skew-symmetric drift terms that preserve the target distribution while inducing rotational flow, thereby enhancing exploration. Although prior work has established their benefits in continuous-time settings, the impact of discretization and the design of optimal perturbations for discrete-time algorithms remain open problems. We develop a framework for optimizing constant (position-independent) irreversible perturbations in the Unadjusted Langevin Algorithm (ULA). Our approach balances two competing objectives: maximizing the spectral gap of the continuous dynamics to accelerate convergence, and minimizing discretization error that drives estimation bias. Motivated by this, we introduce new criteria that jointly evaluate bias and efficiency, and we show how these criteria identify perturbations that improve performance beyond existing constructions. Theoretical analysis is complemented by numerical experiments on Gaussian and nonGaussian targets. These experiments demonstrate that appropriately designed irreversible perturbations can reduce mean-squared error without sacrificing stability, while poorly chosen perturbations can degrade performance. The results highlight the importance of geometry-aware design and motivate systematic optimization strategies for irreversible perturbations. Overall, this work extends the theoretical and practical understanding of irreversible Langevin dynamics, bridging the gap between continuous-time spectral analysis and discrete-time numerical performance. It provides principled tools for constructing efficient MCMC samplers, with potential applications in high-dimensional Bayesian inference and modern machine learning.

Single Camera Motion Compensated Viewpoint Shift

2026-02-13T03:49:03Z

Single Camera Motion Compensated Viewpoint Shift Snowdon, Adam Eye contact is a necessary tool for human connection and in most video conferencing situations, eye contact is not possible. Standard laptop and webcam configurations position the camera at the top of the screen, meaning that when the user looks at other people’s faces in the center of the screen, the camera captures the user looking downward, creating the impression of poor eye contact for remote participants. Solutions involving 3D modeling of the face to synthesize a gaze-corrected view have been explored and exist but are too computationally costly for most personal computers. To address this computational challenge, we draw inspiration from 2D frame interpolation techniques to synthesize a virtual camera view that repositions the user’s apparent gaze toward the camera. Our method uses a single camera located at the top of the user’s screen and requires only a brief setup period. Assuming there is only one user, our approach creates a virtual camera view that transforms the user’s viewpoint from the screen center to the camera position, enabling more realistic eye contact in video conference calls.

A Framework for 3D Mouse Brain Reconstruction: RNA-based Stitching of Adjacent Tissue Slices and Co-Registration of Multimodal Imaging Data

2026-02-13T03:48:56Z

A Framework for 3D Mouse Brain Reconstruction: RNA-based Stitching of Adjacent Tissue Slices and Co-Registration of Multimodal Imaging Data Pan, Jessica N. Mapping the brain’s complex neural networks requires tracing the long-distance pathways of individual axons, a task that demands a comprehensive 3D reconstruction of the brain. Recently, spatially resolved transcriptomics (SRT) methods enable the study of gene expression and biomolecule distribution in each neuron in its spatial context, opening the door to more thoroughly investigating cell-cell interactions between neurons. However, SRT methods are limited to slices of tissue; therefore, computational alignment is essential to reconstruct a cohesive 3D volume while correcting for both batch effects and inherent sample variability. This thesis presents a novel framework that addresses these challenges through three primary contributions. First, a memory-efficient, non-referenced-based algorithm was developed to align the superficial surfaces of adjacent, high-resolution tissue slices. Second, these surface transformations were interpolated through the tissue slices on a proof-of-concept dataset of three adjacent slices. Third, methods for co-transforming fluorescent protein imaging data were explored to fully resolve the cell boundaries between neurons. These three methods are necessary steps towards creating a fully-resolved, multimodal 3D model of the brain.

Optimizing Non-Convex Objectives to Plan More Optimal Motion for Manipulators

2026-02-13T03:49:11Z

Optimizing Non-Convex Objectives to Plan More Optimal Motion for Manipulators Garg, Shruti Non-convex optimization is essential to tackle increasingly complex and practical problems in kinematic motion planning. Although introducing non-convexity often sacrifices guarantees of feasibility and optimality–making solutions more susceptible to local minima or failure to converge–many robotic systems and tasks are non-convex by nature, necessitating at least somewhat non-convex formulations. In this thesis, we aim to mostly constrain non-convexity to the objective. This optimization structure helps preserve certain feasibility guarantees in theory and usability in practice while enhancing optimality of solutions, even if global optimality is not achieved. In the first chapter, we demonstrate the effectiveness of non-convex objectives in scenarios where motion planning involves a non-convex parameterization of the configuration space. We keep constraints strictly convex, with the non-convexity quarantined to the objective. This structure guarantees a feasible solution given a feasible initial guess. We primarily use our method to post-process Graphs of Convex Sets solutions in three domains: constrained bimanual motion, motion with guaranteed non-collision, and planning in SO(3). In each case, the non-convex objective compensates for distortion introduced by the parameterization, resulting in more efficient and natural motion. In the second chapter, we propose teleoperation scheme with full-body motion planning for non-holonomic mobile manipulators. Our key contribution is a Differential Inverse Kinematics (DiffIK) formulation that crafts non-convex objectives to avoid singularities and joint limits leading to more robust feasible motion. Unlike before, the constraints are not strictly convex, so the optimization has no guarantees of feasibility. However, we mitigate the non-convexity in the constraints as much as we can by linearizing around the robot’s current position and approximating the highly non-convex non-holonomic constraint. We explore multiple formulations for singularity avoidance and empirically demonstrate that integrating these objectives into DiffIK improves motion quality for teleoperation for the RBY-1 robot.

CableSplat: Optimized 3D Gaussian Splatting for 1D Deformable Pose Estimation

2026-02-13T03:49:02Z

CableSplat: Optimized 3D Gaussian Splatting for 1D Deformable Pose Estimation Pai, Sameer A key challenge in the robotic manipulation of deformable objects is the lack of accurate and efficient systems for estimating their pose in real-time, especially in the presence of occlusion. In this thesis we propose CableSplat, a novel non-parametric method leveraging 3D Gaussian Splatting to estimate the pose of a linear deformable object given RGB images of the object from multiple viewpoints. To facilitate the evaluation of the performance of this method, we develop both simulated and real-world pipelines to collect calibrated and segmented recordings of cables undergoing various manipulations and transformations. We find that our method is consistently able to estimate cable pose to within an average error of ∼2.5mm across simulated tasks. Furthermore, performance on a scene reconstruction metric drops only slightly between simulated and real-world data, suggesting high-fidelity state estimation even in the real world. CableSplat is therefore a promising candidate for the extension of existing manipulation systems to deformables.

scPhen: Single-Cell Phenotype Predictor for Alzheimer’s Disease

2026-02-13T03:49:10Z

scPhen: Single-Cell Phenotype Predictor for Alzheimer’s Disease Guo, Sophie J. Advances in artificial intelligence (AI) and generative AI for representation learning have transformed our ability to model complex biological systems. Single-cell RNA sequencing (scRNA-seq) provides unprecedented resolution into cellular heterogeneity, offering a powerful substrate for modeling disease circuitry. However, predicting patient-level phenotypes from scRNA-seq remains challenging due to limited sample sizes, variable cell counts, and the computational burden of modeling long-context dependencies. We present scPhen, a flexible, parametric deep-learning framework for phenotype prediction from single-cell transcriptomic data, applied here to Alzheimer’s disease (AD) as a paradigm of complex, heterogeneous pathology. scPhen consists of a cell embedding module and a patient embedding module, designed to capture both fine-grained molecular patterns and higher-order cell–cell relationships. The framework supports multiple architectural backbones, including Transformers, Graph Neural Networks (GNNs), and state-space models such as Mamba, Mamba2, and BiMamba2, allowing exploration of tunable components for optimized performance. Across classification and regression tasks, state-space models, and in particular BiMamba2, demonstrated superior predictive accuracy and computational efficiency compared to Transformer-based and hybrid approaches. We further integrated attention-based multiple instance learning to enable variable cell counts per patient and to prioritize phenotype-informative cellular subsets. Interpretability analyses using Integrated Gradients and cell-level attention scores revealed gene programs and cell populations associated with AD progression, highlighting known neuroinflammatory signatures and suggesting novel molecular targets. By unifying cutting-edge sequence modeling architectures with scalable single-cell analysis, scPhen provides a generalizable, high-resolution approach to phenotype prediction. While demonstrated here in AD, this framework is readily extensible to other complex diseases and multi-modal cellular datasets, bridging computational innovation and biological discovery.

Predicting Task Functional Localizers Using Naturalistic fMRI

2026-02-13T03:49:09Z

Predicting Task Functional Localizers Using Naturalistic fMRI Wilke, Jordan Functional magnetic resonance imaging (fMRI) data collected during naturalistic stimuli has shown promise for predicting individual traits, biomarkers of disease and functional brain localizations, potentially offering advantages over traditional resting-state approaches. This study investigated the use of interpretable deep learning models to predict demographics and functional task localizer activations from fMRI time-series data collected while participants viewed naturalistic stimuli. Using the data of 143 subjects from the Human Connectome Project, I analyzed 7T fMRI scans from participants watching movies to predict sex, age, and functional localizer activations across multiple cognitive tasks. I employed state-of-the-art machine learning architectures, including DICE and Glacier models, specifically chosen for their interpretable design features that build directed connectivity matrices and produce weighted temporal attention maps. These models aimed to capture dynamic brain activity patterns while maintaining the ability to understand which temporal features drive predictions. The results successfully reproduced previous findings for sex classification but showed poor performance for age prediction, with correlations ranging from -0.175 to 0.243. For functional localizer predictions, models initially appeared to achieve high performance with some specific contrasts having correlations around 0.9 and Dice scores generally above 0.6. However, detailed analysis revealed that these models were primarily predicting group averages rather than learning meaningful inter-subject variability, as evidenced by chance-level subject identification accuracy. This finding contrasts with previous works that demonstrated successful prediction of individual differences in functional localizations. The failure to capture inter-subject variability represents a significant limitation, as individual differences in functional regions of interest are crucial for applications such as pre-surgical mapping and disease prediction. My findings suggest that predicting from raw fMRI time-series may require different approaches than those used here, with preprocessed functional connectivity matrices showing promising results, and highlight the importance of sufficient training data to separate signal from noise when learning directly from naturalistic stimuli. Despite these challenges, this work establishes important methodological foundations and identifies key limitations that must be addressed in future research combining naturalistic stimuli with machine learning for fMRI prediction tasks. The findings emphasize the need for models that can capture individual functional differences while maintaining the interpretability necessary for understanding how naturalistic stimuli drive brain-based predictions.

Probabilistic Programming with Low-Level, High-Performance GPU Programmable Inference

2026-02-13T03:49:04Z

Probabilistic Programming with Low-Level, High-Performance GPU Programmable Inference Chung, Karen GPU-compatible probabilistic programming languages (PPLs) have enabled high-performance, data-parallel programmable inference. However, these systems face fundamental trade-offs between expressiveness and performance, as their GPU code generation is automated and black-boxed, limiting optimization opportunities and imposing restrictions on program expressivity. This thesis introduces GenCUDA, a probabilistic programming system that addresses this limitation by embedding the CUDA GPU programming language directly into a C++/CUDA frontend, enabling GPU programmable inference with fine-grained control over runtime and memory profiles. GenCUDA extends the Gen probabilistic programming architecture by providing a dynamic modeling language (DML) that allows users to write performance-critical sections of generative functions as CUDA kernels while maintaining automatic trace management and the generative function interface (GFI). The system supports both sequential and parallel execution contexts through specialized effect handlers that seamlessly compose CPU and GPU code paths. Key technical contributions include: (1) a high-performance GPU distributions library achieving 10-100× speedups over TensorFlow-Probability, (2) memory-efficient trace management via template-optimized parallel effect handlers, and (3) vectorized generative functions that enable massive parallelization of inference algorithms. We demonstrate GenCUDA’s capabilities through comprehensive benchmarks on inference algorithms applied to diverse models including factor graphs, mixture models, and Hidden Markov Models. Results show significant performance improvements over JAX-based implementations: up to 3× speedup for importance sampling on a hierarchical model, 5.7× speedup for parallel Gibbs sampling on factor graphs, and memory efficiency improvements for large-scale mixture models supporting up to 6× as many clusters compared to existing frameworks’ limits. The system maintains the composability and expressiveness of probabilistic programming while unlocking GPU performance optimization techniques such as kernel fusion and memory hierarchy exploitation that are inaccessible to higher-level frameworks. GenCUDA demonstrates that embedding low-level GPU programming within automated probabilistic inference workflows can achieve both performance gains and algorithmic expressivity without sacrificing the modularity of probabilistic programming paradigms.

Simplifying Equivariant GPU Kernels through Tile-based Programming

2026-02-13T03:49:00Z

Simplifying Equivariant GPU Kernels through Tile-based Programming Kotak, Mit E(3)-equivariant neural networks have demonstrated success across a wide range of 3D modeling tasks. Until recently, they were bottlenecked due to their high memory and wall-time requirements. In this thesis we first provide an overview of recent GPU kernel efforts by both academia and industry that address this issue. These approaches tradeoff performance for engineering complexity, while still being algorithmically bottlenecked at 10 % GPU utilization. We instead trade off engineering complexity for performance. This not only lowers the barrier to GPU programming but also builds an abstraction layer to reason about future algorithmic innovations that can improve GPU utilization. Our kernel 𝐵3, based on the tiling- optimizations in just 100 lines of PyTorch-like code. We explore the performance-simplicity tradeoff with two case studies and demonstrate the practicality of our kernel workflow through downstream integration with a production model. We hope this work serves as inspiration to broaden and deepen existing equivariant kernel efforts.

Chemical exposures in drinking water: contaminant analysis and physicochemical behavior

2026-02-13T03:10:41Z

Chemical exposures in drinking water: contaminant analysis and physicochemical behavior Bugher, Nicolette A. Environmental chemical exposures pose an understudied risk to human health. The quality and accessibility of data on occurrence in the environment and physicochemical behavior of industrial chemicals are integral for accurate exposure risk assessment. In this dissertation, analytical chemistry techniques were developed and leveraged to characterize human exposures to contaminants in drinking water and improve methods for assessing such risks. The occurrence of organic industrial pollutants in domestic well waters was investigated, with a particular focus on the impacts of region-specific industrial activity (e.g., hydraulic fracturing), legacy pollution sites (e.g., Superfund sites), and geochemistry. The exposure risk to water contaminants of domestic well users was further interrogated by evaluating trends in contaminant concentrations resulting from the implementation and maintenance of in-home water treatment devices. The results show widespread, low-dose mixtures of organic pollutants, where the efficacy of removal by in-home water treatment varied by water contaminant class and maintenance frequency. Additionally, analytical methods were optimized to quantify a group of organic water contaminants (i.e., probable carcinogens, N-nitrosamines), improving method sensitivity and critically identifying false-positive interferences. Finally, methods were evaluated and deployed for the determination of physicochemical properties of N-nitrosamines. The results of which demonstrate gaps in existing experimental data, provide a valuable methodological intercomparison (two experimental and two computational approaches), and contribute novel partitioning data. This dissertation addresses gaps in occurrence data, analytical method sensitivity, and reliability of physicochemical parameters for risk assessment. The combination of method development and implementation enables the study of exposures to water contaminant mixtures at health-relevant concentrations, representative of prevalent exposure pathways.

From coarse fate choice to precise pattern: post-mitotic progenitor targeting

2026-02-13T03:49:07Z

From coarse fate choice to precise pattern: post-mitotic progenitor targeting Nie, Mel F. Planarians possess remarkable regenerative abilities, driven by pluripotent stem cells called neoblasts. While neoblasts are known to give rise to progenitor cells that form various tissues, whether and the extent to which these progenitors migrate across the animal remains unclear. Irradiation experiments eliminate all neoblasts outside shielded areas, allowing for the visualization of cell migration from the remaining neoblasts, but irradiated animals may not reflect homeostatic progenitor migration patterns. To address this, 5-ethynyl-2’-deoxyuridine (EdU) labeling and plug transplant techniques were used to trace progenitor movement in non-irradiated planarians. Using whole-mount fluorescence in situ hybridization (FISH) and the quantification of EdU-labeled cells, this study demonstrates that progenitor cells are capable of migrating long distances and exhibit a pronounced anterior bias in their movement and integration.

Optimizing Data Layouts for Evolving Cloud Table Storage

2026-02-13T03:10:39Z

Optimizing Data Layouts for Evolving Cloud Table Storage Sudhir, Sivaprasad Modern data analytics platforms increasingly adopt disaggregated architectures, storing data in cost-effective cloud object stores. While this approach enables a clean separation of concerns, allowing each layer to be independently managed and scaled, it introduces significant performance bottlenecks due to expensive data movement. Effective data layouts, which organize data to minimize unnecessary data reads, are thus critical to achieving high query performance. However, existing techniques typically rely on manually specified layouts, collect limited metadata, or lack mechanisms to dynamically adapt to changing data and workloads. This thesis investigates adaptive, metadata-rich, expressive data layouts for cloud table storage. First, we introduce Pando, a correlation-aware layout technique that leverages rich metadata on query predicates to significantly improve data skipping. Next, we propose CopyRight, a partial replication strategy that selectively replicates subsets of data and optimizes each replica differently, efficiently serving heterogeneous query patterns. Finally, we describe Self-Organizing Data Containers (SDCs), a practical table storage layer for the cloud that incrementally reorganizes complex data layouts based on changes in data and workload distributions.

Development of Ensemble Strategies for Generalization in Deepfake Image Detection

2026-02-13T03:49:01Z

Development of Ensemble Strategies for Generalization in Deepfake Image Detection Wagh, Rohan M. The growing accessibility of generative models has enabled the rapid proliferation of deepfake content, posing significant challenges in image-based biometric security and media authenticity. In this thesis, six diverse facial deepfake image datasets are assembled, and four modern detection models are evaluated in a cross-domain scenario. We observe that individual models fail to generalize to images generated by techniques outside the scope of their training data. This often hinders the applicability of a single model in real-world deepfake detection. This thesis proposes ensemble strategies as a means of addressing this lack of generalization. We find that the ensemble models outperform individual models in classifying deepfake images, particularly in terms of accuracy and recall. An exhaustive evaluation of combinations of models shows that ensembles of similar models provide limited benefit, whereas ensembles of complementary models lead to significant improvements in classification performance. Ensembling models based specifically on accuracy and recall metrics also produces models that lower the rate of more harmful false negative predictions. This work highlights the value of ensemble models in improving generalization across diverse image families and provides a framework for building robustness in real-world deepfake detection systems.

Quantum information science and underground facilities

2026-03-08T03:40:02Z

Quantum information science and underground facilities Formaggio, Joseph A As both nuclear physics and particle physics involve the quantum interactions of many sub-atomic particles, there has always existed a strong interplay between these fields and the study of quantum physics and quantum information systems (QIS). This interplay has accelerated in recent years, particularly with the emergence of new, highly sensitive technologies, nascent access to quantum computing environments at the O(10)-O(100)-bit scale, and the use of coherence and entanglement to enhance sensitivity to novel and exotic phenomena. One unusual area of interplay between the two disciplines that has recently emerged is the role of background radiation and background mitigation on highly sensitive systems such as qubits.

Reduced-order model to predict thermal conductivity of dimensionally confined materials

2026-03-08T03:40:08Z

Reduced-order model to predict thermal conductivity of dimensionally confined materials Hosseini, S Aria; Greaney, Alex; Romano, Giuseppe Predicting nanoscale thermal transport in dielectrics requires models, such as the Boltzmann transport equation (BTE), that account for phonon boundary scattering in structures with complex geometries. Although the BTE has been validated against several key experiments, its computational expense limits its applicability. Here, we demonstrate the use of an analytic reduced-order model for predicting the thermal conductivity in dimensionally confined materials, i.e., monolithic and porous thin films, and rectangular and cylindrical nanowires. The approach uses the recently developed “Ballistic Correction Model,” which accounts for materials' full distribution of phonon mean-free-paths. The model is validated against BTE simulations for a selection of base materials, obtaining excellent agreement. By furnishing a precise yet easy-to-use prediction of thermal transport in nanostructures, our work strives to accelerate the identification of materials for energy-conversion and thermal-management applications.

Parametric decay instabilities driven by high power helicon waves in DIII-D

2026-03-08T03:40:07Z

Parametric decay instabilities driven by high power helicon waves in DIII-D Porkolab, M; Pinsker, RI; DeGrandchamp, GH; Baek, SG; Compernolle, B Van; Denk, S; Petty, CC; Tang, SX; Thome, KE High power helicon waves (whistler or very high harmonic fast lower hybrid waves) at a frequency of 476 MHz are being tested for efficient off-axis current drive on DIII-D with the goal of demonstrating profile control in AT plasmas [1-4]. In agreement with earlier theoretical predictions, strong Parametric Decay Instability (PDI) has been observed at injected RF power levels in the range of 0.05-0.5 MW with corresponding electric fields of 10-30 kV/m [5,6]. The dominant driver of the PDI is the E×B and the polarization drift velocity which can drive ion cyclotron quasi-modes and lower hybrid (or IBW) sideband waves unstable [5,6]. Initial experimental results have been obtained with powers up to 0.3 MW showing evidence of strong PDI measured with high-frequency one-turn magnetic probes located at both the outboard and the inboard wall at frequencies set by the usual selection rules [7,8]. Here we review the appropriate analytic formulation to predict such instabilities and present numerical evaluation of frequencies and growth rates relevant to DIII-D plasma parameters. We also assess the convective thresholds for the PDIs, and compared them with experimental observations.

Experimental footprints of a water-rich depletion layer in the Herschel–Bulkley pipe flow of solidifying polyelectrolytes

2026-03-08T03:40:12Z

Experimental footprints of a water-rich depletion layer in the Herschel–Bulkley pipe flow of solidifying polyelectrolytes Nazari, B.; Moghimi, E.; Bousfield, D. W. A fundamental understanding of the transition from fluid-like to gel-like behavior is critical for a range of applications including personal care, pharmaceuticals, food products, batteries, painting, biomaterials, and concrete. The pipe flow behavior of a Herschel–Bulkley fluid is examined by a combination of rheology, ultrasound imaging velocimetry, and pressure measurements together with modeling. The system is a solution of 0.50 wt. % polyelectrolytes of sulfated polysaccharides in water that solidifies on cooling. Fluids with different ionic strengths were pumped at various rates from a reservoir at 80 °C into a pipe submerged in a bath maintained at 20 °C. The fluid velocity, pressure drop ΔP, and temperature were monitored. The same quantities were extracted by solving continuity, energy, and momentum equations. Moreover, the modeling results demonstrate that the local pressure gradient along the pipe dPdx|x is related to the local yield stress near the pipe wall τywall|x, which explains the variations of dPdx|x along the pipe. Experimental results show much lower values for ΔP compared to those from modeling. This discrepancy is exacerbated at higher ionic strengths and smaller flow rates, where fluid shows a higher degree of solidification. The tabulated experimental ΔP data against the solidification onset length Lonset (where the fluid is cool enough to solidify) along with the ultrasound imaging velocimetry associate these discrepancies between experiments and models to a depletion layer of ∼1 μm, reflecting the lubrication effects caused by the water layer at the wall.

Radiation pressure of radio frequency waves on turbulence in edge plasmas

2026-03-08T03:40:06Z

Radiation pressure of radio frequency waves on turbulence in edge plasmas Ram, Abhay K; Hizanidis, Kyriakos The scattering of radio frequency (RF) waves – lower hybrid and helicon waves – by a single cylindrical filament, embedded in a background plasma, is studied using a full-wave analytical theory. While a filament can affect the propagation of RF waves, the radiation force exerted by the waves can influence the filament. The force on a filament is determined using the Maxwell stress tensor. The radiation force can either pull the filament towards the RF source or push it away. The radiation force, in the two frequency ranges, is large enough to impact the motion of a filament and could be measured experimentally. Consequently, it may be possible to modify the edge turbulence using RF waves.

Replies to Moran, Gallois, and Bar-On and Johnson

2026-03-08T03:40:05Z

Replies to Moran, Gallois, and Bar-On and Johnson Byrne, Alex I am very grateful to Dorit Bar-On, Drew Johnson, André Gallois, and Dick Moran for their thoughtful commentaries. Bar-On, Gallois, and Moran are discussed extensively in Transparency and Self-Knowledge (hereafter T&SK), and their work has been an important source of inspiration for my own. In order to make my contribution to this symposium reasonably compact, I have not attempted to reply to every single point. (One especially notable omission is the alternative account of self-knowledge sketched by Bar-On and Johnson.) Instead, I have concentrated on the main objections.

The end of MAD? Technological innovation and the future of nuclear retaliatory capabilities

2026-03-08T03:40:10Z

The end of MAD? Technological innovation and the future of nuclear retaliatory capabilities Glaser, Charles L. This article motivates the special issue, explaining the new debate over whether emerging technologies – including small satellites, machine learning, cyber weapons, and quantum technologies – will enable major powers to undermine each others’ nuclear retaliatory capabilities. The first article analyzes key relevant emerging technologies. Following articles explore how emerging technologies will influence the vulnerability of mobile missiles, ballistic missile submarines, and nuclear command-and-control; and the ability of missile defenses against intercontinental range missiles. The final article explores China’s views on the requirements of nuclear deterrence. Overall, the articles suggest that U.S. prospects for achieving a damage-limitation capability are poor and declining.

Prototyping longevity services: Tech-driven or human-assisted service?

2026-03-08T03:40:05Z

Prototyping longevity services: Tech-driven or human-assisted service? Lee, Sheng-Hung; Coughlin, Joseph F; Yang, Maria The study investigates the design of longevity services through an experimental comparison of tech-driven and human-assisted service encounters, focusing on six key features: learnability, efficiency, safety, trustworthiness, confidence, and satisfaction. The controlled experiment, which involved 12 gender-balanced participants from Boston, USA, employed four qualitative methods, including surveys, the Think-aloud technique, semi-structured interviews, and transcript analysis supported by computer-assisted qualitative data analysis software (CAQDAS) and its AI-empowered coding function. The study concluded with two insights: 1. Tech-driven services can improve safety, trust, confidence, and satisfaction; and 2. both service encounters are context-sensitive, shaped by participants’ demographics, personality, culture, and environmental factors. Although the small sample size limits the study’s generalizability, the participants’ stories and perceptions offered valuable insights into their implicit needs and subtle behaviors in learning, experiencing, and addressing sensitive, private, and vulnerable topics related to longevity planning.

The Iraq Petroleum Company’s Infrastructure of “Desert Control” during the British Mandate in the Middle East

2026-03-08T03:40:01Z

The Iraq Petroleum Company’s Infrastructure of “Desert Control” during the British Mandate in the Middle East Freeman, Margaret This article discusses the infrastructure of the Iraq Petroleum Company (IPC) in the interwar British Mandatory Middle East as belonging to a larger British imperial project for “desert control” through architecture. Britain’s so-called “desert control” was, more accurately, a programme for control over the pastoralist Bedouin tribespeople who were the primary inhabitants of the Mandatory territories’ desert zones. This article identifies the two pillars of Britain’s “desert control” strategy: the use of Bedouin police forces, and the architectural annexation and restriction of water resources from Bedouin tribes. It argues that Mandate Britain’s “desert control” programme was replicated and adapted by the IPC for its own needs to protect its commercial infrastructural investment, the Iraq–Mediterranean Pipeline, in the British Mandatory territories. It compares two building typologies, the Mandate’s “desert outposts” and the IPC pipeline’s pumping stations, as sites where the Bedouin were alternately welcomed into and excluded from imperial and commercial projects in the interest of controlling them.

Surrogate modelling of surface roughness for asphalt pavements using artificial neural networks: a mechanistic-empirical approach

2026-03-08T03:39:59Z

Surrogate modelling of surface roughness for asphalt pavements using artificial neural networks: a mechanistic-empirical approach Li, Haoran; AzariJafari, Hessam; Kirchain, Randolph; Santos, João; Khazanovich, Lev Pavement surface smoothness (or roughness) is crucial for traffic safety, driving comfort, and fuel efficiency. As a widely applied roughness indicator, an accurate forecasting of the International Roughness Index (IRI) and its deterioration is essential for the design, maintenance, and management of asphalt pavements. Previous studies have used field measurement data or AASHTOWare Pavement ME Design simulations for the development of machine learning (ML) models to streamline the IRI modelling. However, these models frequently lack the accuracy and robustness of the measurement data or high-fidelity computational simulations they are intended to surrogate. To address this issue, we employed a new adaptive sampling technique to generate an informative yet efficient pavement damage database from Pavement ME simulations. Utilising Artificial Neural Networks (ANNs), we engineered two types of surrogate ML models: (a) Model I, an ANN-based IRI predictive model, and (b) Model II, a hybrid model combining ANN-based predictions of rutting, fatigue damage, and thermal cracking with closed-form relationships between these indicators and IRI. Our findings show that Model II outperforms Model I in IRI modelling accuracy both globally and locally. Moreover, Model II matches IRI simulations of Pavement ME while providing enhanced efficiency and adaptability to a broader spectrum of design considerations.

Honoring practices of community-based educators: lessons learned from the collaborative design of a creative mobile app

2026-03-08T03:39:52Z

Honoring practices of community-based educators: lessons learned from the collaborative design of a creative mobile app Rusk, Natalie; Jain, Rupal; Martin, Caitlin K.; Roque, Ricarose; Freitas, João Adriano; Molaodi, Linford This paper shares reflections and stories from a collaborative design process between the Lifelong Kindergarten group at the MIT Media Lab and a global network of community-based educators to develop a creative coding app called OctoStudio, which supports children and families to create and share interactive projects on mobile devices. The app design is grounded in practices that community-based educators who are primarily from the Global South have developed around strengths, needs, and interests of children and their communities, as well as constraints and affordances of local infrastructure. We use the lens of minimal computing – which focuses on community context and constraints in decisions about technology – to describe our collaborative work on OctoStudio. We describe trade-offs involved in the design decisions, and highlight insights from the process of collaboration to develop tools and practices that are more responsive and meaningful to communities who are often excluded from design decisions that impact them.

Expected Constant Round Byzantine Broadcast under Dishonest Majority

2026-02-12T03:06:52Z

Expected Constant Round Byzantine Broadcast under Dishonest Majority Wan, Jun; Xiao, Hanshen; Shi, Elaine; Devadas, Srinivas Byzantine Broadcast (BB) is a central question in distributed systems, and an important challenge is to understand its round complexity. Under the honest majority setting, it is long known that there exist randomized protocols that can achieve BB in expected constant rounds, regardless of the number of nodes n. However, whether we can match the expected constant round complexity in the corrupt majority setting --- or more precisely, when f > n/2 --- remains unknown, where f denotes the number of corrupt nodes. In this paper, we are the first to resolve this long-standing question. We show how to achieve BB in expected O((n/(n-f))2) rounds. Our results hold under a weakly adaptive adversary who cannot perform ``after-the-fact removal' of messages already sent by a node before it becomes corrupt. We also assume trusted setup and the Decision Linear (DLIN) assumption in bilinear groups.

Minimum Plane Bichromatic Spanning Trees

2026-02-12T03:06:50Z

Minimum Plane Bichromatic Spanning Trees Akitaya, Hugo; Biniaz, Ahmad; Demaine, Erik; Kleist, Linda; Stock, Frederick; T?th, Csaba D. For a set of red and blue points in the plane, a minimum bichromatic spanning tree (MinBST) is a shortest spanning tree of the points such that every edge has a red and a blue endpoint. A MinBST can be computed in O(n log n) time where n is the number of points. In contrast to the standard Euclidean MST, which is always plane (noncrossing), a MinBST may have edges that cross each other. However, we prove that a MinBST is quasi-plane, that is, it does not contain three pairwise crossing edges, and we determine the maximum number of crossings. Moreover, we study the problem of finding a minimum plane bichromatic spanning tree (MinPBST) which is a shortest bichromatic spanning tree with pairwise noncrossing edges. This problem is known to be NP-hard. The previous best approximation algorithm, due to Borgelt et al. (2009), has a ratio of O(sqrt(n)). It is also known that the optimum solution can be computed in polynomial time in some special cases, for instance, when the points are in convex position, collinear, semi-collinear, or when one color class has constant size. We present an O(log n)-factor approximation algorithm for the general case.

Nested Dissection Meets IPMs: Planar Min-Cost Flow in Nearly-Linear Time

2026-02-12T03:07:19Z

Nested Dissection Meets IPMs: Planar Min-Cost Flow in Nearly-Linear Time Dong, Sally; Gao, Yu; Goranci, Gramoz; Lee, Yin Tat; Sachdeva, Sushant; Peng, Richard; Ye, Guanghao We present a nearly-linear time algorithm for finding a minimum-cost flow in planar graphs with polynomially bounded integer costs and capacities. The previous fastest algorithm for this problem is based on interior point methods (IPMs) and works for general sparse graphs in O(n1.5 polylog n)) time [Daitch-Spielman, STOC'. Intuitively, ?(n1.5) is a natural runtime barrier for IPM-based methods, since they require ?n iterations, each routing a possibly-dense electrical flow. To break this barrier, we develop a new implicit representation for flows based on generalized nested dissection [Lipton-Rose-Tarjan, SINUM'79] and approximate Schur complements [Kyng-Sachdeva, FOCS'. This implicit representation permits us to design a data structure to route an electrical flow with sparse demands in roughly ?n update time, resulting in a total runtime of O(n polylog n). Our results immediately extend to all families of separable graphs.

Enhancing Electric Vehicle Security and Privacy through Decentralized Identity Management

2026-02-12T03:06:49Z

Enhancing Electric Vehicle Security and Privacy through Decentralized Identity Management Aydeger, Abdullah; Zeydan, Engin; Mangues-Bafalluy, Josep; Arslan, Suayb; Turk, Yekta In the next decade, electric vehicles (EVs) are expected to contribute to reducing climate change and transforming road mobility significantly. However, the security and privacy of EV charging systems present considerable challenges that need to be addressed. This paper introduces a novel approach by integrating blockchain-based self-sovereign identity (SSI) to enhance the security and privacy of EV charging systems. By leveraging decentralized and immutable nature of blockchain, the proposed SSI framework can ensure secure and private data exchanges between EVs, charging stations, and backend systems. This three-way integration addresses the vulnerabilities identified in existing EV charging methods, such as conductive, inductive, and battery swapping, and complies with cybersecurity regulations like UNECE R155. This paper provides a comprehensive analysis, practical case study, and evaluation of the security and privacy enhancements achieved through the proposed SSI framework, offering valuable insights for industry professionals and researchers. We have conducted extensive end-to-end testing to evaluate the performance of our blockchain-based SSI framework in the EV charging ecosystem, focusing on identity verification, credential management and service orchestration. The results show that the system enables fast wallet creation, efficient metadata retrieval and low-latency service deployment, ensuring seamless identity management and service orchestration.

Local Distributed Rounding: Generalized to MIS, Matching, Set Cover, and Beyond

2026-02-12T03:06:46Z

Local Distributed Rounding: Generalized to MIS, Matching, Set Cover, and Beyond Faour, Salwa; Ghaffari, Mohsen; Grunau, Christoph; Kuhn, Fabian; Rozho?, V?clav We develop a general deterministic distributed method for locally rounding fractional solutions of graph problems for which the analysis can be broken down into analyzing pairs of vertices. Roughly speaking, the method can transform fractional/probabilistic label assignments of the vertices into integral/deterministic label assignments for the vertices, while approximately preserving a potential function that is a linear combination of functions, each of which depends on at most two vertices (subject to some conditions usually satisfied in pairwise analyses). The method unifies and significantly generalizes prior work on deterministic local rounding techniques [Ghaffari, Kuhn FOCS'21; Harris FOCS'19; Fischer, Ghaffari, Kuhn FOCS'17; Fischer DISC' to obtain polylogarithmic-time deterministic distributed solutions for combinatorial graph problems. Our general rounding result enables us to locally and efficiently derandomize a range of distributed algorithms for local graph problems, including maximal independent set (MIS), maximum-weight independent set approximation, and minimum-cost set cover approximation.

Feasibility Study on Heat Pipes for Radio Frequency Antennas

2026-02-12T03:07:33Z

Feasibility Study on Heat Pipes for Radio Frequency Antennas Jung, Minuk; Watterson, Amy; Wallace, Gregory M The applicability of a heat pipe is investigated for the cooling of radio frequency antennas in fusion reactors operating at high temperatures. A heat pipe is a passive cooling device that transfers a large amount of heat through the liquid-vapor phase change and pumps the working fluid by the surface tension of the wick structure without moving parts. As the heat pipe is expected to operate near 1000 K, refractory metals or ceramics should be used for wall materials, and liquid metals are primarily considered as the working fluid. However, liquid metals are electrically conductive, and the strong magnetic field perpendicular to the flow direction imposes significant magnetohydrodynamic (MHD) flow resistance in addition to viscous friction, which impairs heat transfer performance. Since a strong magnetic field is inevitable in magnetic confinement fusion reactors, materials with low electrical conductivity should be applied to wall coatings to reduce the MHD effect. Heat flux limitations at a magnetic field of 10 T and a condenser coolant temperature of 773 K are estimated using COMSOL multiphysics, which can capture the fully developed MHD wick flow, laminar/turbulent vapor flow, and heat transfer simultaneously. For simplicity, the generic heat pipe geometry of a straight horizontal cylinder with a length of 2 ft (0.6096 m) is employed. Optimal geometrical parameters are evaluated to meet radial evaporator/condenser heat fluxes greater than 0.1 MW/m2, even under a strong MHD effect.

Regional incidence and persistence of high-growth firms: testing ideas from the entrepreneurial ecosystems literature

2026-02-12T03:07:25Z

Regional incidence and persistence of high-growth firms: testing ideas from the entrepreneurial ecosystems literature Coad, Alex; Domnick, Clemens; Santoleri, Pietro; Srhoj, Stjepan Policymakers and scholars often assume that a higher incidence of high-growth firms (HGFs) is synonymous with vibrant regional economic dynamics, and that HGF shares are persistent over time as entrepreneurial ecosystems (EEs) have slowly changing features. In this paper we test these hypotheses, which are deeply rooted in the EE literature. Results do not provide strong support for the hypothesis that more developed regions feature higher HGF shares. We do find evidence consistent with HGF shares displaying persistency over time. However, we show that more developed regions do not have higher persistence in their HGF shares, and that the strength in persistence does not increase across the HGFs distribution, which does not support path-dependency as the main mechanism behind the observed persistence. Overall, we call for a more nuanced interpretation of both regional HGF shares and the EEs literature.

Knives out: response to critics

2026-02-12T03:07:34Z

Knives out: response to critics Khoo, Justin Writing a book can feel like a solitary endeavor. You labor for (in my case) years, sometimes talking about parts of the project with others, but mostly toiling alone to work out the consequences of commitments you made months and years prior. I'm grateful for the opportunity to engage with three brilliant interlocutors about these ideas, which for so long seemed to matter to no one besides myself (and maybe my publisher).

Toward Ontological Alignment: Coordinating Student Ideas with the Representational System of a Computational Modeling Unit for Science Learning

2026-02-12T03:07:10Z

Toward Ontological Alignment: Coordinating Student Ideas with the Representational System of a Computational Modeling Unit for Science Learning Wagh, Aditi; Rosenbaum, Leah F.; Fuhrmann, Tamar; Eloy, Adelmo; Blikstein, Paulo; Wilkerson, Michelle Computational modeling tools present unique opportunities and challenges for student learning. Each tool has a representational system that impacts the kinds of explorations students engage in. Inquiry aligned with a tool’s representational system can support more productive engagement toward target learning goals. However, little research has examined how teachers can make visible the ways students’ ideas about a phenomenon can be expressed and explored within a tool’s representational system. In this paper, we elaborate on the construct of ontological alignment—that is, identifying and leveraging points of resonance between students’ existing ideas and the representational system of a tool. Using interaction analysis, we identify alignment practices adopted by a science teacher and her students in a computational agent-based modeling unit. Specifically, we describe three practices: (1) Elevating student ideas relevant to the tool’s representational system; (2) Exploring and testing links between students’ conceptual and computational models; and (3) Drawing on evidence resonant with the tool’s representational system to differentiate between theories. Finally, we discuss the pedagogical value of ontological alignment as a way to leverage students’ ideas in alignment with a tool’s representational system and suggest the presented practices as exemplary ways to support students’ computational modeling for science learning.

Stand Up and Split. Desiring Desertion in Jean Giono and Emmanuelle Lambert

2026-02-12T03:07:22Z

Stand Up and Split. Desiring Desertion in Jean Giono and Emmanuelle Lambert Perreau, Bruno To face the powers that be, contemporary French writer Virginie Despentes proposes a straightforward solution: “stand up and split!” But where to go and with whom? How do we stop the proliferation of contested norms if we clear the decks? In a context of ecological crisis, desiring desertion is not rare even if we have only one world to inhabit. This article analyzes the desire to desert from two texts: Le Déserteur et autres récits (Citation1966 [1973]) by Jean Giono and La Désertion (Citation2018a) by Emmanuelle Lambert. It demonstrates that desertion does not make a clean sweep of the past but rather accepts the desert at the heart of existence. That is, both presence and disappearance.

Fuel Behavior Implications of Reactor Design Choices in Pressurized Water SMRs

2026-02-12T03:07:18Z

Fuel Behavior Implications of Reactor Design Choices in Pressurized Water SMRs Halimi, Assil; Shirvan, Koroush Small pressurized water reactors can feature boron-free operation, natural circulation mode, reduced-height assemblies, and/or long refueling cycles. This paper attempts to explore core design optimization for each of these design evolutions. In consequence, five core design layouts are developed incorporating boron-free operation with continuous control rod insertion, natural circulation with low burnup/low power density design, natural circulation with high burnup/low power density design, forced circulation with standard core power density design, and forced circulation with high power density design. These cores’ performance is compared to a standard four-loop pressurized water reactor. The design process aims to improve the fuel cycle cost under safety constraints through core design optimization using the CASMO4E/SIMULATE3 reactor physics codes and the FRAPCON4.1 fuel performance assessment tool. Core modeling assumes standard 17×17 PWR fuel assemblies loaded with low enriched uranium up to 5 wt% or low enriched uranium plus (i.e. below 10 wt% enrichment) pellets with gadolinium oxide as the burnable poison. Satisfactory core and fuel performances are obtained for all the designed cores under steady state and considered overpower transients. For low power density operation, long cycle lengths are achieved reaching 2.5-year and 5-year cycles, and peak rod-average burnup is pushed to 83 MWd/kgU. Other cycle lengths are maintained at 18 months. Boron-free operation exhibits the ability to achieve longer cycle lengths at the cost of higher peaking factors leading to high local power and fuel temperatures, which prevents sizable power uprates and is deemed uneconomical. Fuel assembly height reduction allows coolant velocity retrofit, which enables higher core power density without violating the structural integrity of the fuel assembly. As a result, a core power density of 123 kW/L is reached where total cladding hoop strain becomes the limiting parameter.

Shades of authoritarian digital sovereignty: divergences in Russian and Chinese data localisation regimes

2026-02-12T03:07:21Z

Shades of authoritarian digital sovereignty: divergences in Russian and Chinese data localisation regimes Khasanova, Liliya; Tai, Katharin The concept of sovereignty is now referred to in cyberspace-related policy by a range of governments, both authoritarian and democratic. At the same time, the most prominent proponents of state – or sovereignty-centric models of internet governance are Russia and China, whose positions are often characterised as a shared ‘Sino-Russian’ model. This paper subjects this idea of a shared Sino-Russian approach to empirical scrutiny by conducting a comparative analysis of rules, regulations and policies on data localisation in both countries. By delimiting the research question to regulations on data localisation and cross-border data transfers in both countries, we identify an important set of similarities and differences between the Russian and Chinese approaches. They share some features associated with authoritarian regimes, such as uncertainty around the selective enforcement of broadly formulated rules and a centralised assessment of outbound data transfers. However, we also find significant differences in the level of institutional centralisation, degrees of responsiveness within the policymaking process, and economic logics driving data localisation and cross-border transfer regulations. Based on these findings, we argue that despite a perception that Russia and China adhere to a similar model of authoritarian digital sovereignty, there are significant disparities in their data localisation regimes.

Race, profit, and algorithms: Neighborhood-level analysis of iBuyers’ profit margin

2026-02-12T03:07:08Z

Race, profit, and algorithms: Neighborhood-level analysis of iBuyers’ profit margin So, Wonyoung iBuyers are firms that use automated valuation models (AVMs), streamline home buying processes, and provide all-cash offers to purchase homes. Although the previous literature has explored the roles and limitations of iBuyers in the housing market, empirical research on the racial implications of these algorithmic home buying processes remains understudied. Using a spatial lag model, this study shows the spatial clustering of iBuyer profit margins, that iBuyers gain more profits when they resell to individuals than institutions, and that some iBuyers have a statistically significant correlation between their profit margins and the proportion of marginalized racial groups within a census tract, while controlling for individual housing characteristics, neighborhood housing quality and demand, and neighborhood amenities and socioeconomic factors. These findings suggest that the more adeptly iBuyers can forecast housing values, the greater the potential to maximize profits from homeowners in communities of color. Consequently, this research contributes to the understanding of how technological mechanisms operate within a purportedly race-neutral framework and advocates for the development and deployment of algorithmic systems guided by the principles of antisubordination, rather than relying solely on notions of “fairness” and anticlassification.

Powering Through the Turn: Finding Time for Concept Exploration Before Industry Stagnation

2026-02-12T03:07:29Z

Powering Through the Turn: Finding Time for Concept Exploration Before Industry Stagnation Noble, Connery; Cameron, Bruce G This study examines how the tension between exploration and exploitation affects early-stage development within the engineering teams of large corporations. Using survey data collected from over 900 system engineers and managers, it was observed that exploration decreased as an organization’s market growth declined, but dire market projections prompted a refocus on exploration. In addition, engineers routinely desire more concept exploration time than they perceive that they have available. The authors argue that engineering teams should more intentionally consider their innovation strategy, and that companies with stagnant market growth should invest in concept exploration before they get to a period of market decline.

Vital Biodiversity Systems: A Companion Paper

2026-02-12T03:01:00Z

Vital Biodiversity Systems: A Companion Paper Westerlaken, Michelle; Bischoff, Amanda; Mertens, Krishen; Pertusa, Alejandro Regenerative and diverse ecosystems are essential to living futures. Healthy ecosystems are more resilient to climate change and are better able to absorb and store carbon. Communities and corporations worldwide are currently establishing how environmental data can best support these processes. This Companion Paper provides the rationale for the Design Brief and synthesizes findings from four years of research across academia, corporate sustainability teams, and community stakeholders. It argues that biodiversity data systems are not neutral repositories but designed artefacts that embed assumptions and values. To redirect innovation, the paper supports the Brief by expanding on its key design principles, criteria, constraints, and propositions that together chart a pathway for ‘vital biodiversity systems’: platforms that embed the aliveness of the ecosystems they mediate.

Molecular dynamics simulations and structural bioinformatics of bacterial integral alpha-helical membrane enzymes and their AlphaFold2-predicted water-soluble QTY analogues

2026-02-12T03:07:30Z

Molecular dynamics simulations and structural bioinformatics of bacterial integral alpha-helical membrane enzymes and their AlphaFold2-predicted water-soluble QTY analogues Sajeev-Sheeja, Akash; Karagöl, Alper; Karagöl, Taner; Zhang, Shuguang The study of integral membrane proteins has long been challenging because of their poor solubility in aqueous environments. We previously used QTY code to enhance the hydrophilicity in alpha-helices, beta-barrels, and monoclonal antibodies by systematically pairwise replacing the hydrophobic amino acids L (leucine) to Q (glutamine), V(valine)/I(isoleucine) to T (threonine), and F (phenylalanine) to Y (tyrosine). The superposed AlphaFold2-predicted structures of alpha-helical transmembrane enzyme variants with >41% amino acid substitutions displayed remarkable similarity to native structures (RMSD 0.3Å-0.7 Å). We conducted molecular dynamics (MD) simulations, which revealed that, even in the absence of a lipid bilayer, the QTY-modified enzymes retained stable dynamics comparable to their membrane-bound forms. Root mean square fluctuation (RMSF) values remained below 2 Å across the transmembrane and core regions, and residue-wise root mean square deviation (RMSD) values were minimal (<3 Å), indicating that the structural integrity of the protein core was largely preserved. These results suggest that the QTY variants, designed for soluble environments, effectively mimic the stability and conformational rigidity of natural membrane-bound enzymes. Our findings show that the QTY code is a simple method for designing water-soluble membrane protein enzymes in different biological scenarios, and it may encourage further experiments to validate our structural bioinformatics research.

City of ‘social saints’: the role of place in driving impact entrepreneurship in Turin, Italy

2026-02-12T03:07:27Z

City of ‘social saints’: the role of place in driving impact entrepreneurship in Turin, Italy Burke, Mary Kathleen; Sydow, Alisa; Torchia, Daniel; Corazza, Laura This paper theorizes impact entrepreneurship (IE) in relation to place by examining dynamics at the individual, community, and organizational levels. While existing IE literature emphasizes entrepreneurship aimed at addressing grand challenges, it often adopts an aggregate view that overlooks how locally embedded entrepreneurs access and mobilize social and economic resources. We introduce a novel, multidimensional framework to show how sense of place, community embeddedness and IE interrelate to shape approaches to current social/environmental challenges. Adopting a qualitative approach, this paper investigates how different actors in Turin, Italy, contribute to IE through building on a legacy of social sector institutions. We find that individuals identifying with a place-based vocation of social impact find communities with a shared volition to work together and across organizations. We contribute to understanding how individuals’ senses of place can be leveraged into wider community efforts to support IE in the region. The paper advances the IE concept to account for the individual perspectives influencing local organizing practices and visions for IE rooted in place.

Plane Delivery: Towards a Physical Grammar for Large-Scale Digital Fabrication

2026-02-12T03:07:12Z

Plane Delivery: Towards a Physical Grammar for Large-Scale Digital Fabrication Sass, Lawrence There will come a day when computers and robots will participate regularly in designing, fabricating, and delivering homes as customized kits of parts (Sass Citation2008). They will not replace builders. Instead, one possible future is where computers and robots operate as intelligent assistants, discovering, reasoning, and inferring the best solutions using large language models (LLMs). This language will be vector-based on points, lines, and planes of the type Stiny described (Stiny Citation2006). A standard design and builder language is a first step towards automation. The proposed system is of a Lego-style approach to physical house production, used to manage costs, enhance design variety, improve design quality, and, most importantly, facilitate building.

Techno-statecraft and industrial strategy: semiconductor development in Arizona

2026-02-12T03:07:24Z

Techno-statecraft and industrial strategy: semiconductor development in Arizona Kollar, Justin The resurgence of U.S. industrial strategy discourse is not a centralised return of the state but a territorially fragmented form of techno-statecraft. This article analyzes Arizona's semiconductor expansion as a case in which subnational actors – agencies, utilities, universities, and developers – mobilise infrastructure, land-use policy, and regulatory coordination to attract global capital. Rather than a coherent national plan, Arizona's strategy reflects speculative governance oriented toward risk absorption and territorial readiness. The article situates this conjuncture within longer histories of militarised growth and infrastructural overbuild, contributing to debates on state capitalism, industrial strategy, and the spatial politics of techno-industrial transformation.

Ideology, Equity, and Structure: Comments on Tzu-wei Hung’s ‘Equity and Marxist Buddhism’

2026-02-12T03:07:35Z

Ideology, Equity, and Structure: Comments on Tzu-wei Hung’s ‘Equity and Marxist Buddhism’ Haslanger, Sally In his essay, ‘Equity and Marxist Buddhism’, Tzu-wei Hung argues that Marxist Buddhism brings a commitment to social justice together with a distinctive form of virtue theory. In my commentary, I raise several questions from a Marxian perspective: (1) Might it be argued that Marxist Buddhism is (in the critical sense) ideological (similar to religion) because the spiritual goal of ‘transcendence’ distracts us from the need to fight for emancipation? (2) Can justice as equity be achieved by promoting individual altruism? (3) Aren’t both mainstream accounts of justice and Marxist Buddhism aspirational and so need to rely on non-ideal theory to achieve justice?

A data-driven and context-aware approach for demand forecasting in the beverage industry

2026-02-12T03:07:07Z

A data-driven and context-aware approach for demand forecasting in the beverage industry Ma, Benedict Jun; Jackson, Ilya; Huang, Maggie; Villegas, Sebastian; Macias-Aguayo, Jaime Accurate demand forecasting is essential for logistics and supply chain management as it enables efficient inventory planning, reduces operational costs, and ensures high service levels across the network. However, in practice, diverse demand patterns of items make this task challenging, and a one-size-fits-all forecasting approach is inadequate. This paper proposes a data-driven and context-aware forecasting framework and tests it by using both endogenous data from a large private-label beverage manufacturer and exogenous features (such as holidays and temperature). Our method begins by classifying SKUs based on demand volume, volatility, and intermittency, and then refining the derived clusters by taking volume distribution into account. Totally, we obtain four distinct clusters, which are (i) stable and high volume, (ii) stable with low volume, (iii) erratic and intermittent, and (iv) lumpy. To explore the appropriate forecasting models for different demand patterns, we employ statistical models (exponential smoothing, ARIMA, and Croston), machine learning models (XGBoost), deep learning models (TiDE and N-BEATS), and even qualitative approaches such as collaborative planning, forecasting, and replenishment (CPFR). Our experimental results suggest which forecasting models are recommended for each demand pattern, and insightful implications are provided for the managers.

Adversarial Network Optimization under Bandit Feedback: Maximizing Utility in Non-Stationary Multi-Hop Networks

2026-02-11T04:36:10Z

Adversarial Network Optimization under Bandit Feedback: Maximizing Utility in Non-Stationary Multi-Hop Networks Dai, Yan; Huang, Longbo Stochastic Network Optimization (SNO) concerns scheduling in stochastic queueing systems and has been widely studied in network theory. Classical SNO algorithms require network conditions to be stationary w.r.t. time, which fails to capture the non-stationary components in increasingly many real-world scenarios. Moreover, most existing algorithms in network optimization assume perfect knowledge of network conditions before decision, which again rules out applications where unpredictability in network conditions presents. Motivated by these issues, this paper considers Adversarial Network Optimization (ANO) under bandit feedback. Specifically, we consider the task of i) maximizing some unknown and time-varying utility function associated with scheduler's actions, where ii) the underlying network topology is a non-stationary multi-hop network whose conditions change arbitrarily with time, and iii) only bandit feedback (the effect of actually deployed actions) is revealed after decision-making. We propose the UMO2 algorithm, which does not require any pre-decision knowledge or counterfactual feedback, ensures network stability, and also matches the utility maximization performance of any ''mildly varying'' reference policy up to a polynomially decaying gap. To our knowledge, no previous algorithm can handle multi-hop networks or achieve utility maximization guarantees in ANO problems with bandit feedback, whereas ours is able to do both. Technically, our method builds upon a novel integration of online learning techniques into the Lyapunov drift-plus-penalty method. Specifically, we propose meticulous analytical techniques to jointly balance online learning and Lyapunov arguments, which is used to handle the complex inter-dependency among queues in multi-hop networks. To tackle the learning obstacles due to potentially unbounded queue sizes and negative queue differences, we design a new online linear optimization algorithm that automatically adapts to the unknown (potentially negative) loss magnitudes. Finally, we also propose a bandit convex optimization algorithm with novel queue-dependent learning rate scheduling that suites drastically varying queue lengths in utility maximization. Our new insights and techniques in online learning can also be of independent interest. SIGMETRICS Abstracts ’25, Stony Brook, NY, USA

IDAT: A Multi-Modal Dataset and Toolkit for Building and Evaluating Interactive Task-Solving Agents

2026-02-11T04:36:14Z

IDAT: A Multi-Modal Dataset and Toolkit for Building and Evaluating Interactive Task-Solving Agents Mohanty, Shrestha; Arabzadeh, Negar; Tupini, Andrea; Sun, Yuxuan; Skrynnik, Alexey; Zholus, Artem; C?t?, Marc-Alexandre; Kiseleva, Julia Seamless interaction between AI agents and humans using natural language remains a key goal in AI research. This paper addresses the challenges of developing interactive agents capable of understanding and executing grounded natural language instructions through the IGLU competition. Despite advancements, challenges such as a scarcity of appropriate datasets and the need for effective evaluation platforms persist. We introduce a scalable data collection tool for gathering interactive grounded language instructions within a Minecraft-like environment, resulting in a Multi-Modal dataset with around 9,000 utterances and over 1,000 clarification questions. Additionally, we present a Human-in-the-Loop interactive evaluation platform for qualitative analysis and comparison of agent performance through multi-turn communication with human annotators. We offer to the community these assets referred to as IDAT (IGLU Dataset And Toolkit) which aim to advance the development of intelligent, interactive AI agents and provide essential resources for further research. SIGIR ’25, Padua, Italy

A Theory to Estimate, Bound, and Manage Systemic Cyber-Risk

2026-02-11T04:36:11Z

A Theory to Estimate, Bound, and Manage Systemic Cyber-Risk Pal, Ranjan; Duan, Konnie; Sequeira, Rohan The market to manage critical infrastructure cyber-risks using cyber insurance (CI) has been growing steadily (but not fast enough) as it is still skeptical of the extent of economic and societal impact of systemic risk across networked supply chains in interdependent IT-driven enterprises. To demystify this skepticism, we first study in this paper the role of (a) the statistical nature of multiple enterprise cyber-risks contributing to aggregate supply chain risk and (b) the graph structure of the underlying enterprise supply chain network, in the statistical spread of aggregate cyber-risk. We provide statistical tail bounds on the aggregate cyber-risk that a risk managing firm such as a cyber insurer is exposed to in a supply chain. Subsequently, we study the problem of aggregate cyber-risk management by cyber re-insurance firms via portfolio design to optimally diversify aggregate/systemic cyber-risk sourced from multiple CIs insuring enterprises on a supply chain. We propose the first mathematical framework for re-insurers to test the operational sustainability of systemic cyber-risk diversification portfolios with respect to the standard Value-at-Risk (VaR) metric for general aggregate cyber risk distributions. We also propose a statistical copula methodology to make systemic cyber-risk portfolio diversification sustainable for re-insurers in scenarios where the sustainability test fails. We validate our theory via Monte Carlo simulations. SIGSIM-PADS ’25, Santa Fe, NM, USA

14.41 Public Finance and Public Policy, Fall 2010

2026-02-10T18:32:40Z

14.41 Public Finance and Public Policy, Fall 2010 Gruber, Jonathan Explores the role of government in the economy, applying tools of basic microeconomics to answer important policy questions such as government response to global warming, school choice by K-12 students, Social Security versus private retirement savings accounts, government versus private health insurance, setting income tax rates for individuals and corporations.

Dynamic Incentive Allocation for City-Scale Deep Decarbonization

2026-02-11T04:35:56Z

Dynamic Incentive Allocation for City-Scale Deep Decarbonization Sitaraman, Anupama; Lechowicz, Adam; Bashir, Noman; Liu, Xutong; Hajiesmaili, Mohammad; Shenoy, Prashant Greenhouse gas emissions from the residential sector represent a large fraction of global emissions and must be significantly curtailed to achieve ambitious climate goals. To stimulate the adoption of relevant technologies such as rooftop PV and heat pumps, governments and utilities have designed incentives that encourage adoption of decarbonization technologies. However, studies have shown that many of these incentives are inefficient since a substantial fraction of spending does not actually promote adoption. Further, these incentives are not equitably distributed across socioeconomic groups. In this paper, we present a novel data-driven approach that adopts a holistic, emissions-based, and city-scale perspective on decarbonization. We propose an optimization model that dynamically allocates a total incentive budget to households to directly maximize the resultant carbon emissions reduction -- this is in contrast to prior work, which focuses on metrics such as the number of new installations. We leverage techniques from the multi-armed bandits problem to estimate human factors, such as a household's willingness to adopt new technologies given a certain incentive. We apply our proposed dynamic incentive framework to a city in the Northeast U.S., using real household energy data, grid carbon intensity data, and future price scenarios. We compare our learning-based technique to two baselines, one "status-quo' baseline using incentives offered by a state and utility, and one simple heuristic baseline. With these baselines, we show that our learning-based technique significantly outperforms both the status-quo baseline and the heuristic baseline, achieving up to 37.88% higher carbon reductions than the status-quo baseline and up to 28.76% higher carbon reductions compared to the heuristic baseline. Additionally, our incentive allocation approach is able to achieve significant carbon reduction even in a broad set of environments, with varying values for electricity and gas prices, and for carbon intensity of the grid. Finally, we show that our framework can accommodate equity-aware constraints to preserve an equitable allocation of incentives across socioeconomic groups while achieving 83.34% of the carbon reductions of the optimal solution on average.

The Economics of Large Language Models: Token Allocation, Fine-Tuning, and Optimal Pricing

2026-02-11T04:36:05Z

The Economics of Large Language Models: Token Allocation, Fine-Tuning, and Optimal Pricing Bergemann, Dirk; Bonatti, Alessandro; Smolin, Alex We develop an economic framework to analyze the optimal pricing and product design of Large Language Models (LLM). Our framework captures several key features of LLMs: variable operational costs of processing input and output tokens; the ability to customize models through fine-tuning; and high-dimensional user heterogeneity in terms of task requirements and error sensitivity. In our model, a monopolistic seller offers multiple versions of LLMs through a menu of products. The optimal pricing structure depends on whether token allocation across tasks is contractible and whether users face scale constraints. When it is possible to contract on the entire assignment of tokens to tasks, the seller's problem ("Token Allocations") is an infinite-dimensional screening problem, which is well-known to be difficult. We are nonetheless able to make progress in two important classes of environments: binary environment and two dimensional value-scale heterogeneity, in which case users with similar aggregate value-scale characteristics choose similar levels of fine-tuning and token consumption. When only the total number of tokens is contractible ("Token Packages"), we leverage the tractability of a constant elasticity of substitution framework to drastically simplify the problem: the buyer's type-a function mapping each task to a value of precision- is an index. This index for the value of precision allows the seller to solve a one-dimensional screening problem. The optimal mechanism can be implemented through menus of two-part tariffs, with higher markups for more intensive users. Our results rationalize observed industry practices such as tiered pricing based on model customization and usage levels. EC ’25, July 7–10, 2025, Stanford, CA, USA

Alternates, Assemble! Selecting Optimal Alternates for Citizens’ Assemblies

2026-02-11T04:36:09Z

Alternates, Assemble! Selecting Optimal Alternates for Citizens’ Assemblies Assos, Angelos; Baharav, Carmel; Flanigan, Bailey; Procaccia, Ariel Citizens' assemblies are an increasingly influential form of deliberative democracy, where randomly selected people discuss policy questions. The legitimacy of these assemblies hinges on their representation of the broader population, but participant dropout often leads to an unbalanced composition. In practice, dropouts are replaced by preselected alternates, but existing methods do not address how to choose these alternates. To address this gap, we introduce an optimization framework for alternate selection. Our algorithmic approach, which leverages learning-theoretic machinery, estimates dropout probabilities using historical data and selects alternates to minimize expected misrepresentation. Our theoretical bounds provide guarantees on sample complexity (with implications for computational efficiency) and on loss due to dropout probability mis-estimation. Empirical evaluation using real-world data demonstrates that, compared to the status quo, our method significantly improves representation while requiring fewer alternates. EC ’25, July 7–10, 2025, Stanford, CA, USA

Does Firm Size Influence the Collection of Sensitive Data?: A Study of Child-Orientated Apps

2026-02-11T04:36:07Z

Does Firm Size Influence the Collection of Sensitive Data?: A Study of Child-Orientated Apps Cecere, Grazia; Tucker, Catherine; Lefrere, Vincent How does firm size affect the privacy protections offered to customers? On the one hand, it could be that larger firms use their size to amass more data. On the other hand, smaller firms may be less careful in their data protection practices, because they have a different perception of risk. Using data from the Google Play Store over a three-year period, we explore this empirical question in the U.S. children's app market. Our findings indicate that larger app developers consistently implement stronger privacy protections, requesting less sensitive data compared to smaller developers. These results hold across empirical approaches, including instrumental variables and the propensity-score matching approach. Additionally, our analysis shows that mergers between developers and sudden increases in size of the user-bases of the product are associated with reduced data collection. We show that newly created and updated apps produced by large developers collect less data compared to existing apps. Our findings indicate a trend toward standardized privacy practices across different national regulatory regimes. This research highlights the potential for growth-driven improvements in data privacy practices among app developers, regardless of their regulatory context. EC ’25, July 7–10, 2025, Stanford, CA, USA

Inertial Coordination Games

2026-02-11T04:36:06Z

Inertial Coordination Games Koh, Andrew; Li, Ricky; Uzui, Kei Coordination lies at the heart of many economic phenomena. A well-known example is currency crises, in which traders decide whether to launch a speculative attack. On one hand, shocks to the currency's fundamentals can propagate: as more traders attack, the central bank's foreign reserves are depleted, which in turn encourages further attacks as traders seek to exploit a weakening currency. On the other hand, shocks can also fizzle out: traders may quickly learn that the central bank's balance sheet is strong, causing pessimism to dissipate and attacks to subside. When do shocks propagate, and when do they fizzle out? In particular, how do these outcomes depend on the speed at which traders learn about the fundamental? Motivated by these questions, we propose a model of inertial coordination games—dynamic coordination games where players repeatedly decide whether to take a risky action. The payoff from this risky action depends on (i) a persistent fundamental; and (ii) an endogenous component that depends on others' past play. Players receive private signals about the persistent fundamental over time and form beliefs about the current state. Notably, the current state depends on past play, which in turn depends on past beliefs about play yet farther back into the past. Thus, expectations about histories shape behavior in the present which, in turn, drives the evolution of future states and future play. Our main result develops a tight connection between the speed of learning and limit aggregate play: the risk-dominant action is played in the limit if and only if posterior precisions grow sub-quadratically. This has sharp implications for the long-run propagation of shocks. With slow (sub-quadratic) learning, limit play exhibits history independence: initial shocks have no lasting effect, and limit play is determined solely by fundamentals. By contrast, with fast (super-quadratic) learning, limit play is history dependent: initial shocks can be self-fulfilling, and whether they propagate depends jointly on fundamentals, the size of the shock, and the speed of learning. Our results offer a novel perspective on whether 'history' or 'expectations' shape long-run coordination outcomes: in our model, expectations about histories is what matters for whether self-fulfilling spirals occur. Finally, we show that the speed of learning also shapes the path of play, focusing on the case of sub-quadratic learning. When signals are precise, aggregate play exhibits a sudden transition from nearly all players choosing the non-risk-dominant action to nearly all players choosing the risk-dominant action. In contrast, when signals are noisy, the transition is gradual, with the share of players choosing the risk-dominant action increasing gradually over time. This suggests that "spikes" in aggregate behavior (such as a sudden and massive sell-off of a currency) can be consistent with transition to limit equilibrium play, and need not indicate an "equilibrium shift." EC ’25, July 7–10, 2025, Stanford, CA, USA

Galley: Modern Query Optimization for Sparse Tensor Programs

2026-02-11T04:36:13Z

Galley: Modern Query Optimization for Sparse Tensor Programs Deeds, Kyle; Ahrens, Willow; Balazinska, Magdalena; Suciu, Dan The tensor programming abstraction is a foundational paradigm which allows users to write high performance programs via a high-level imperative interface. Recent work on sparse tensor compilers has extended this paradigm to sparse tensors (i.e., tensors where most entries are not explicitly represented). With these systems, users define the semantics of the program and the algorithmic decisions in a concise language that can be compiled to efficient low-level code. However, these systems still require users to make complex decisions about program structure and memory layouts to write efficient programs. This work presents .Galley, a system for declarative tensor programming that allows users to write efficient tensor programs without making complex algorithmic decisions. Galley is the first system to perform cost based lowering of sparse tensor algebra to the imperative language of sparse tensor compilers, and the first to optimize arbitrary operators beyond Σ and *. First, it decomposes the input program into a sequence of aggregation steps through a novel extension of the FAQ framework. Second, Galley optimizes and converts each aggregation step to a concrete program, which is compiled and executed with a sparse tensor compiler. We show that Galley produces programs that are 1-300x faster than competing methods for machine learning over joins and 5-20x faster than a state-of-the-art relational database for subgraph counting workloads with a minimal optimization overhead.

Virtualizing Cloud Data Infrastructures with BRAD

2026-02-11T04:36:16Z

Virtualizing Cloud Data Infrastructures with BRAD Yu, Geoffrey; Wu, Ziniu; Kossmann, Ferdi; Li, Tianyu; Markakis, Markos; Ngom, Amadou; Zhang, Sophie; Kraska, Tim; Madden, Samuel Organizations usually manage their data using multiple specialized cloud database engines (e.g., Aurora, BigQuery, etc.). However, designing and managing multi-engine infrastructures is hard; there can be many designs, each with different performance and costs. Changing the design afterwards (e.g., due to growth) is even more challenging since application code usually ends up tightly coupled to the engines. We propose data infrastructure virtualization. The key idea is to declare a set of virtual database engines (VDBEs), which specify an engine's application-facing properties (e.g., query interface, performance) and its tables, but do not prescribe a concrete engine. An automated planner then decides how to best realize the VDBEs onto physical engines based on the workload. Clients connect to VDBE endpoints and are oblivious to the underlying physical engines-allowing for seamless infrastructure changes. We implemented VDBEs and an automated planner in BRAD: the first data infrastructure virtualization runtime. Our demo will showcase VDBEs and BRAD's automated planner under different workloads. SIGMOD-Companion ’25, Berlin, Germany

An Efficient Subcritical Multiplication Mode for Monte Carlo Solvers

2026-02-11T04:36:46Z

An Efficient Subcritical Multiplication Mode for Monte Carlo Solvers Forget, Benoit This paper presents an efficient Monte Carlo mode for simulating subcritical systems with external sources. While solving these systems as a fixed source is possible, the length of the histories grows significantly as the system nears criticality, making run time significant. Instead, a hybrid method is proposed that leverages the traditional eigensolver while including elements of the external source. The method builds on prior work,but proposes an approach that maintains the size of the source bank and also provides a natural way of scaling tallies with the true multiplication factor. The method is demonstrated on a subcritical sphere with varying point source position and energy spectrum, as well as an approach to criticality problem. The results demonstrate good agreement with the fixed-source mode, with much improved particle tracking rates for near-critical problems.

Afterword: Reflections from Afar, with Hope for our Collective Future

2026-02-11T04:36:44Z

Afterword: Reflections from Afar, with Hope for our Collective Future Henderson, Diana E Appearing in the wake of a decade of rapid growth in Indian screen Shakespeare as an academic subspeciality, ‘Adapting Shakespearean Romance in Indian Cinema’ reveals how cross-cultural comparison and attention to popular reception can profitably modify inherited critical assumptions for all Shakespeare’s readership. Taking ‘romance’ as a key term, this afterword considers the possibilities and potential problems of recasting its dominant meaning as thematic, focusing on modern love, rather than as a dramatic subgenre. In a time of increasing political censorship and existential threats to gender studies, greater engagement and exchange between those in other areas of Shakespeare studies with this rich cinematic corpus and its aligned subfield of cross-disciplinary criticism provides reasons for hope and renewed community.

Resolving the Contested Future of the GSEs: The Stakes Are High

2026-02-11T04:36:37Z

Resolving the Contested Future of the GSEs: The Stakes Are High Golding, Edward; Wachter, Susan Seventeen years after entering conservatorship, Fannie Mae and Freddie Mac remain central to the future of U.S. housing finance. This paper evaluates the feasibility of their exit from conservatorship without Congressional action, assessing repayment of the federal bailout, capital adequacy under current regulatory frameworks, and the durability of structural reforms. It puts forth a utility model that preserves liquidity, affordability, and mission alignment while mitigating risks of increased mortgage costs. Treasury mechanisms—including commitment fees and stock warrant monetization—are examined as tools to support affordable housing and fulfill charter mandates. A carefully structured exit, supported by robust oversight and capital standards, can balance adequate financial returns with public purpose. A regulatory framework that maintains stable lending standards and pricing over the business cycle is essential to reducing investor-required returns and enhancing affordability, thereby resolving the contested future of the GSEs.

Stabilizing far-from-equilibrium (Mo,Ti)S2 thin films by metal sulfurization at reduced temperature

2026-02-11T04:36:42Z

Stabilizing far-from-equilibrium (Mo,Ti)S2 thin films by metal sulfurization at reduced temperature Li, Yifei; Reidy, Kate; Penn, Aubrey; Lee, Seng Huat; Wang, Baoming; Ye, Kevin; Mao, Zhiqiang; Ross, Frances M; Jaramillo, R We report the synthesis of large-area, high-Ti-content, Mo1−xTixS2 alloy thin films in the 2H phase at temperature as low as 500 °C using a scalable two-step method of metal film deposition, followed by sulfurization in H2S. Film processing at higher temperature accelerates Ti segregation, film coarsening, and the formation of TiS2 in the 1T phase. Crystal growth at higher temperature results in the formation of multiple binary sulfide phases, in agreement with the equilibrium phase diagram. Making highly metastable, smooth, and uniform single-phase alloy films, therefore, hinges on developing low-temperature processing. Our results are relevant to the development of technologies based on designer transition metal dichalcogenide alloys, including in photonic integrated circuits and gas sensing.

Experimental study of lower hybrid wave power absorption on EAST

2026-02-11T04:36:40Z

Experimental study of lower hybrid wave power absorption on EAST Baek, S-G; Li, MH; Bonoli, PT; Ding, BJ; Wallace, GM; Chen, JL; Duan, YM; Gong, XZ; Qian, JP; Wang, L; Yang, H; Zang, Q; Zhang, JY; Zhang, XJ Lower hybrid power absorption analysis is presented on the EAST high-density plasmas using the power modulation technique. The change in the plasma and magnetic energy is monitored to evaluate the power absorption coefficient by linearizing the change for the first 10 msec for the given input power. The power absorption coefficient evaluated is approximately 0.44 (0.35) for 4.6 GHz (2.45 GHz) at n̄e = 3.5x1019 m-3 GENRAY/CQL3D current drive modeling suggests a combination of antenna spectrum, accessibility, and edge losses could primarily be responsible for the observed level of power absorption. Evidence of first-pass parasite LH power flow causing impurity sputtering is also reported, suggesting a need for optimum power coupling. Implications of the experimental findings are discussed.

Limiting role of dislocations in high-current AlGaN/GaN hot electron transistors

2026-02-11T04:36:38Z

Limiting role of dislocations in high-current AlGaN/GaN hot electron transistors Daulton, J. W.; Molnar, R. J.; Brinkerhoff, J. A.; Weir, T. J.; Hollis, M. A.; Zaslavsky, A. III-nitride-based hot electron transistors (HETs) hold significant promise as high-speed, high-power devices. In our previous work, we demonstrated high current density and common-emitter gain at room temperature. Here, we measure multiple devices at cryogenic temperatures, extending the Gummel characteristics past the onset of intervalley scattering at 77 K. We demonstrate a Gummel current gain of 4.7 at a collector current density of 2.6 MA/cm2 at 77 K as well as a peak current density exceeding 3 MA/cm2. From these data, we determine that dislocation-associated inhomogeneities play a limiting role in AlGaN/GaN HETs, controlling the current gain, density, knee voltage, and base-collector leakage. A comparison of two nominally identical devices suggests that even a modest reduction in dislocation density would result in a substantial improvement in HET performance.

SeerCuts: Explainable Attribute Discretization

2026-02-10T03:07:40Z

SeerCuts: Explainable Attribute Discretization Lai, Eugenie; Croitoru, Inbal; Bitton, Noam; Shalem, Ariel; Youngmann, Brit; Galhotra, Sainyam; Rezig, El Kindi; Cafarella, Michael This demonstration showcases SeerCuts - a tool that suggests useful and semantically meaningful discretization strategies (partitions) for numerical attributes. SeerCuts is a generic, interactive framework where users specify attributes to discretize and their utility measure for a downstream task of choice. It uses GPT-4o to assess the semantic meaningfulness of candidate partitions and employs an efficient search strategy to explore the vast space of discretization options. With hierarchical clustering to group related partitions and a multi-armed bandit policy to identify useful partitions with only a few samples, SeerCuts quickly finds meaningful and useful partitions. In the demo, we will provide an overview of SeerCuts and allow the audience to explore various datasets and tasks, including data visualization and comprehensive modeling. The users will be able to evaluate how SeerCuts identifies meaningful discretization strategies and compare the tradeoff between different discretization options. SIGMOD-Companion ’25, Berlin, Germany

PalimpChat: Declarative and Interactive AI analytics

2026-02-10T03:07:41Z

PalimpChat: Declarative and Interactive AI analytics Liu, Chunwei; Vitagliano, Gerardo; Rose, Brandon; Printz, Matthew; Samson, David Andrew; Cafarella, Michael Thanks to the advances in generative architectures and large language models, data scientists can now code pipelines of AI operations to process large collections of unstructured data. Recent progress has seen the rise of declarative AI frameworks (e.g., Palimpzest, Lotus, and DocETL) to build optimized and increasingly complex pipelines, but these systems often remain accessible only to expert programmers. In this demonstration, we present PalimpChat, a chat-based interface to Palimpzest that bridges this gap by letting users create and run sophisticated AI pipelines through natural language alone. By integrating Archytas, a ReAct-based reasoning agent, and Palimpzest's suite of relational and LLM-based operators, PalimpChat provides a practical illustration of how a chat interface can make declarative AI frameworks truly accessible to non-experts. Our demo system is publicly available online. At SIGMOD'25, participants can explore three real-world scenarios-scientific discovery, legal discovery, and real estate search-or apply PalimpChat to their own datasets. In this paper, we focus on how PalimpChat, supported by the Palimpzest optimizer, simplifies complex AI workflows such as extracting and analyzing biomedical data. SIGMOD-Companion ’25, Berlin, Germany

CauSumX: Summarized Causal Explanations For Group-By-Average Queries

2026-02-10T03:07:39Z

CauSumX: Summarized Causal Explanations For Group-By-Average Queries Levy, Nativ; Cafarella, Michael; Gilad, Amir; Roy, Sudeepa; Youngmann, Brit Group-by-average SQL queries are a cornerstone of data analysis, often employed to uncover patterns and trends within datasets. However, interpreting the results of these queries can be challenging and time-intensive, particularly when working with large, high-dimensional datasets. Automating the generation of explanations for such queries can greatly enhance analysts' ability to derive meaningful insights while reducing human effort. Effective explanations must balance succinctness and depth, offering insights into different patterns across aggregate results, while crucially reflecting cause-effect relationships rather than mere correlations. This ensures that users can make informed, data-driven decisions grounded in reality. In this demonstration, we present CauSumX, a system that produces concise and causal explanations for group-by-average queries. Leveraging background causal knowledge, CauSumX identifies the key causal factors driving variations in the outcome variable across different groups. The system employs an efficient algorithm based on a recently published paper. We will demonstrate the utility of CauSumX for generating useful summarized causal explanations by interacting with the SIGMOD'25 participants, who will act as data analysts aiming to explain their query results. SIGMOD-Companion ’25, Berlin, Germany

CausaLens: A System for Summarizing Causal DAGs

2026-02-10T03:07:04Z

CausaLens: A System for Summarizing Causal DAGs Chen, Noam; Zeng, Anna; Cafarella, Michael; Kenig, Batya; Markakis, Markos; Mishali, Oren; Youngmann, Brit; Salimi, Babak Causal inference aids researchers in discovering causal relationships, leading to scientific insights. Pearl's causal model uses causal DAGs to estimate causal effects, so DAG correctness is essential for reliable causal conclusions. However, for high-dimensional data, the causal DAGs are often complex beyond human verifiability. Graph summarization is a logical next step, but current methods for general-purpose graph summarization are inadequate for causal DAG summarization, as they are not designed to preserve causal information. In this demonstration, we present a system called CausaLens that summarizes a given causal DAG and balances graph simplification for better understanding and retention of essential causal information for reliable inference directly on the summary DAG. We illustrate that causal inference on the summary DAG is more robust to misspecification in the initial causal DAG compared to performing inference directly on the initial causal DAG, thereby enhancing the robustness of causal inference. We will demonstrate the utility of CausaLens for generating useful summary causal DAGs by interacting with the SIGMOD'25 participants, who will act as data analysts aiming to perform causal analysis on high dimensional datasets. SIGMOD-Companion ’25, Berlin, Germany

First Workshop on Novel Optimizations for Visionary AI Systems (NOVAS)

2026-02-10T03:06:46Z

First Workshop on Novel Optimizations for Visionary AI Systems (NOVAS) Vitagliano, Gerardo; Liu, Chunwei; Cao, Lei; Sun, Huan; Papotti, Paolo The first NOVAS workshop (Novel Optimizations for Visionary AI Systems) is aimed at hosting novel work at the intersection between artificial intelligence and data management. This area has emerged with the rise of transformer-based architectures, which have revolutionized data processing across modalities. While these models benefit from massive pre-training and large-context inference, there are significant challenges related to scalability, determinism, and resource constraints. These issues-long studied in the data management community-have sparked a convergence between generative AI and traditional database research. The workshop will be held on June 22nd, in conjunction with SIGMOD/PODS 2025. The workshop solicits regular and short papers on topics including hardware and execution optimizations, high-level programming abstractions, integration of LLMs with relational databases, and new transformer architectures for structured data. By bridging together the different communities of machine learning, data systems, and information retrieval, NOVAS aims at becoming the venue to discuss, share ideas and early results, and spark new research collaborations for the next-generation of data-driven AI systems. SIGMOD-Companion ’25, Berlin, Germany

Data-Efficient Discovery of Hyperelastic TPMS Metamaterials with Extreme Energy Dissipation

2026-02-10T03:07:38Z

Data-Efficient Discovery of Hyperelastic TPMS Metamaterials with Extreme Energy Dissipation Perroni-Scharf, Maxine; Ferguson, Zachary; Butruille, Thomas; Portela, Carlos; Konakovi? Lukovi?, Mina Triply periodic minimal surfaces (TPMS) are a class of metamaterials with a variety of applications and well-known primitive morphologies. We present a new method for discovering novel microscale TPMS structures with exceptional energy-dissipation capabilities, achieving double the energy absorption of the best existing TPMS primitive structure. Our approach employs a parametric representation, allowing seamless interpolation between structures and representing a rich TPMS design space. As simulations are intractable for efficiently optimizing microscale hyperelastic structures, we propose a sample-efficient computational strategy for rapid discovery with limited empirical data from 3D-printed and tested samples that ensures high-fidelity results. We achieve this by leveraging a predictive uncertainty-aware Deep Ensembles model to identify which structures to fabricate and test next. We iteratively refine our model through batch Bayesian optimization, selecting structures for fabrication that maximize exploration of the performance space and exploitation of our energy-dissipation objective. Using our method, we produce the first open-source dataset of hyperelastic microscale TPMS structures, including a set of novel structures that demonstrate extreme energy dissipation capabilities, and show several potential applications of these structures. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

Splat and Replace: 3D Reconstruction with Repetitive Elements

2026-02-10T03:07:03Z

Splat and Replace: 3D Reconstruction with Repetitive Elements Violante, Nicolas; Meuleman, Andr?as; Gauthier, Alban; Durand, Fredo; Groueix, Thibault; Drettakis, George We leverage repetitive elements in 3D scenes to improve novel view synthesis. Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have greatly improved novel view synthesis but renderings of unseen and occluded parts remain low-quality if the training views are not exhaustive enough. Our key observation is that our environment is often full of repetitive elements. We propose to leverage those repetitions to improve the reconstruction of low-quality parts of the scene due to poor coverage and occlusions. We propose a method that segments each repeated instance in a 3DGS reconstruction, registers them together, and allows information to be shared among instances. Our method improves the geometry while also accounting for appearance variations across instances. We demonstrate our method on a variety of synthetic and real scenes with typical repetitive elements, leading to a substantial improvement in the quality of novel view synthesis. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

Variational Elastodynamic Simulation

2026-02-10T03:07:26Z

Variational Elastodynamic Simulation Mattos Da Silva, Leticia; Sell?n, Silvia; Pacheco-Tallaj, Natalia; Solomon, Justin Numerical schemes for time integration are the cornerstone of dynamical simulations for deformable solids. The most popular time integrators for isotropic distortion energies rely on nonlinear root-finding solvers, most prominently, Newton’s method. These solvers are computationally expensive and sensitive to ill-conditioned Hessians and poor initial guesses; these difficulties can particularly hamper the effectiveness of variational integrators, whose momentum conservation properties require reliable root-finding. To tackle these difficulties, this paper shows how to express variational time integration for a large class of elastic energies as an optimization problem with a “hidden” convex substructure. This hidden convexity suggests uses of optimization techniques with rigorous convergence analysis, guaranteed inversion-free elements, and conservation of physical invariants up to tolerance/numerical precision. In particular, we propose an Alternating Direction Method of Multipliers (ADMM) algorithm combined with a proximal operator step to solve our formulation. Empirically, our integrator improves the performance of elastic simulation tasks, as we demonstrate in a number of examples. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

Strain-Tunable Thermal Conductivity in Largely Amorphous Polyolefin Fibers via Alignment-Induced Vibrational Delocalization

2026-02-10T03:08:05Z

Strain-Tunable Thermal Conductivity in Largely Amorphous Polyolefin Fibers via Alignment-Induced Vibrational Delocalization Developing fast, reversible, and recyclable thermal switches is essential to advance adaptive thermal management. Here, we present a strain-tunable thermal switch based on largely amorphous olefin block copolymer (OBC) fibers, achieving a continuous switching ratio above 2 over 1000 cycles, as well as very short response times below 0.22 s. Using Raman spectroscopy, we quantify vibrational delocalization with increasing strain and demonstrate its direct connection to the observed thermal conductivity changes. We show that unlike prior assumptions linking propagating heat carriers primarily to crystalline domains, alignment in amorphous systems can enable phonon-like modes that dominate transport. To our best knowledge, this work is the first to experimentally probe vibrational delocalization using Raman spectroscopy and to demonstrate that alignment alone can govern the dominant carrier in disordered polymers. These findings establish design strategies for fatigue-resistant, high-performance, and recyclable polymer thermal switches for advanced thermal energy transport applications.

Introducing synchromodality: One missing link between transportation and supply chain management

2026-02-10T03:07:55Z

Introducing synchromodality: One missing link between transportation and supply chain management Acero, Beatriz; Saenz, Maria Jesus; Luzzini, Davide This study develops and tests the synchromodality construct, a novel supply chain concept that integrates the flexible use of different transport modes based on real-time information. At a time when global supply chains are complex and subject to uncertainty, synchromodality has emerged at the forefront of research and practice as a tool to ensure efficient delivery performance and thus supply chain competitiveness. Despite synchromodality is attracting the attention of leading companies and policy makers, only scholars within the transport research community have engaged with the topic so far. We believe a supply chain management perspective is missing, but essential, to develop the full potential of synchromodality. Our study shows that synchromodality capabilities encapsulate four key elements: visibility, integration, multi-modal transport, and flexibility. Thanks to a three-stage research approach exploiting multiple methods, this study conceptualizes, develops, and validates the first synchromodality measurement model, which reflects the multidimensional nature of the concept. We hope to set the stage for a number of potential future research opportunities that can explore synchromodality implementation and outcomes.

Shape Space Spectra

2026-02-10T03:07:07Z

Shape Space Spectra Chang, Yue; Benchekroun, Otman; Chiaramonte, Maurizio M.; Chen, Peter Yichen; Grinspun, Eitan Eigenanalysis of differential operators, such as the Laplace operator or elastic energy Hessian, is typically restricted to a single shape and its discretization, limiting reduced order modeling (ROM). We introduce the first eigenanalysis method for continuously parameterized shape families. Given a parametric shape, our method constructs spatial neural fields that represent eigen-functions across the entire shape space. It is agnostic to the specific shape representation, requiring only an inside/outside indicator function that depends on shape parameters. Eigenfunctions are computed by minimizing a variational principle over nested spaces with orthogonality constraints. Since eigenvalues may swap dominance at points of multiplicity, we jointly train multiple eigenfunctions while dynamically reordering them based on their eigenvalues at each step. Through causal gradient filtering, this reordering is reflected in backpropagation. Our method enables applications to operate over shape space, providing a single ROM that encapsulates vibration modes for all shapes, including previously unseen ones. Since our eigenanalysis is differentiable with respect to shape parameters, it facilitates eigenfunction-aware shape optimization. We evaluate our approach on shape optimization for sound synthesis and locomotion, as well as reduced-order modeling for elastodynamic simulation.

Dynamic Mesh Processing on the GPU

2026-03-08T03:23:09Z

Dynamic Mesh Processing on the GPU Mahmoud, Ahmed H.; Porumbescu, Serban D.; Owens, John D. We present a system for dynamic triangle mesh processing entirely on the GPU. Our system features an efficient data structure that enables rapid updates to mesh connectivity and attributes. By partitioning the mesh into small patches, we process all dynamic updates for each patch within the GPU's fast shared memory. This approach leverages speculative processing for conflict handling, minimizing rollback costs, maximizing parallelism, and reducing locking overhead. Additionally, we introduce a new programming model for dynamic mesh processing. This model provides concise semantics for dynamic updates, abstracting away concerns about conflicting updates during parallel execution. At the core of our model is the cavity operator, a general mesh update operator that facilitates any dynamic operation by removing a set of mesh elements and inserting new ones into the resulting void. We applied our system to various GPU applications, including isotropic remeshing, surface tracking, mesh decimation, and Delaunay edge flips. On large inputs, our system achieves an order-of-magnitude speedup compared to multi-threaded CPU solutions and is more than two orders of magnitude faster than state-of-the-art single-threaded CPU solutions. Furthermore, our data structure outperforms state-of-the-art GPU static data structures in terms of both speed and memory efficiency.

Exciton Fine Structure in 2D Perovskites: The Out‐of‐Plane Excitonic State

2026-03-08T03:39:49Z

Exciton Fine Structure in 2D Perovskites: The Out‐of‐Plane Excitonic State Posmyk, Katarzyna; Dyksik, Mateusz; Surrente, Alessandro; Maude, Duncan K; Zawadzka, Natalia; Babiński, Adam; Molas, Maciej R; Paritmongkol, Watcharaphol; Mączka, Mirosław; Tisdale, William A; Plochocka, Paulina; Baranowski, Michał 2D Ruddlesden-Popper metal-halide perovskites feature particularly strong excitonic effects, making them a fascinating playground for studying exciton physics. A complete understanding of the properties of this quasi-particle is crucial to fully exploit the tremendous potential of 2D perovskites (2DP) in light emission applications. Despite intense investigations, some of the exciton properties remain elusive to date, for example, the energy-ordering of the exciton states within the so-called fine structure manifold. Using optical spectroscopy, it demonstrates that in the archetypical 2DP (PEA)2PbI4, in contradiction to theoretical predictions, the energy of the bright out-of-plane exciton state is higher than that of two in-plane states. Having elucidated the order of exciton fine structure, it determines the g-factor of the dark exciton transition, together with the values of the electron and hole g-factors in the direction parallel to the c-axis of the crystal. In this way, it provides for the first time, a complete picture of the exciton fine structure in (PEA)2PbI4 2DP.

Discovery of enhanced lattice dynamics in a single-layered hybrid perovskite

2026-03-08T03:39:47Z

Discovery of enhanced lattice dynamics in a single-layered hybrid perovskite Zhang, Zhuquan; Zhang, Jiahao; Liu, Zi-Jie; Dahod, Nabeel S; Paritmongkol, Watcharaphol; Brown, Niamh; Stollmann, Alexia; Lee, Woo Seok; Chien, Yu-Che; Dai, Zhenbang; Nelson, Keith A; Tisdale, William A; Rappe, Andrew M; Baldini, Edoardo Layered hybrid perovskites exhibit emergent physical properties and exceptional functional performances, but the coexistence of lattice order and structural disorder severely hinders our understanding of these materials. One unsolved problem regards how the lattice dynamics are affected by the dimensional engineering of the inorganic frameworks and their interaction with the molecular moieties. Here, we address this question by using a combination of spontaneous Raman scattering, terahertz spectroscopy, and molecular dynamics simulations. This approach reveals the structural dynamics in and out of equilibrium and provides unexpected observables that differentiate single- and double-layered perovskites. While no distinct vibrational coherence is observed in double-layered perovskites, an off-resonant terahertz pulse can drive a long-lived coherent phonon mode in the single-layered system. This difference highlights the dramatic change in the lattice environment as the dimension is reduced, and the findings pave the way for ultrafast structural engineering and high-speed optical modulators based on layered perovskites.

Bright Excitonic Fine Structure in Metal-Halide Perovskites: From Two-Dimensional to Bulk

2026-03-08T03:40:20Z

Bright Excitonic Fine Structure in Metal-Halide Perovskites: From Two-Dimensional to Bulk Posmyk, Katarzyna; Zawadzka, Natalia; Łucja Kipczak; Dyksik, Mateusz; Surrente, Alessandro; Maude, Duncan K; Kazimierczuk, Tomasz; Babiński, Adam; Molas, Maciej R; Bumrungsan, Wakul; Chooseng, Chanisara; Paritmongkol, Watcharaphol; Tisdale, William A; Baranowski, Michał; Plochocka, Paulina The optical response of two-dimensional (2D) perovskites, often referred to as natural quantum wells, is primarily governed by excitons, whose properties can be readily tuned by adjusting the perovskite layer thickness. We have investigated the exciton fine structure splitting in the archetypal 2D perovskite (PEA)2(MA)n−1PbnI3n+1 with varying numbers of inorganic octahedral layers n = 1, 2, 3, and 4. We demonstrate that the in-plane excitonic states exhibit splitting and orthogonally oriented dipoles for all confinement regimes. The evolution of the exciton states in an external magnetic field provides further insights into the g-factors and diamagnetic coefficients. With increasing n, we observe a gradual evolution of the excitonic parameters characteristic of a 2D to three-dimensional transition. Our results provide valuable information concerning the evolution of the optoelectronic properties of 2D perovskites with the changing confinement strength.

Persistent enhancement of exciton diffusivity in CsPbBr3 nanocrystal solids

2026-03-08T03:39:47Z

Persistent enhancement of exciton diffusivity in CsPbBr3 nanocrystal solids Shcherbakov-Wu, Wenbi; Saris, Seryio; Sheehan, Thomas John; Wong, Narumi Nagaya; Powers, Eric R; Krieg, Franziska; Kovalenko, Maksym V; Willard, Adam P; Tisdale, William A In semiconductors, exciton or charge carrier diffusivity is typically described as an inherent material property. Here, we show that the transport of excitons among CsPbBr3 perovskite nanocrystals (NCs) depends markedly on how recently those NCs were occupied by a previous exciton. Using transient photoluminescence microscopy, we observe a striking dependence of the apparent exciton diffusivity on excitation laser power that does not arise from nonlinear exciton-exciton interactions or thermal heating. We interpret our observations with a model in which excitons cause NCs to transition to a long-lived metastable configuration that markedly increases exciton transport. The exciton diffusivity observed here (>0.15 square centimeters per second) is considerably higher than that observed in other NC systems, revealing unusually strong excitonic coupling between NCs. The finding of a persistent enhancement in excitonic coupling may help explain other photophysical behaviors observed in CsPbBr3 NCs, such as superfluorescence, and inform the design of optoelectronic devices.

All-Perovskite Multicomponent Nanocrystal Superlattices

2026-03-08T03:39:46Z

All-Perovskite Multicomponent Nanocrystal Superlattices Sekh, Taras V; Cherniukh, Ihor; Kobiyama, Etsuki; Sheehan, Thomas J; Manoli, Andreas; Zhu, Chenglian; Athanasiou, Modestos; Sergides, Marios; Ortikova, Oleksandra; Rossell, Marta D; Bertolotti, Federica; Guagliardi, Antonietta; Masciocchi, Norberto; Erni, Rolf; Othonos, Andreas; Itskos, Grigorios; Tisdale, William A; Stöferle, Thilo; Rainò, Gabriele; Bodnarchuk, Maryna I; Kovalenko, Maksym V Nanocrystal superlattices (NC SLs) have long been sought as promising metamaterials, with nanoscale-engineered properties arising from collective and synergistic effects among the constituent building blocks. Lead halide perovskite (LHP) NCs come across as outstanding candidates for SL design, as they demonstrate collective light emission, known as superfluorescence, in single- and multicomponent SLs. Thus far, LHP NCs have only been assembled in single-component SLs or coassembled with dielectric NC building blocks acting solely as spacers between luminescent NCs. Here, we report the formation of multicomponent LHP NC-only SLs, i.e., using only CsPbBr3 NCs of different sizes as building blocks. The structural diversity of the obtained SLs encompasses the ABO6, ABO3, and NaCl structure types, all of which contain orientationally and positionally locked NCs. For the selected model system, the ABO6-type SL, we observed efficient NC coupling and Förster-like energy transfer from strongly confined 5.3 nm CsPbBr3 NCs to weakly confined 17.6 nm CsPbBr3 NCs, along with characteristic superfluorescence features at cryogenic temperatures. Spatiotemporal exciton dynamics measurements reveal that binary SLs exhibit enhanced exciton diffusivity compared to single-component NC assemblies across the entire temperature range (from 5 to 298 K). The observed coherent and incoherent NC coupling and controllable excitonic transport within the solid NC SLs hold promise for applications in quantum optoelectronic devices.

Electrical manipulation of dissipation in microwave photon–magnon hybrid system through the spin Hall effect

2026-03-08T03:40:19Z

Electrical manipulation of dissipation in microwave photon–magnon hybrid system through the spin Hall effect Hou, Justin T; Chou, Chung-Tao; Han, Jiahao; Fan, Yabin; Liu, Luqiao Hybrid dynamic systems combine advantages from different subsystems for realizing information processing tasks in both classical and quantum domains. However, the lack of controlling knobs in tuning system parameters becomes a severe challenge in developing scalable, versatile hybrid systems for useful applications. Here, we report an on-chip microwave photon–magnon hybrid system where the dissipation rates and the coupling cooperativity can be electrically influenced by the spin Hall effect. Through magnon–photon coupling, the linewidths of the resonator photon mode and the hybridized magnon polariton modes are effectively changed by the spin injection into the magnetic wires from an applied direct current, which exhibit different trends in samples with low and high coupling strengths. Moreover, the linewidth modification by the spin Hall effect shows strong dependence on the detuning of the two subsystems, in contrast to the classical behavior of a standalone magnonic device. Our results point to a direction of realizing tunable, on-chip, scalable magnon-based hybrid dynamic systems, where spintronic effects provide useful control mechanisms.

Human Factors Observations in Flightcrew Response to System Failure Events in Transport Category Aircraft from 2000 to 2024

2026-02-06T03:00:48Z

Human Factors Observations in Flightcrew Response to System Failure Events in Transport Category Aircraft from 2000 to 2024 Perez Gago, Cecilia; Hansman, R. John

When Will (Game) Wars End?

2026-03-08T03:40:17Z

When Will (Game) Wars End? Bhatia, Manan; Chin, Byron; Mani, Nitya; Mossel, Elchanan We study several variants of the classical card game war. As anyone who has played this game knows, the game can take some time to terminate, but it usually does. Here, we analyze a number of asymptotic variants of the game, where the number of cards is n, and show that all have an expected termination time of order 𝑛2. This is the same expected termination time as the game where at each turn a fair coin toss decides which player wins a card, known as Gambler’s Ruin and was studied by Blaise Pascal, Pierre de Fermat and others in the seventeenth century.

Semiconductor-free, monolithically 3D-printed logic gates and resettable fuses

2026-03-08T03:40:18Z

Semiconductor-free, monolithically 3D-printed logic gates and resettable fuses Cañada, Jorge; Velásquez-García, Luis Fernando Additive manufacturing has the potential to enable the inexpensive, single-step fabrication of fully functional electromechanical devices. However, while the 3D printing of mechanical parts and passive electrical components is well developed, the fabrication of fully 3D-printed active electronics, which are the cornerstone of intelligent devices, remains a challenge. Existing examples of 3D-printed active electronics show potential but lack integrability and accessibility. This work reports the first active electronics fully 3D-printed via material extrusion, i.e. one of the most accessible and versatile additive manufacturing processes. The technology is proof-of-concept demonstrated through the implementation of the first fully 3D-printed, semiconductor-free, solid-state logic gates, and the first fully 3D-printed resettable fuses. The devices take advantage of a positive temperature coefficient phenomenon found to affect narrow traces of 3D-printed copper-reinforced, polylactic acid. Although the reported devices don’t perform competitively against semiconductor-enabled integrated circuits, the customisability and accessibility intrinsic to material extrusion additive manufacturing make this technology promisingly disruptive. This work serves as a steppingstone for the semiconductor-free democratisation of electronic device fabrication and is of immediate relevance for the manufacture of custom, intelligent devices far from traditional manufacturing centres.

Rapid large-scale building damage level classification after earthquakes using deep learning with Lidar and satellite optical data

2026-03-08T03:40:21Z

Rapid large-scale building damage level classification after earthquakes using deep learning with Lidar and satellite optical data Liu, Chang; Ge, Linlin; Bai, Ting In post-earthquake scenarios, the swift assessment of building damage levels is pivotal for efficient emergency response and recovery planning. Nevertheless, conventional in-situ damage evaluations consume time. Current satellite-based deep learning methods save time but often lack detail, usually classifying damage as either collapsed or intact. This two-level information is not enough for rescue or recovery planning. Light Detection and Ranging (Lidar)-based deep learning methods, which provide three-dimensional (3D) information, could address this issue of damage details. Therefore, this paper proposes a deep learning-based building damage level classification method using both Lidar and satellite data. The proposed method classifies damage into four levels, including no/minor damage, partially collapsed, totally collapsed, and story failure. The developed network builds upon RandLA-Net, incorporating surface normal vectors to enhance accuracy. A colourised Lidar dataset was created for the network. The network underscores the advantage of incorporating surface normal information. A framework is also proposed based on the damage level outcomes of the developed network, which aids in emergency response efforts. Consequently, this paper demonstrates the practical utility of deep learning networks in rapidly assessing detailed building damage levels after earthquakes. Its practical contribution is guiding decision-making during the critical phases of post-earthquake response and recovery.

Cortical somatostatin innervation follows a unique experience-independent developmental trajectory

2026-03-08T03:39:45Z

Cortical somatostatin innervation follows a unique experience-independent developmental trajectory Boivin, Josiah R; Schmerl, Bettina; Martin, Kendyll B; Lee, Chia-Fang; Nedivi, Elly Despite the critical role of inhibition in regulating developmental plasticity, there are significant gaps in our understanding of inhibitory synapse development, particularly for the vast majority of inhibitory synapses that reside on dendrites. Dendritic inhibitory synapses, canonically arising from somatostatin (SST)-expressing neurons, are challenging to detect electrophysiologically and difficult to visualize without a molecular tag. Here, we integrate a genetic synapse labeling strategy with epitope-preserving magnified analysis of proteome (eMAP), a combination of tissue expansion and clearing, to reveal the development of SST innervation in the primary visual cortex of male and female mice. Unlike excitatory innervation, which follows a deep to shallow progression and undergoes pruning, we find that SST bouton formation occurs simultaneously across all cortical layers and is not subject to a period of net pruning. SST bouton and synapse formation occur most dramatically in the days following eye opening and during the opening of the critical period for ocular dominance plasticity. Yet, despite a coincidence with these visual milestones, neither SST bouton nor synapse formation depend on visual experience. This is in contrast to excitatory and non-SST inhibitory synapses, whose development has been shown to depend heavily on visual experience. Thus, SST cortical innervation follows a unique developmental trajectory that is independent of sensory experience and is optimally timed to regulate processes that are fundamental to cortical circuit maturation. Significance statement During development, neurons form extensive synaptic connections while maintaining a delicate balance of excitation and inhibition. It is critical to understand how different subpopulations of synapses form during development, because perturbations in this precisely coordinated process can cause neurodevelopmental disorders. Here, we reveal at unprecedented resolution the development of cortical inhibitory innervation from somatostatin-expressing neurons, which canonically target dendrites. We show that somatostatin neurons follow different rules than other cell types during development, and somatostatin innervation is well-timed to contribute to developmental processes that are central to healthy cortical function. Our results provide new insights on how somatostatin neurons, a critically influential cell type, integrate into cortical circuitry during development.

Neruda through copper-coloured glasses: the role of place attachment in the embeddedness of Chilean entrepreneurship

2026-03-08T03:40:17Z

Neruda through copper-coloured glasses: the role of place attachment in the embeddedness of Chilean entrepreneurship Burke, M. Kathleen; Conley, Mark A.; Jack, Sarah L. Despite scholarly interest in how emotional and instrumental place attachments motivate entrepreneurship, the influences on embeddedness remain underexplored. Building on the notion that entrepreneurship becomes embedded in a locality, we argue that this process is packed with place-based interpretations of the material and imagined reality. Engaging with the empirical setting of Chile, the world’s largest copper producer, we embark on a study examining the interactions between the place attachment, embeddedness and natural resource-based entrepreneurship. We uncover these interactions through analysing several works of poetry by Nobel laureate Pablo Neruda, which focus on the diverging place attachment styles between local and multinational agents. Through reflecting on the poems, we show how historical changes within the Chilean mining industry and broader societal changes are visible in Neruda’s imagery of place attachments, emotions and concerns for local conditions. We problematize embeddedness and entrepreneurship through illuminating the place attachments shaping local actors’ entrepreneurial imagination, thus contributing to knowledge about being embedded in natural resource-based entrepreneurship contexts. We provide new insights into how place attachment can evolve alongside different forms of embedded entrepreneurship.

Human-centric manufacturing culture: a research study of MedTech manufacturers in Ireland

2026-03-08T03:40:19Z

Human-centric manufacturing culture: a research study of MedTech manufacturers in Ireland Rhodes, Donna H; Cuddy, Sara; Jeffers, Malcolm; O’Rourke, Fiona Digital manufacturing is rapidly evolving; however, this transformation is predominantly technology centric. Human-centric manufacturing shifts the paradigm for the digital manufacturing enterprise towards a human focus to realising its envisioned digital future. In that context, Digital Manufacturing Ireland (DMI), Ireland’s expert body for driving digital adoption across manufacturing, initiated a research study in collaboration with two research partners, MIT and IAAE, in support of this important focus for future manufacturing. This paper discusses results of the DMI 2023 human-Centric Manufacturing Culture Study, which engaged manufacturing leaders from 11 MedTech companies with major manufacturing sites in Ireland. Overall findings are discussed, with a focus on 12 emergent themes grouped in four categories: imperatives, values, strategies, and practices. Planned collaboration initiatives and anticipated future research are described. This paper also highlights considerations regarding new thinking needed by manufacturing leaders, along with recommendations as to what leaders can begin to do differently.

Decomposition of Frobenius pushforwards of line bundles on wonderful compactifications

2026-03-08T03:39:43Z

Decomposition of Frobenius pushforwards of line bundles on wonderful compactifications Cai, Merrick; Krylov, Vasily De Concini and Procesi introduced varieties known as wonderful compactifications, which are smooth projective compactifications of semisimple adjoint groups G. We study the Frobenius pushforwards of line bundles on the wonderful compactifications, and in particular we decompose them into a direct sum of vector subbundles and explicitly describe the ranks. We are especially interested in when these subbundles are line bundles, and in the case of 𝐺=𝖯𝖲𝖫𝑛, we offer lower bounds on the multiplicities (as direct summands) for these line bundles.

BrepDiff: Single-Stage B-rep Diffusion Model

2026-03-08T03:22:50Z

BrepDiff: Single-Stage B-rep Diffusion Model Lee, Mingi; Zhang, Dongsu; Jambon, Cl?ment; Kim, Young Min The Boundary Representation (B-rep) is a widely used 3D model representation of most consumer products designed with CAD software. However, its highly irregular and sparse set of relationships poses significant challenges for designing a generative model tailored to B-reps. Existing approaches use multi-stage approaches to satisfy the complex constraints sequentially. As a result, the final geometry cannot incorporate user edits due to the non-deterministic dependencies between cascaded stages. In contrast, we propose BrepDiff, a single-stage diffusion model for B-rep generation. We present a masked UV grid representation consisting of structured point samples from faces, serving as input for a diffusion transformer. By introducing an asynchronous and shifted noise schedule, we improve the training signal, enabling the diffusion model to better capture the distribution of UV grids. The explicitness of our masked UV grid representation enables users to intuitively understand and freely design surface geometry without being constrained by topological validity. The interconnectivity can be derived from the face layout, which is later processed into a valid solid volume during post-processing. Our approach achieves performance on par with state-of-the-art cascaded models while offering complex and diverse manipulations of geometry and topology, such as shape completion, merging, and interpolation. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

SwiftSketch: A Diffusion Model for Image-to-Vector Sketch Generation

2026-03-08T03:22:35Z

SwiftSketch: A Diffusion Model for Image-to-Vector Sketch Generation Arar, Ellie; Frenkel, Yarden; Cohen-Or, Daniel; Shamir, Ariel; Vinker, Yael Recent advancements in large vision-language models have enabled highly expressive and diverse vector sketch generation. However, state-of-the-art methods rely on a time-consuming optimization process involving repeated feedback from a pretrained model to determine stroke placement. Consequently, despite producing impressive sketches, these methods are limited in practical applications. In this work, we introduce SwiftSketch, a diffusion model for image-conditioned vector sketch generation that can produce high-quality sketches in less than a second. SwiftSketch operates by progressively denoising stroke control points sampled from a Gaussian distribution. Its transformer-decoder architecture is designed to effectively handle the discrete nature of vector representation and capture the inherent global dependencies between strokes. To train SwiftSketch, we construct a synthetic dataset of image-sketch pairs, addressing the limitations of existing sketch datasets, which are often created by non-artists and lack professional quality. For generating these synthetic sketches, we introduce ControlSketch, a method that enhances SDS-based techniques by incorporating precise spatial control through a depth-aware ControlNet. We demonstrate that SwiftSketch generalizes across diverse concepts, efficiently producing sketches that combine high fidelity with a natural and visually appealing style. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

Lifting the Winding Number: Precise Discontinuities in Neural Fields for Physics Simulation

2026-03-08T03:22:40Z

Lifting the Winding Number: Precise Discontinuities in Neural Fields for Physics Simulation Chang, Yue; Liu, Mengfei; Wang, Zhecheng; Chen, Peter Yichen; Grinspun, Eitan Cutting thin-walled deformable structures is common in daily life, but poses significant challenges for simulation due to the introduced spatial discontinuities. Traditional methods rely on mesh-based domain representations, which require frequent remeshing and refinement to accurately capture evolving discontinuities. These challenges are further compounded in reduced-space simulations, where the basis functions are inherently geometry- and mesh-dependent, making it difficult or even impossible for the basis to represent the diverse family of discontinuities introduced by cuts. Recent advances in representing basis functions with neural fields offer a promising alternative, leveraging their discretization-agnostic nature to represent deformations across varying geometries. However, the inherent continuity of neural fields is an obstruction to generalization, particularly if discontinuities are encoded in neural network weights. We present Wind Lifter, a novel neural representation designed to accurately model complex cuts in thin-walled deformable structures. Our approach constructs neural fields that reproduce discontinuities precisely at specified locations, without “baking in” the position of the cut line. To achieve this, we augment the input coordinates of the neural field with the generalized winding number of any given cut line, effectively lifting the input from two to three dimensions. Lifting allows the network to focus on the easier problem of learning a 3D everywhere-continuous volumetric field, while a corresponding restriction operator enables the final output field to precisely resolve strict discontinuities. Crucially, our approach does not embed the discontinuity in the neural network’s weights, opening avenues to generalization of cut placement. Our method achieves real-time simulation speeds and supports dynamic updates to cut line geometry during the simulation. Moreover, the explicit representation of discontinuities makes our neural field intuitive to control and edit, offering a significant advantage over traditional neural fields, where discontinuities are embedded within the network’s weights, and enabling new applications that rely on general cut placement. SIGGRAPH Conference Papers ’25, Vancouver, BC, Canada

Making Concurrent Hardware Verification Sequential

2026-03-08T03:39:50Z

Making Concurrent Hardware Verification Sequential Bourgeat, Thomas; Liu, Jiazheng; Chlipala, Adam; Arvind Compared to familiar hardware-description languages like Verilog, rule-based languages like Bluespec offer opportunities to import modularity features from software programming. While Verilog modules are about connecting wires between submodules, Bluespec modules resemble objects in object-oriented programming, where interactions with a module occur only through calls to its methods. However, while software objects can typically be characterized one method at a time, the concurrent nature of hardware makes it essential to consider the repercussions of invoking multiple methods simultaneously. Prior formalizations of rule-based languages conceptualized modules by describing their semantics considering arbitrary sets of simultaneous method calls. This internalized concurrency significantly complicates correctness proofs. Rather than analyzing methods one-at-a-time, as is done when verifying software object methods, validating the correctness of rule-based modules necessitated simultaneous consideration of arbitrary subsets of method calls. The result was a number of proof cases that grew exponentially in the size of the module’s API. In this work, we side-step the exponential blowup through a set of judicious language restrictions. We introduce a new Bluespec-inspired formal language, Fjfj, that supports sequential characterization of modules, while preserving the concurrent hardware nature of the language. We evaluated Fjfj by implementing it in Coq, proving the key framework principle: the refinement theorem. We demonstrated Fjfj’s expressivity via implementation and verification of three examples: a pipelined processor, a parameterized crossbar, and a network switch.

Lilo: A Higher-Order, Relational Concurrent Separation Logic for Liveness

2026-03-08T03:22:44Z

Lilo: A Higher-Order, Relational Concurrent Separation Logic for Liveness Lee, Dongjae; Lee, Janggun; Yoon, Taeyoung; Cho, Minki; Kang, Jeehoon; Hur, Chung-Kil Concurrent separation logic (CSL) has excelled in verifying safety properties across various applications, yet its application to liveness properties remains limited. While existing approaches like TaDA Live and Fair Operational Semantics (FOS) have made significant strides, they still face limitations. TaDA Live struggles to verify certain classes of programs, particularly concurrent objects with non-local linearization points, and lacks support for general liveness properties such as "good things happen infinitely often". On the other hand, FOS’s scalability is hindered by the absence of thread modular reasoning principles and modular specifications. This paper introduces Lilo, a higher-order, relational CSL designed to overcome these limitations. Our core observation is that FOS helps us to maintain simple primitives for our logic, which enable us to explore design space with fewer restrictions. As a result, Lilo adapts various successful techniques from literature. It supports reasoning about non-terminating programs by supporting refinement proofs, and also provides Iris-style invariants and modular specifications to facilitate modular verification. To support higher-order reasoning without relying on step-indexing, we develop a technique called stratified propositions inspired by Nola. In particular, we develop novel abstractions for liveness reasoning that bring these techniques together in a uniform way. We show Lilo’s scalability through case studies, including the first termination-guaranteeing modular verification of the elimination stack. Lilo and examples in this paper are mechanized in Coq.

Stochastic Lazy Knowledge Compilation for Inference in Discrete Probabilistic Programs

2026-03-08T03:22:51Z

Stochastic Lazy Knowledge Compilation for Inference in Discrete Probabilistic Programs Bowers, Maddy; Lew, Alexander K.; Tenenbaum, Joshua B.; Solar-Lezama, Armando; Mansinghka, Vikash K. We present new techniques for exact and approximate inference in discrete probabilistic programs, based on two new ways of exploiting lazy evaluation. First, we show how knowledge compilation, a state-of-the art technique for exact inference in discrete probabilistic programs, can be made lazy, enabling asymptotic speed-ups. Second, we show how a probabilistic program’s lazy semantics naturally give rise to a division of its random choices into subproblems, which can be solved in sequence by sequential Monte Carlo with locally-optimal proposals automatically computed via lazy knowledge compilation. We implement our approach in a new tool, Pluck, and evaluate its performance against state-of-the-art approaches to inference in discrete probabilistic languages. We find that on a suite of inference benchmarks, lazy knowledge compilation can be faster than state-of-the-art approaches, sometimes by orders of magnitude.

Probabilistic Programming with Vectorized Programmable Inference

2026-03-08T03:39:39Z

Probabilistic Programming with Vectorized Programmable Inference Becker, McCoy R.; Huot, Mathieu; Matheos, George; Wang, Xiaoyan; Chung, Karen; Smith, Colin; Ritchie, Sam; Saurous, Rif A.; Lew, Alexander K.; Rinard, Martin C.; Mansinghka, Vikash K. We present GenJAX, a new language and compiler for vectorized programmable probabilistic inference. GenJAX integrates the vectorizing map (vmap) operation from array programming frameworks such as JAX into the programmable inference paradigm, enabling compositional vectorization of features such as probabilistic program traces, stochastic branching (for expressing mixture models), and programmable inference interfaces for writing custom probabilistic inference algorithms. We formalize vectorization as a source-to-source program transformation on a core calculus for probabilistic programming ($\gen$), and prove that it correctly vectorizes both modeling and inference operations. We have implemented our approach in \href{https://github.com/probcomp/genjax}{the GenJAX language and compiler}, and have empirically evaluated this implementation on several benchmarks and case studies. Our results show that our implementation supports a wide and expressive set of programmable inference patterns and delivers performance comparable to hand-optimized JAX code.

Waste-Efficient Work Stealing

2026-03-08T03:39:38Z

Waste-Efficient Work Stealing Singer, Kyle; Agrawal, Kunal; Schardl, Tao B. Although randomized work stealing is effective at automatically load-balancing task-parallel programs, it can waste computational resources when scheduling programs that lack sufficient parallelism to use all available threads. For such programs, threads will waste cycles attempting to steal parallel tasks when none are available. This waste can reduce the machine’s efficiency by wasting computational resources and energy and needlessly burdening the operating system. This paper introduces WEWS, a simple, practical, and provably efficient extension to randomized work stealing that mitigates waste. WEWS dynamically adjusts the number of active threads to reduce the waste of randomized work stealing. WEWS executes a parallel computation with the same asymptotic running time as traditional randomized work stealing while bounding the waste to O(min{PT∞, T1 + P2}) instructions. WEWS also follows the work-first principle to perform well in practice. WEWS requires no special support from the operating system or hardware, which simplifies its implementation. We implemented WEWS within the OpenCilk runtime system and compared it to other common waste-mitigation strategies. Across 10 parallel benchmarks, we find that WEWS has minimal impact on parallel running times while, on programs with limited parallelism, substantially reducing waste. PPoPP ’26, Sydney, NSW, Australia

UniTe: A Universal Tensor Abstraction for Capturing Spatial Relationships

2026-03-08T03:39:41Z

UniTe: A Universal Tensor Abstraction for Capturing Spatial Relationships Ray, Jessica; Collin, Teodoro; Sze, Vivienne; Reuther, Albert; Amarasinghe, Saman Tensors are an integral part of numerous domains, and while significant effort has been put into the design of tensor data structures in isolation, little attention has been paid to the relationships that exist across tensors and how this affects their representation and use. In this paper, we focus on spatial relationships across tensors in a program, where such tensors are defined relative to a common reference coordinate system. These relationships are complicated by the fact that the tensors may differ in their representations, such as having variations in their axes, spacings, origins, and overall shape. Due to the lack of existing abstractions and language support for these types of tensor semantics, users are currently forced to manually perform the bookkeeping necessary to account for these varying relationships and representations. Unfortunately, we cannot rely on a simple library to capture these relationships, as computations on these types of tensors often happen at the innermost levels of programs; we find that the overheads associated with an unoptimized implementation quickly accumulate, leading to performance up to nearly 65x slower than a reference C implementation on a series of image and video compression benchmarks. In this paper, we introduce the novel UniTe abstraction, which captures spatial relationships across all such tensors in a program. We also introduce two domain-specific languages and optimizing compilers, CoLa for Python and SHiM for C/C++, built off of UniTe. Both CoLa and SHiM provide users an intuitive set of tensor primitives based on spatial relationships, hiding the complexity that goes into maintaining the tensors and computing accesses across them. In addition, we discuss the optimizations necessary to remove the associated abstraction overhead, and describe their implementations. On the benchmarks, we show that both CoLa and SHiM successfully remove the overheads, achieving performance parity with existing C implementations.

Triplet Exciton Sensitization of Silicon Mediated by Defect States in Hafnium Oxynitride

2026-03-08T03:39:56Z

Triplet Exciton Sensitization of Silicon Mediated by Defect States in Hafnium Oxynitride Nagaya, Narumi; Alexiu, Alexandra; Perkinson, Collin F; Nix, Oliver M; Koh, Dooyong; Bawendi, Moungi G; Tisdale, William A; Van Voorhis, Troy; Baldo, Marc A Singlet exciton fission has the potential to increase the efficiency of crystalline silicon solar cells beyond the conventional single junction limit. Perhaps the largest obstacle to achieving this enhancement is uncertainty about energy coupling mechanisms at the interfaces between silicon and exciton fission materials such as tetracene. Here, the previously reported silicon‐hafnium oxynitride‐tetracene structure is studied and a combination of magnetic‐field‐dependent silicon photoluminescence measurements and density functional theory calculations is used to probe the influence of the interlayer composition on the triplet transfer process across the hafnium oxynitride interlayer. It is found that hafnium oxide interlayers do not show triplet exciton sensitization of silicon, and that nitrogen content in hafnium oxynitride layers is correlated with enhanced sensitization. Calculation results reveal that defects in hafnium oxynitride interlayers with higher nitrogen content introduce states close to the band‐edge of silicon, which can mediate the triplet exciton transfer process. Some defects introduce additional deleterious mid‐gap states, which may explain observed silicon photoluminescence quenching. These results show that band‐edge states can mediate the triplet exciton transfer process, potentially through a sequential charge transfer mechanism.

Layered Metal–Organic Chalcogenides: 2D Optoelectronics in 3D Self-Assembled Semiconductors

2026-03-08T03:40:10Z

Layered Metal–Organic Chalcogenides: 2D Optoelectronics in 3D Self-Assembled Semiconductors Paritmongkol, Watcharaphol; Feng, Zhifu; Refaely-Abramson, Sivan; Tisdale, William A; Kastl, Christoph; Maserati, Lorenzo Molecular self-assembly offers an effective and scalable way to design nanostructured materials with tunable optoelectronic properties. In the past 30 years, organic chemistry has delivered a plethora of metal-organic structures based on the combination of organic groups, chalcogens, and a broad range of metals. Among these, several layered metal-organic chalcogenides (MOCs)─including "mithrene" (AgSePh)─recently emerged as interesting platforms to host 2D physics embedded in 3D crystals. Their combination of broad tunability, easy processability, and promising optoelectronic performance is driving a renewed interest in the more general material group of "low-dimensional" hybrids. In addition, the covalent MOC lattice provides higher stability compared with polar materials in operating devices. Here, we provide a perspective on the rise of 2D MOCs in terms of their synthesis approaches, 2D quantum confined exciton physics, and potential future applications in UV and X-ray photodetection, chemical sensors, and electrocatalysis.

Exciton fission enhanced silicon solar cell

2026-03-08T03:39:57Z

Exciton fission enhanced silicon solar cell Nagaya, Narumi; Lee, Kangmin; Perkinson, Collin F; Li, Aaron; Lee, Youri; Zhong, Xinjue; Lee, Sujin; Weisburn, Leah P; Wang, Janet Z; Baikie, Tomi K; Bawendi, Moungi G; Van Voorhis, Troy; Tisdale, William A; Kahn, Antoine; Seo, Kwanyong; Baldo, Marc A While silicon solar cells dominate global photovoltaic energy production, their continued improvement is hindered by the single-junction limit. One potential solution is to use molecular singlet exciton fission to generate two electrons from each absorbed high-energy photon. We demonstrate that the long-standing challenge of coupling molecular excited states to silicon solar cells can be overcome using sequential charge transfer. Combining zinc phthalocyanine, aluminum oxide, and a shallow junction crystalline silicon microwire solar cell, the peak charge generation efficiency per photon absorbed in tetracene is (138% ± 6%), comfortably surpassing the quantum efficiency limit for conventional silicon solar cells and establishing a new, scalable approach to low-cost, high-efficiency photovoltaics.

1D Silver Organochalcogenide Semiconductors: Color Tunable Luminescence, Polarized Emission, and Long-Range Exciton Diffusion

2026-03-08T03:40:03Z

1D Silver Organochalcogenide Semiconductors: Color Tunable Luminescence, Polarized Emission, and Long-Range Exciton Diffusion Sakurada, Tomoaki; Pathoor, Nithin; Matsumoto, Takuma; Khamlue, Rattapon; Chatsiri, Petcharaphorn; Valenta, Jan; Kawamoto, Tadashi; Omagari, Shun; Tisdale, William A; Paritmongkol, Watcharaphol; Cho, Yeongsu; Vacha, Martin Metal organochalcogenides (MOCs) represent a promising class of organic-inorganic hybrid semiconductors with unique light-matter interactions. Their hybrid nature enables extensive structural and optoelectronic tunability via ligand engineering. In this study, we systematically modulated the electronic properties of ligands using Cl and Me functional groups, achieving precise control over the optoelectronic properties of Ag-based MOCs. Structural analysis revealed that these MOCs adopt a one-dimensional (1D) chain structure with organic ligands surrounding a Ag-chalcogen core. Density functional theory (DFT) calculations demonstrated that MOCs exhibit characteristics of 1D semiconductors with strongly dispersive conduction and valence bands aligned along the crystal rod directions. Experimentally, the MOCs displayed bright luminescence, with peaks centered between 560 and 690 nm. The substitution of Cl with Me groups in the benzene ligands induced a red shift in both absorption and photoluminescence, corroborated by experimental and theoretical analyses. Further optical measurements indicated that the emission from the MOCs is strongly polarized along the chain directions. Notably, Se-based MOCs exhibited enhanced exciton diffusivity along the chain axis with a diffusion length of 130 nm, which is among the highest reported for covalent systems. The observed trend in carrier diffusivity among individual compounds is attributed to differences in the effective masses of the carriers, as determined by DFT calculations. Our findings offer valuable insights into the systematic structural and property tuning of hybrid semiconductors and highlight the unique characteristics of the 1D MOC family.

Excitonic Anisotropy in Single‐Crystalline 2D Silver Phenylchalcogenides

2026-03-08T03:40:01Z

Excitonic Anisotropy in Single‐Crystalline 2D Silver Phenylchalcogenides Lee, Woo Seok; Cho, Yeongsu; Posmyk, Katarzyna; Peksa, Paulina; Dyksik, Mateusz; Samulewicz, Nicholas; Plochocka, Paulina; Baranowski, Michał; Kulik, Heather J; Tisdale, William A 2D materials exhibiting in‐plane anisotropy enable new applications in directional energy transport and polarized optical response. Silver phenylchalcogenides (AgEPh) – including mithrene (AgSePh), tethrene (AgTePh), and thiorene (AgSPh) – represent an exciting new addition to this family, with optical response spanning the visible to near‐UV. Here, excitonic anisotropy is predicted and characterized in this family of materials using a combination of ab initio theory and optical micro‐spectroscopy of single‐crystalline flakes. Using density functional theory and GW with the Bethe–Salpeter equation calculations, it is revealed that all AgEPh compounds exhibit anisotropic electronic band structure and host multiple delocalized excitons with in‐plane anisotropy. Room‐temperature polarization‐resolved optical micro‐spectroscopy shows that orthogonally polarized excitons with similar energy lead to nearly isotropic absorption in AgSPh, whereas energy separation between excitonic resonances in AgSePh and AgTePh leads to strong absorption and emission anisotropy. Cryogenic reflectance micro‐spectroscopy further reveals exciton fine structure in AgSePh, reconciling the discrepancies between room‐temperature experiments and theoretical predictions. Finally, it is demonstrated that the optical response of thicker AgEPh crystals is influenced by photonic effects arising from finite crystal size. Overall, this work advances the understanding of the relationship between anisotropic structure, composition, and excitonic properties in AgEPh, providing a foundation for technological integration.

Revolutionize cold chain: an AI/ML driven approach to overcome capacity shortages

2026-03-08T03:40:12Z

Revolutionize cold chain: an AI/ML driven approach to overcome capacity shortages Jackson, Ilya; Namdar, Jafar; Saénz, Maria Jesús; Elmquist III, Richard Augustus; Dávila Novoa, Luis Rodrigo This research investigates how Artificial Intelligence (AI) and Machine Learning (ML) forecasting methodologies can be leveraged for cold chain capacity planning, specifically utilising Prophet and Seasonal Autoregressive Integrated Moving Average parametrised through grid search. In collaboration with Americold, the world's second-largest refrigerated logistic service provider, the study explores the challenges and opportunities in applying AI/ML techniques to complex operations covering 385 customers and a capacity of 73,296 pallet positions. We train and test several AI/ML and traditional statistical models using extensive data for every customer over 3.5 years. Based on the results, MAPE of 5.28% was achieved on the whole site level, and SARIMA outperformed ML models in most cases. Next, we show that developing and applying a Customer Segmentation Matrix has enabled more accurate forecasting and planning across various customer segments, addressing the issue of forecasting inaccuracies. This approach effectively improves forecasting inaccuracies, underscoring the significance of tailoring AI/ML models for demand forecasting within the cold-chain industry. Ultimately, this research presents an AI-driven approach that transcends mere forecasting, offering a practical pathway to manage capacity in light of the constraints.

Beyond binary group categorization: towards a dynamic view of human groups

2026-03-08T03:39:57Z

Beyond binary group categorization: towards a dynamic view of human groups Kish Bar-On, Kati Society is a composite of interacting people and groups. These groups play a significant role in maintaining social status, establishing group identity and social identity, and enforcing norms. As such, groups are essential for understanding human behavior. Nevertheless, the study of groups in everyday group life yields many diverse and sometimes contradicting theories of group behavior, and researchers tend to agree that we have yet to understand the emergence of groups out of aggregates of individuals. The current paper aims to shed new light on the convoluted interrelation between groups and individuals by focusing on individuals’ social identities and group categorization. It does so by exploring the dynamic nature of the self and its implications on identity and group membership, and introducing a framework recognizing the fluidity of groups and group categorization. Incorporating historical insights with contemporary theories, this paper argues for a flexible understanding of group dynamics that surpasses rigid in-group and out-group classifications, proposing instead that group affiliations exist along a continuum that reflects the ever-changing social landscape.

Reparative Urban Science: Challenging the Myth of Neutrality and Crafting Data-Driven Narratives

2026-03-08T03:40:00Z

Reparative Urban Science: Challenging the Myth of Neutrality and Crafting Data-Driven Narratives So, Wonyoung I offer how urban planning should approach technology within the context of systemic racism, advocating for a reparative approach to address the issues of urban technology perpetuating today’s racial inequality and hindering efforts to redress historical oppression. I identify three mechanisms – formalization, context removal and legitimization, and penalization and extraction – that illustrate how urban technology perpetuates historical inequalities, often penalizing marginalized groups under the pretext of neutrality and fairness. Then, I discuss methodologies of reparative urban science, aiming to use urban technology to challenge race-neutral ideologies and create data-driven narratives for reparations.

What determines EV architecture? An analysis of the most influential battery electric vehicle design decisions from market data

2026-03-08T03:39:58Z

What determines EV architecture? An analysis of the most influential battery electric vehicle design decisions from market data Khan, Mumin; Cameron, Bruce The penetration and variety of Battery Electric Vehicles (BEVs) in the automotive sector have been growing rapidly. While there is substantial research on hybrid ICE-battery vs. battery-only choices, little work has examined whether a dominant design for BEVs is emerging, as predicted by the innovation literature. This study provides a comprehensive exploration of BEV architectures, examining the influence of individual architectural decisions on vehicle performance and market prevalence. This study utilizes multivariate linear regression to analyze a curated dataset of global BEV models from 2022 and 2023, focusing on candidate architectural decisions such as battery cathode composition, battery voltage choice, number of motors, and drive layout. Our research aims to identify potential dominant designs by assessing their impact on performance metrics. The analysis then leverages statistical tools to evaluate the correlation between these architectural decisions and vehicle performance, using range as a primary indicator of consumer appeal. Findings from this research indicate significant variance in the adoption of specific BEV architectures, suggesting that the market has not yet consolidated down to a dominant design. We observe, however, that range is most strongly influenced by the architectural decisions for battery capacity, drive type, and motor type.

Chemical and Chemical-Mechanical Polishing of Surface Roughness on L-PBF/GRCop-42 Cu-Cr-Nb Additive Manufactured 10-GHz RF Structures

2026-03-08T03:39:53Z

Chemical and Chemical-Mechanical Polishing of Surface Roughness on L-PBF/GRCop-42 Cu-Cr-Nb Additive Manufactured 10-GHz RF Structures Seltzman, AH; Wukitch, SJ Laser-based powder bed fusion (L-PBF) allows additive manufacture (AM) of lower hybrid current drive (LHCD) radio-frequency (RF) launchers from Glenn Research Copper, a Cr2Nb precipitation-hardened alloy (GRCop-42) in configurations unachievable with conventional machining. Rough surfaces in AM components increase RF losses and lead to arcing in high-power vacuum RF applications. Chemical polishing, chemical-mechanical polishing, or a combination of both were utilized to planarize the internal surfaces of RF structures, resulting in surface roughness as low as Ra = 0.2 µm. Refinement in polishing techniques now enables GRCop-42 alloys (4 at. % Cr, 2 at. % Nb) to achieve similar surface roughness to GRCop-84 (8 at. % Cr, 4 at. % Nb) and equivalent cavity losses to extruded oxygen-free copper waveguides at 10 GHz.

The Name of Moses in an Egyptian Context—A Hypothetical Etymology

2026-03-08T03:40:06Z

The Name of Moses in an Egyptian Context—A Hypothetical Etymology Adair, Aaron The etymological origins of the name “Moses” have been unclear, but an Egyptian candidate is the most likely hypothesis. In this article, a new proposal is given that finds the best candidate as the Demotic word, mšꜥ, but only in the late Persian period or later would it fit the Hebrew Mōše. Evidence from Greek orthography and testimony from Manetho provide a stronger basis for this proposal over prior candidates. However, this results in a Hellenistic-era inclusion of “Moses” into the Exodus narrative.

Making Sense of Models: Connecting Science and Math Through Decoding and Modifying Computational Models

2026-03-08T03:39:44Z

Making Sense of Models: Connecting Science and Math Through Decoding and Modifying Computational Models Lee, Irene A.; Sagartz, Mary; Meyer, Patricia; Anderson, Emma The Making Sense of Models (MSM) curriculum was designed to bridge math and science learning through agent-based modeling and rich computational thinking investigations that do not require teaching computer programming in middle school classrooms. The MSM curriculum supports students in the NGSS skill of reasoning about how and why a phenomenon happens. After developing decoding skills, students are able to assess the validity of a model based on comparing mechanisms in the model to what they learned about the phenomenon being modeled. In this article, the authors describe the decoding approach and how the MSM curriculum supports students’ ability to reason about scientific models and the real world.

Cycle 7 VLBI Acceptance Report

2026-03-05T18:39:37Z

Cycle 7 VLBI Acceptance Report Crew, Geoff This report summarizes the acceptance process for VLBI which was carried out in early 2020 in preparation for the 2020 VLBI Campaigns. Even though the ALMA operations are suspended, the EHTC campaign has been cancelled, and the GMVA is going forwards without ALMA, it is still a useful exercise to report on the Acceptance testing that was done. This is especially true since (for a variety of reasons) the testing was more extensive this past January. It has also been several years since the initial Acceptance of VLBI for Cycle 4, and new features are finally to become available in Cycle 8, so it is reasonable to capture the state of things at this time. Going forward, it has been suggested as desirable for the Acceptance be added to the normal ALMA Acceptance process. This report thus serves to detail the sort of checks that can and should be made in the future. Some Action items are also noted for the near term. On the bright side, the system was totally ready. This report reviews the setup and on-site checks that can be made in a stand-alone ( no co-observing peers) mode. Then we present results from four VLBI sessions. For each, a CASA reduction is compared with a reduction of the VLBI data which (in three of the four cases) are correlated with participating EHTC sites. This report was prepared for the formal acceptance of the software required for ALMA Observing Cycle 7. Notionally it is ALMA Technical Note #20, but not published (yet) as such.

Thermal analog computing: Application to matrix-vector multiplication with inverse-designed metastructures

2026-02-04T03:08:34Z

Thermal analog computing: Application to matrix-vector multiplication with inverse-designed metastructures Silva, Caio; Romano, Giuseppe The rising computational demand of modern workloads has renewed interest in energy-efficient paradigms, such as neuromorphic and analog computing. A fundamental operation in these systems is matrix-vector multiplication (MVM), ubiquitous in signal processing and machine learning. Here, we demonstrate MVM using inverse-designed metastructures that exploit heat conduction as the signal carrier. The proposed approach is based on a generalization of effective thermal conductivity to systems with multiple input and output ports: The input signal is encoded as a set of applied temperatures, while the output is represented by the power collected at designated terminals. The metastructures are obtained via density-based topology optimization, enabled by a differentiable thermal transport solver and automatic differentiation, achieving an accuracy greater than 99% in most cases across a pool of matrices with dimensions 2 ×2 and 3 ×3. We apply this methodology—termed thermal analog computing—to realize matrices relevant to practical tasks, including the discrete Fourier transform and convolutional filters. These findings open avenues for analog information processing in thermally active environments, including temperature-gradient sensing in microelectronics and thermal control systems.

When competition becomes contagious: Strategic arms racing spillovers, alliance politics, and the Sino-American nuclear competition

2026-02-04T03:08:48Z

When competition becomes contagious: Strategic arms racing spillovers, alliance politics, and the Sino-American nuclear competition Seitz, Samuel M.; Ji, Elliot S. The development of new conventional counterforce systems and improved missile defence systems enables non-nuclear states to directly influence the strategic nuclear balance. These dynamics increase the possibility of strategic arms racing spillovers, where arms racing in one dyad yields capabilities that threaten third parties’ arsenals and thus creates a type of security dilemma. It also increases the risk of non-nuclear allies entrapping their nuclear patrons in strategic arms racing. We illustrate this argument via the case of North and South Korea’s arms racing.

Building communities of critical inquiry in the language classroom

2026-02-04T03:08:44Z

Building communities of critical inquiry in the language classroom Dessein, Eva; Ledford, Julian A Addressing issues of power, difference, and social stratification is essential in language education, where systemic inequities shape classroom experiences. This study examines the design and impact of three targeted modules implemented in beginner and intermediate French courses at two U.S. institutions. Grounded in critical pedagogical principles, the modules focused on language and power, inclusive language practices, and cultural and intercultural awareness. They aimed to foster critical inquiry through individual reflection and engagement with socially relevant topics. Analysis of student reflections and survey responses indicates that the modules supported learners in critically examining how language reinforces or challenges inequities, particularly in relation to gender biases and colonial legacies. Students reported increased awareness of linguistic hierarchies, a stronger sense of agency, and deeper reflection on language’s sociopolitical dimensions. The modules also encouraged engagement with inclusive language and cultural diversity. While the interventions promoted critical awareness and personal growth, findings point to limited peer interaction and community-building. This suggests a need for more structured opportunities for dialogic learning. Overall, the study highlights the transformative potential of critical pedagogy in language education and the importance of designing inclusive curricula that prepare students to reflect on and challenge systemic inequities.

Managing technology-related disruptions and vulnerabilities in highly automated warehouse systems: an integrative review and research agenda

2026-02-04T03:08:41Z

Managing technology-related disruptions and vulnerabilities in highly automated warehouse systems: an integrative review and research agenda Rodríguez-García, Miguel; Kembro, Joakim Hans; Betts, Kellen; Ponce-Cueto, Eva Recent technological developments in warehousing have introduced new risks. This paper presents an integrative review that combines insights from highly automated warehouse systems (HAWS) and risk management, providing a comprehensive understanding of technology-related warehouse disruptions and vulnerabilities. We identify five major disruptions that can affect HAWS: cyberattacks, technology sabotage, technology failures, power and network outages, and human-machine interaction issues. Moreover, we identify 48 technology-related vulnerabilities across all disruptions. In particular, HAWS have become vulnerable to cyberattacks due to the increasing number of warehouse technology suppliers, greater complexity of multi-robot networks such as AMRs, reliance on cloud-based systems, and cascading effect of cyberattacks due to higher levels of interconnectivity in HAWS networks. Our review also shows that risk management strategies in HAWS are unevenly covered in the literature. In response, we propose a research agenda with 17 pathways aimed at enhancing prevention, detection, mitigation, and recovery strategies for HAWS. Managers also benefit from the identified disruptions and vulnerabilities, as they serve as a reference point for understanding their specific technology-related risks in HAWS. In addition, managers can use our review of current risk management practices as a benchmark and our research agenda to think about areas that they could develop further.

Preliminary Investigation of Gamma Radiation on the Chemical and Physical Characteristics of an Organic Coolant

2026-02-04T03:08:46Z

Preliminary Investigation of Gamma Radiation on the Chemical and Physical Characteristics of an Organic Coolant Vasquez, Angel; Seshadri, Arunkumar; Shirvan, Koroush; Buongiorno, Jacopo Organic-cooled reactor concepts offer potential advantages over traditional light water reactors, including operation at elevated temperatures and reduced pressures. However, radiation-induced degradation of organic coolants remains a critical concern requiring thorough investigation. This study examines the effects of gamma irradiation (1-MGy dose) on Dowtherm A (27% biphenyl, 73% diphenyl ether) under varying atmospheric conditions (ambient air versus argon) and temperatures (room temperature versus 250°C). Chemical characterization using Fourier transform infrared spectroscopy, ultraviolet-visible spectroscopy (UV-Vis), and gas chromatography-mass spectrometry revealed the formation of higher molecular weight byproducts, including terphenyls and quaterphenyls, along with notable biphenyl degradation. Physical property measurements using differential scanning calorimetry, rheometry, and thermal conductivity analysis demonstrated significant changes in the thermophysical properties, including decreased heat capacity and viscosity, with increased thermal conductivity observed under argon irradiation conditions. Pronounced photodarkening occurred in all the irradiated samples, with atmospheric conditions significantly influencing degradation pathways. UV-Vis analysis indicated that oxygen presence during irradiation suppresses certain chromophoric species formation. These findings provide crucial insights into radiation-induced degradation mechanisms and their impact on coolant performance, informing future organic coolant system design and optimization strategies for advanced reactor applications.

Housing data politics in the United States: Inequitable open data, informal networks, and strategic neutrality

2026-02-04T03:08:43Z

Housing data politics in the United States: Inequitable open data, informal networks, and strategic neutrality Aizman, Asya; So, Wonyoung; Navalkha, Chenab; D’Ignazio, Catherine Open housing data—property transactions, eviction filings, 311 complaints, and rental registries—have been a crucial resource for policymaking and real estate professionals. Meanwhile, housing data actors increasingly collect, analyze, and use data to address housing inequality, including efforts related to eviction prevention and land use reform, among others. This paper examines the motivations and practices of grassroots and institutional housing data actors. From a field scan of 67 entities engaged in housing data work across 12 U.S. states and 18 municipalities, we conducted 18 in-depth interviews to explore how housing data actors operate, their political goals, and data processes. We put forward a two‑axis framework that positions housing data actors according to their organizational structure (institutional/grassroots) and their stated data ideology (neutral/political). This framework contributes to understanding how different actors navigate complex issues such as embedded power dynamics and ethics in housing data. This two-axis view supplies a vocabulary for tracing how normative commitments and material constraints shape housing data pipelines and, ultimately, housing outcomes across the broader housing information ecosystem.

Unconstrained Sovereignty: Delegation of Authority and Reversibility

2026-02-04T03:08:49Z

Unconstrained Sovereignty: Delegation of Authority and Reversibility Grinberg, Mariya The concept of sovereignty shapes our understanding of the world. Yet our current understanding of sovereignty conflates delegation of authority with loss of sovereignty. Delegation is relatively cheap, quick, and leads to an assured outcome; it’s an affirmation of sovereignty. Use of force, however, is required to regain lost sovereignty. I propose a definition of sovereignty that draws a clear distinction between sovereignty and delegated authority. Adopting this definition shows that sovereignty applies across time and space, it is indivisible, institutions do not place permanent constraints on supreme authority, and popular sovereignty is not a well-grounded concept.

Concurrent Balanced Augmented Trees

2026-02-03T05:03:30Z

Concurrent Balanced Augmented Trees Wrench, Evan; Singh, Ajay; Roh, Younghun; Fatourou, Panagiota; Jayanti, Siddhartha; Ruppert, Eric; Wei, Yuanhao Augmentation makes search trees tremendously more versatile, allowing them to support efficient aggregation queries, order-statistic queries, and range queries in addition to insertion, deletion, and lookup. In this paper, we present the first lock-free augmented balanced search tree supporting generic augmentation functions. Our algorithmic ideas build upon a recent augmented unbalanced search tree presented by Fatourou and Ruppert [DISC, 2024]. We implement both data structures, solving some memory reclamation challenges in the process, and provide an experimental performance analysis of them. We also present optimized versions of our balanced tree that use delegation to achieve better scalability and performance (by more than 2x in most workloads). Our experiments show that our augmented balanced tree completes updates 2.2 to 30 times faster than the unbalanced augmented tree, and outperforms unaugmented trees by up to several orders of magnitude on 120 threads. PPoPP ’26, Sydney, NSW, Australia

Building Intelligent Agents with Neuro-Symbolic Concepts

2026-02-03T05:03:25Z

Building Intelligent Agents with Neuro-Symbolic Concepts Mao, Jiayuan; Tenenbaum, Joshua; Wu, Jiajun This article presents a concept-centric paradigm for building agents that can learn continually and reason flexibly. The concept-centric agent utilizes a vocabulary of neuro-symbolic concepts. These concepts, such as object, relation, and action concepts, are grounded on sensory inputs and actuation outputs. They are also compositional, allowing for the creation of novel concepts through their structural combination. To facilitate learning and reasoning, the concepts are typed and represented using a combination of symbolic programs and neural network representations. Leveraging such neuro-symbolic concepts, the agent can efficiently learn and recombine them to solve various tasks across different domains, ranging from 2D images, videos, 3D scenes, and robotic manipulation tasks. This concept-centric framework offers several advantages, including data efficiency, compositional generalization, continual learning, and zero-shot transfer.

Foundational Verification of Running-Time Bounds for Interactive Programs

2026-02-03T05:03:33Z

Foundational Verification of Running-Time Bounds for Interactive Programs Tockman, Andy; Singh, Pratap; Erbsen, Andres; Gruetter, Samuel; Chlipala, Adam Some important domains of software demand concrete bounds on how long functions may run, for instance for real-time cyberphysical systems where missed deadlines may damage industrial machinery. Such programs may interact with external devices throughout execution, where time deadlines ought to depend on, for instance, sensor readings (e.g. we only scramble to close a valve immediately when a sensor reports that a tank is about to overflow). We present the first software-development toolchain that delivers first-principles proofs of meaningful time bounds for interactive machine code, while allowing all per-application programming and verification to happen at the source-code level. We allow C-like programs to be proved against separation-logic specifications that also constrain their running time, and such proofs are composed with verification of a compiler to RISC-V machine code. All components are implemented and proved inside the Rocq proof assistant, producing final theorems whose statements depend only on machine-language formal semantics and some elementary specification constructions for describing running time. As a capstone case study, we extended a past verification (of a real microcontroller-based cyberphysical system) to bound time between arrival of network packets and actuation of an attached device. CPP ’26, Rennes, France

Network-RBV for Critical Minerals: How Standards, Permits, and Licensing Shape Midstream Bottlenecks

2026-02-03T05:04:11Z

Network-RBV for Critical Minerals: How Standards, Permits, and Licensing Shape Midstream Bottlenecks Kegenbekov, Zhandos; Alipova, Alima; Jackson, Ilya Critical mineral supply chains underpin electric mobility, power electronics, clean hydrogen, and advanced manufacturing. Drawing on the resource-based view (RBV), the relational view, and dynamic capabilities, we conceptualize advantage not as ownership of ore bodies but as orchestration of multi-tier resource systems: upstream access, midstream processing know-how, standards and permits, and durable inter-organizational ties. In a world of high concentration at key stages (refining, separation, engineered materials), full “decoupling” is economically costly and technologically constraining. We argue for structured cooperation among the United States, European Union, China, and other producers and consumers, combined with selective domestic capability building for bona fide security needs. Methodologically, we conduct a structured conceptual synthesis integrating RBV, relational view, dynamic capabilities, and network-of-network research, combined with a structured comparative policy analysis of U.S./EU/Chinese instruments anchored in official documents. We operationalize the argument via technology–material dependency maps that identify midstream bottlenecks and the policy/standard levers most likely to expand qualified, compliant capacity.

REEV SENSE IMUs for Spatiotemporal Gait Analysis in Post-Stroke Patients: Validation Against Optical Motion Capture

2026-02-03T05:04:02Z

REEV SENSE IMUs for Spatiotemporal Gait Analysis in Post-Stroke Patients: Validation Against Optical Motion Capture Marsan, Thibault; Clauzade, Sacha; Zhang, Xiang; Grandin, Nicolas; Urman, Tatiana; Linton, Evan; Sibachir, Samy; Ricciardi, Catherine E.; Temporelli, Robin Objective gait assessment is essential for post-stroke rehabilitation monitoring, yet optical motion capture systems remain inaccessible to most clinical settings due to cost and infrastructure constraints. This study assessed the validity of the REEV SENSE IMU for measuring spatiotemporal gait parameters in post-stroke individuals and evaluated assistive device effects on measurement accuracy. Twenty chronic post-stroke participants were enrolled, and fourteen completed the study (ten without an assistive device, four using a cane) after applying pre-defined exclusion criteria (walking speed <0.28 m/s, n = 6). Participants walked at self-selected speed while simultaneously being recorded by REEV SENSE IMUs and optical motion capture. Spatiotemporal parameters from matched heel strikes were compared using intraclass correlation coefficients (ICC), mean relative error (MRE), and Bland–Altman analysis. Temporal parameters demonstrated excellent reliability: contact time (ICC 0.96–0.99, MRE 2.77–5.45%), stride duration (ICC 0.95–0.99, MRE 2.57–2.62%), and cadence (ICC 0.98–0.99, MRE 1.80–1.93%). Spatial parameters showed greater variability, with stride length degrading substantially in slow-walking conditions (Cane group: ICC 0.76, MRE 8.60%). REEV SENSE provides reliable temporal parameter measurement comparable to commercial systems, positioning it as a practical tool for clinical gait monitoring in post-stroke rehabilitation. However, spatial parameter accuracy requires cautious interpretation in slow-walking regimes, necessitating independent validation when clinical decisions depend on precise stride length estimates.

Who Am I? Eyebrow Follicles Minimize Donor-Derived DNA for Germline Testing After Hematopoietic Stem Cell Transplantation

2026-02-03T05:03:59Z

Who Am I? Eyebrow Follicles Minimize Donor-Derived DNA for Germline Testing After Hematopoietic Stem Cell Transplantation Mertens, Matthias; Sadlo, Mona; Kühl, Jörn-Sven; Metzeler, Klaus; Zschenderlein, Louisa; Edelmann, Jeanett; Lehmann, Claudia; Thull, Sarah; Karakaya, Mert; Velmans, Clara; Tumewu, Theresa; Böhme, Matthias; Klötzer, Christina; Weigert, Anne; Vucinic, Vladan; Hentschel, Julia; Mertens, Mareike Germline genetic testing plays a critical role in diagnosing inherited predispositions and increasingly guides therapeutic and surveillance choices—but becomes technically challenging after allogeneic hematopoietic stem cell transplantation (HSCT), when donor-derived DNA contaminates host tissues. To address this, we compared donor-derived DNA across three accessible tissues—buccal swab, nail, and eyebrow follicles—in recipients after hematopoietic stem cell transplantation using two orthogonal assays (34-SNP next-generation sequencing and a 27-marker short tandem repeat panel) and modeled clinical covariates that influence chimerism. Eyebrow follicles showed consistently low donor DNA (median 1% by NGS; 3% by STR) whereas buccal swabs and nails carried substantially higher donor fractions (+25 and +22 percentage points versus eyebrow, respectively; both p < 0.01). Across methods, STR yielded on average ≈6 percentage points higher donor fractions than NGS at low-level chimerism. Several transplant covariates correlated with chimerism: matched-related donors and a perfect HLA match (10/10) were each associated with lower donor DNA (≈12–14 and 15–20 percentage points, respectively); longer times since hematopoietic stem cell transplantation correlated with lower levels for nail samples, and donor–recipient sex match correlated with higher donor DNA (~7–8 percentage points). Even low-level chimerism can distort germline variant interpretation. We propose a pragmatic protocol for post-hematopoietic stem cell transplantation germline testing that prioritizes eyebrow follicles as the default tissue. An SNP-based quality control assay is used to flag unsafe donor fractions (≥ 5–10%) before comprehensive germline analysis, reducing the risk that chimeric donor DNA distorts germline variant interpretation.

The effect of settlements on the stresses in building frames

2026-02-03T04:59:35Z

The effect of settlements on the stresses in building frames Granberg, Robert J. Thesis: B.S., Massachusetts Institute of Technology, Department of Building Engineering and Construction, 1935; Includes bibliographical references.

Irradiation grafting of styrene onto dacron fibers and films

2026-02-03T04:59:32Z

Irradiation grafting of styrene onto dacron fibers and films Schnetzer, L. J.; Hendren, J. W. Thesis: B.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1959; Includes bibliographical references (leaves 43-44).

An investigation of sound transmission irregularity in a one dimensional enclosure

2026-02-03T04:59:29Z

An investigation of sound transmission irregularity in a one dimensional enclosure Foster, Isaac C. Thesis: B.S., Massachusetts Institute of Technology, Department of Physics, 1949

Deposition and characterization of very low pressure CVD silicon/silicon-germanium heteroepitaxial structures

2026-02-03T03:48:14Z

Deposition and characterization of very low pressure CVD silicon/silicon-germanium heteroepitaxial structures Tsai, Curtis. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 1992; Includes bibliographical references (leaves 135-146).

An experimental study of the law of parity conservation in electromagnetic interactions.

2026-02-03T03:48:30Z

An experimental study of the law of parity conservation in electromagnetic interactions. Hegblom, Edwin Richard. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Physics, 1965

Analysis of a dynamic sales call policy model.

2026-02-03T04:59:26Z

Analysis of a dynamic sales call policy model. Karash, Richard Ivan. Thesis: B.S., Massachusetts Institute of Technology, Department of Physics, 1968; Bibliography: leaf 97.

MEKIN numerical modeling and simulation of the SPERT-III E-core transient tests

2026-02-03T04:58:28Z

MEKIN numerical modeling and simulation of the SPERT-III E-core transient tests Tan, Lip-Bu. Thesis: M.S., Massachusetts Institute of Technology, Department of Nuclear Engineering, 1980; Includes bibliographical references.

Comparative tests of the Boston Elevated Co's surface cars

2026-02-03T04:59:24Z

Comparative tests of the Boston Elevated Co's surface cars Jones, Philip C.; Katsainos, Nicholas M. Thesis: B.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1912

Rules for ring closure and aspects of organolithium chemistry

2026-02-03T03:48:24Z

Rules for ring closure and aspects of organolithium chemistry Dupont, William Alan. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Chemistry, 1980; Vita.; Includes bibliographical references.

Dracut nickel ore ; Geology and concentration, ore no. 2592

2026-02-03T04:58:59Z

Dracut nickel ore ; Geology and concentration, ore no. 2592 Burton, Eugene.; Spalding, William Livingston. Thesis: B.S., Massachusetts Institute of Technology, Department of Mining Engineering and Metallurgy, 1905

Evaluation of recirculating well technology with a cost comparison to pump and treat technology for containment of the CS-10 contaminant plume at the Massachusetts Military Reservation

2026-02-03T04:58:24Z

Evaluation of recirculating well technology with a cost comparison to pump and treat technology for containment of the CS-10 contaminant plume at the Massachusetts Military Reservation Smith, Mathew D. (Mathew Darin) Thesis: M. Eng., Massachusetts Institute of Technology, Department of Civil and Environmental Engineering, 1997; Includes bibliographical references (leaves 43-45).

The crystallization of sucrose

2026-02-03T04:58:17Z

The crystallization of sucrose Brown, Ernest K. Thesis: M.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1929; Includes bibliographical references (leaf 81).

Quantum Circuits Are Just a Phase

2026-02-03T05:03:32Z

Quantum Circuits Are Just a Phase Heunen, Chris; Lemonnier, Louis; McNally, Christopher; Rice, Alex Quantum programs today are written at a low level of abstraction---quantum circuits akin to assembly languages - and the unitary parts of even advanced quantum programming languages essentially function as circuit description languages. This state of affairs impedes scalability, clarity, and support for higher-level reasoning. More abstract and expressive quantum programming constructs are needed. To this end, we introduce a simple syntax for generating unitaries from "just a phase"; we combine a (global) phase operation that captures phase shifts with a quantum analogue of the "if let" construct that captures subspace selection via pattern matching. This minimal language lifts the focus from gates to eigendecomposition, conjugation, and controlled unitaries; common building blocks in quantum algorithm design. We demonstrate several aspects of the expressive power of our language in several ways. Firstly, we establish that our representation is universal by deriving a universal quantum gate set. Secondly, we show that important quantum algorithms can be expressed naturally and concisely, including Grover's search algorithm, Hamiltonian simulation, Quantum Fourier Transform, Quantum Signal Processing, and the Quantum Eigenvalue Transformation. Furthermore, we give clean denotational semantics grounded in categorical quantum mechanics. Finally, we implement a prototype compiler that efficiently translates terms of our language to quantum circuits, and prove that it is sound with respect to these semantics. Collectively, these contributions show that this construct offers a principled and practical step toward more abstract and structured quantum programming.

A Digital Engineering Framework for Piston Pin Bearings via Multi-Physics Thermo-Elasto-Hydrodynamic Modeling

2026-02-03T05:04:01Z

A Digital Engineering Framework for Piston Pin Bearings via Multi-Physics Thermo-Elasto-Hydrodynamic Modeling Shu, Zhiyuan; Tian, Tian The piston pin operates under severe mechanical and thermal conditions, making accurate lubrication prediction essential for engine durability. This study presents a comprehensive digital engineering framework for piston pin bearings, built upon a fully coupled thermo-elasto-hydrodynamic (TEHD) formulation. The framework integrates: (1) a Reynolds-equation hydrodynamic solver with temperature-/pressure-dependent viscosity and cavitation; (2) elastic deformation obtained from FEA (finite element analysis)-based compliance matrices; (3) a break-in module that iteratively adjusts surface profiles before steady-state simulation; (4) a three-body heat transfer model resolving heat conduction, convection, and solid–liquid interfacial heat exchange. Applied to a heavy-duty diesel engine, the framework reproduces experimentally observed behaviors, including bottom-edge rounding at the small end and the slow unidirectional drift of the floating pin. By integrating multi-physics modeling with design-level flexibility, this work aims to provide a robust digital twin for the piston-pin system, enabling virtual diagnostics, early-stage failure prediction, and data-driven design optimization for engine development.

The Wafold: Curvature-Driven Termination and Dimensional Compression in Black Holes

2026-02-03T05:04:13Z

The Wafold: Curvature-Driven Termination and Dimensional Compression in Black Holes Viaña, Javier This work explores a geometric description of black holes in which spacetime terminates on a curvature-triggered hypersurface rather than extending to an interior singularity. We study the implications of a scenario in which, upon reaching a critical curvature threshold, the three-dimensional spatial geometry compresses into a thin, closed boundary identified here as the wafold. Beyond this, the manifold would no longer continue, and all mass–energy and information would be confined to the hypersurface itself. This framework combines two well-explored paths: (1) curvature-driven geometric compression, in which extreme curvature forces the bulk degrees of freedom to become supported on a thin hypersurface (without altering the underlying dimensionality of spacetime), and (2) the motivation underlying the holographic principle, namely that black-hole entropy scales with surface area rather than volume, suggesting that information is governed by a boundary geometry rather than a bulk volume. We elaborate a dimensional conversion law that would be required to describe the collapse of spatial volume into surface area as a conserved flux of geometric capacity across the wafold, and we analyze the resulting consequences of treating this hypersurface as the terminal boundary of the manifold.

Design and Implementation of an Automated Thermal Imaging Device for Lower Limb Prosthetic Applications

2026-02-03T05:04:10Z

Design and Implementation of an Automated Thermal Imaging Device for Lower Limb Prosthetic Applications Pizarro, Daniel; Huegel, Joel C.; Diaz, Elias; Alemon, Beatriz; Herr, Hugh; Felix-Herran, Luis C. Since elevated temperature and humidity may occur at the prosthetic socket–skin interface, it is essential to collect thermal data from the residual limb, as this information serves as an indicator of adverse effects such as irritation, postural problems, and significant damage to health. These data are obtained non-invasively through the execution of a thermal imaging (TI) procedure. However, the precision and repeatability of a TI procedure rely significantly on its execution technique. This work presents the design and implementation of a mechatronic device that automates a thermal imaging technique. The application of the device is in lower-limb prosthetics evaluation. The proposed system improves data acquisition consistency by reducing execution time and minimizing human error, thereby enhancing the reproducibility and reliability of thermal measurements. The introduced device, Thermal Imaging Booth, proposes an automated solution for TI standardization in clinical and research settings. By minimizing inconsistencies, this system improves the diagnostic potential of thermography, facilitating its adoption in biomedical applications.

Peepco: Batch-Based Consistency Optimization

2026-02-01T07:32:05Z

Peepco: Batch-Based Consistency Optimization Kuraj, Ivan; Feser, John; Polikarpova, Nadia; Solar-Lezama, Armando We present batch-based consistency, a new approach for consistency optimization that allows programmers to specialize consistency with application-level integrity properties. We implement the approach with a two-step process: we statically infer optimal consistency requirements for executions of bounded sets of operations, and then, use the inferred requirements to parameterize a new distributed protocol to relax operation reordering at run time when it is safe to do so. Our approach supports standard notions of consistency. We implement batch-based consistency in Peepco, demonstrate its expressiveness for partial data replication, and examine Peepco’s run-time performance impact in different settings.

Finch: Sparse and Structured Tensor Programming with Control Flow

2026-02-01T07:31:58Z

Finch: Sparse and Structured Tensor Programming with Control Flow Ahrens, Willow; Collin, Teodoro; Patel, Radha; Deeds, Kyle; Hong, Changwan; Amarasinghe, Saman From FORTRAN to NumPy, tensors have revolutionized how we express computation. However, tensors in these, and almost all prominent systems, can only handle dense rectilinear integer grids. Real world tensors often contain underlying structure, such as sparsity, runs of repeated values, or symmetry. Support for structured data is fragmented and incomplete. Existing frameworks limit the tensor structures and program control flow they support to better simplify the problem. In this work, we propose a new programming language, Finch, which supports both flexible control flow and diverse data structures. Finch facilitates a programming model which resolves the challenges of computing over structured tensors by combining control flow and data structures into a common representation where they can be co-optimized. Finch automatically specializes control flow to data so that performance engineers can focus on experimenting with many algorithms. Finch supports a familiar programming language of loops, statements, ifs, breaks, etc., over a wide variety of tensor structures, such as sparsity, run-length-encoding, symmetry, triangles, padding, or blocks. Finch reliably utilizes the key properties of structure, such as structural zeros, repeated values, or clustered non-zeros. We show that this leads to dramatic speedups in operations such as SpMV and SpGEMM, image processing, and graph analytics.

Smooth, Integrated Proofs of Cryptographic Constant Time for Nondeterministic Programs and Compilers

2026-02-01T07:32:16Z

Smooth, Integrated Proofs of Cryptographic Constant Time for Nondeterministic Programs and Compilers Conoly, Owen; Erbsen, Andres; Chlipala, Adam Formal verification of software and compilers has been used to rule out large classes of security-critical issues, but risk of unintentional information leakage has received much less consideration. It is a key requirement for formal specifications to leave some details of a system's behavior unspecified so that future implementation changes can be accommodated, and yet it is nonetheless expected that these choices would not be made based on confidential information the system handles. This paper formalizes that notion using omnisemantics and plain single-copy assertions, giving for the first time a specification of what it means for a nondeterministic program to be constant-time or more generally to avoid leaking (a part of) its inputs. We use this theory to prove data-leak-free execution of core cryptographic routines compiled from Bedrock2 C to RISC-V machine code, showing that the smooth specification and proof experience omnisemantics provides for nondeterminism extends to constant-time properties in the same setting. We also study variants of the key program-compiler contract, highlighting pitfalls of tempting simplifications and subtle consequences of how inputs to nondeterministic choices are constrained. Our results are backed by modular program-logic and compiler-correctness theorems, and they integrate into a neat end-to-end theorem in the Coq proof assistant.

NeuroChat: A Neuroadaptive AI Chatbot for Customizing Learning Experiences

2026-02-01T07:32:11Z

NeuroChat: A Neuroadaptive AI Chatbot for Customizing Learning Experiences Baradari, D?nya; Kosmyna, Nataliya; Petrov, Oscar; Kaplun, Rebecah; Maes, Pattie Generative AI is reshaping education by enabling personalized, on-demand learning experiences. However, current AI systems lack awareness of the learner’s cognitive state, limiting their adaptability. In parallel, electroencephalography (EEG)-based neuroadaptive systems have shown promise in enhancing engagement through real-time physiological feedback. This paper introduces NeuroChat, a neuroadaptive AI tutor that integrates real-time EEG-based engagement tracking with a large language model to adapt its conversational responses. By continuously monitoring learners’ cognitive engagement, NeuroChat dynamically adjusts content complexity, tone, and response style in a closed-loop interaction. In a within-subjects study (n = 24), NeuroChat significantly increased both EEG-measured and self-reported engagement compared to a non-adaptive chatbot. However, no significant differences in short-term learning outcomes were observed. These findings demonstrate the feasibility of real-time brain–AI interaction for education and highlight opportunities for deeper personalization, longer-term adaptation, and richer learning assessment in future neuroadaptive systems. CUI ’25, Waterloo, ON, Canada

From Synthetic to Human: The Gap Between AI-Predicted and Actual Pro-Environmental Behavior Change After Chatbot Persuasion

2026-02-01T07:32:08Z

From Synthetic to Human: The Gap Between AI-Predicted and Actual Pro-Environmental Behavior Change After Chatbot Persuasion Doudkin, Alexander; Pataranutaporn, Pat; Maes, Pattie Pro-environmental behavior (PEB) is vital to combat climate change, yet turning awareness into intention and action remains elusive. We explore large language models (LLMs) as tools to promote PEB, comparing their impact across 3,600 participants: real humans (n=1,200), simulated humans based on actual participant data (n=1,200), and fully synthetic personas (n=1,200). All three participant groups faced either personalized chatbots, standard chatbots, or static statements, employing four persuasion strategies (moral foundations, future self-continuity, action orientation, or ”freestyle” chosen by the LLM). Results reveal a ”synthetic persuasion paradox”: synthetic and simulated participants significantly change their post-intervention PEB stance, while human attitudes barely shift. Simulated participants better approximate human behavior but still overestimate effects. This disconnect underscores LLM’s potential for pre-evaluating PEB interventions but warns of its limits in predicting human responses. We call for refined synthetic modeling and sustained and extended human trials to align conversational AI’s promise with tangible sustainability outcomes. CUI ’25, Waterloo, ON, Canada

Adaptive Approximation Schemes for Matching Queues

2026-02-01T07:32:06Z

Adaptive Approximation Schemes for Matching Queues AmaniHamedani, Alireza; Aouad, Ali; Saberi, Amin We study a continuous-time, infinite-horizon dynamic bipartite matching problem. Suppliers arrive according to a Poisson process; while waiting, they may abandon the queue at a uniform rate. Customers on the other hand must be matched upon arrival. The objective is to minimize the expected long-term average cost subject to a throughput constraint on the total match rate. Previous literature on dynamic matching focuses on ”static” policies, where the matching decisions do not depend explicitly on the state of the supplier queues, achieving constant-factor approximations. By contrast, we design ”adaptive” policies, which leverage queue length information, and obtain near-optimal polynomial-time algorithms for several classes of instances. First, we develop a bi-criteria fully polynomial-time approximation scheme for dynamic matching on networks with a constant number of queues—that computes a (1−є)-approximation of the optimal policy in time polynomial in both the input size and 1/є. A key new technique is a hybrid LP relaxation, which combines static and state-dependent LP approximations of the queue dynamics, after a decomposition of the network. Networks with a constant number of queues are motivated by deceased organ donation schemes, where the supply types can be divided according to blood and tissue types. The above algorithm, combined with a careful cell decomposition gives a polynomial-time approximation scheme for dynamic matching on Euclidean networks of fixed dimension. The Euclidean case is of interest in ride-hailing and spatial service platforms, where the goal is to fulfill as many trips as possible while minimizing driving distances. STOC ’25, Prague, Czechia

Output-Sensitive Approximate Counting via a Measure-Bounded Hyperedge Oracle, or: How Asymmetry Helps Estimate ��-Clique Counts Faster

2026-02-01T07:32:12Z

Output-Sensitive Approximate Counting via a Measure-Bounded Hyperedge Oracle, or: How Asymmetry Helps Estimate ��-Clique Counts Faster Censor-Hillel, Keren; Even, Tomer; Vassilevska Williams, Virginia Dell, Lapinskas and Meeks [DLM SICOMP 2022] presented a general reduction from approximate counting to decision for a class of fine-grained problems that can be viewed as hyperedge counting or detection problems in an implicit hypergraph, thus obtaining tight equivalences between approximate counting and decision for many key problems such as k-clique, k-sum and more. Their result is a reduction from approximately counting the number of hyperedges in an implicit k-partite hypergraph to a polylogarithmic number of calls to a hyperedge oracle that returns whether a given subhypergraph contains an edge. The main result of this paper is a generalization of the DLM result for output-sensitive approximate counting, where the running time of the desired counting algorithm is inversely proportional to the number of witnesses. Our theorem is a reduction from approximately counting the (unknown) number of hyperedges in an implicit k-partite hypergraph to a polylogarithmic number of calls to a hyperedge oracle called only on subhypergraphs with a small “measure”. If a subhypergraph has ui nodes in the ith node partition of the k-partite hypergraph, then its measure is ∏i ui. Using the new general reduction and by efficiently implementing measure-bounded colorful independence oracles, we obtain new improved output-sensitive approximate counting algorithms for k-clique, k-dominating set and k-sum. In graphs with nt k-cliques, for instance, our algorithm (1± є)-approximates the k-clique count in time Õє(nω(k−t−1/3,k−t/3,k−t+2/3) +n2), where ω(a,b,c) is the exponent of na× nb by nb× nc matrix multiplication. For large k and t>2, this is a substantial improvement over prior work, even if ω=2. STOC ’25, Prague, Czechia

Lightweight and Locality-Aware Composition of Black-Box Subroutines

2026-02-01T07:32:14Z

Lightweight and Locality-Aware Composition of Black-Box Subroutines Bansal, Manya; Sharlet, Dillon; Ragan-Kelley, Jonathan; Amarasinghe, Saman Subroutines are essential building blocks in software design: users encapsulate common functionality in libraries and write applications by composing calls to subroutines. Unfortunately, performance may be lost at subroutine boundaries due to reduced locality and increased memory consumption. Operator fusion helps recover performance lost at composition boundaries. Previous solutions fuse operators by manually rewriting code into monolithic fused subroutines, or by relying on heavy-weight compilers to generate code that performs fusion. Both approaches require a semantic understanding of the entire computation, breaking the decoupling necessary for modularity and reusability of subroutines. In this work, we attempt to identify the minimal ingredients required to fuse computations, enabling composition of subroutines without sacrificing performance or modularity. We find that, unlike previous approaches that require a semantic understanding of the computation, most opportunities for fusion require understanding only data production and consumption patterns. Exploiting this insight, we add fusion on top of black-box subroutines by proposing a lightweight enrichment of subroutine declarations to expose data-dependence patterns. We implement our approach in a system called Fern, and demonstrate Fern's benefits by showing that it is competitive with state-of-the-art, high-performance libraries with manually fused operators, can fuse across library and domain boundaries for unforeseen workloads, and can deliver speedups of up to $5\times$ over unfused code.

Prolonged photostability in hexagonal boron nitride quantum emitters

2026-02-01T07:32:52Z

Prolonged photostability in hexagonal boron nitride quantum emitters Li, Sylvia Xin; Ichihara, Takeo; Park, Hyoju; He, Guangwei; Kozawa, Daichi; Wen, Yi; Koman, Volodymyr B; Zeng, Yuwen; Kuehne, Matthias; Yuan, Zhe; Faucher, Samuel; Warner, Jamie H; Strano, Michael S Single-photon emitters are crucial building blocks for optical quantum technologies. Hexagonal boron nitride (hBN) is a promising two-dimensional material that hosts bright, room-temperature single-photon emitters. However, photo instability is a persistent challenge preventing practical applications of these properties. Here, we reveal the ubiquitous photobleaching of hBN vacancy emitters. Independent of the source or the number of hBN layers, we find that the photobleaching of a common emission at 1.98 ± 0.05 eV can be described by two consistent time constants, namely a first bleaching lifetime of 5 to 10 s, and a second bleaching lifetime in the range of 150 to 220 s. Only the former is environmentally sensitive and can be significantly mitigated by shielding O2, whereas the latter could be the result of carbon-assisted defect migration. Annular dark-field scanning transmission electron microscopy of photobleached hBN allows for visualizing vacancy defects and carbon substitution at single atom resolution, supporting the migration mechanism along with X-ray photoelectron spectroscopy. Thermal annealing at 850 °C of liquid exfoliated hBN eliminates both bleaching processes, leading to persistent photostability. These results represent a significant advance to potentially engineer hBN vacancy emitters with the photostability requisite for quantum applications.

Discretized hexagonal boron nitride quantum emitters and their chemical interconversion

2026-02-01T07:32:50Z

Discretized hexagonal boron nitride quantum emitters and their chemical interconversion Kozawa, Daichi; Li, Sylvia Xin; Ichihara, Takeo; Rajan, Ananth Govind; Gong, Xun; He, Guangwei; Koman, Volodymyr B; Zeng, Yuwen; Kuehne, Matthias; Silmore, Kevin S; Parviz, Dorsa; Liu, Pingwei; Liu, Albert Tianxiang; Faucher, Samuel; Yuan, Zhe; Warner, Jamie; Blankschtein, Daniel; Strano, Michael S Quantum emitters in two-dimensional hexagonal boron nitride (hBN) are of significant interest because of their unique photophysical properties, such as single-photon emission at room temperature, and promising applications in quantum computing and communications. The photoemission from hBN defects covers a wide range of emission energies but identifying and modulating the properties of specific emitters remain challenging due to uncontrolled formation of hBN defects. In this study, more than 2000 spectra are collected consisting of single, isolated zero-phonon lines (ZPLs) between 1.59 and 2.25 eV from diverse sample types. Most of ZPLs are organized into seven discretized emission energies. All emitters exhibit a range of lifetimes from 1 to 6 ns, and phonon sidebands offset by the dominant lattice phonon in hBN near 1370 cm−1. Two chemical processing schemes are developed based on water and boric acid etching that generate or preferentially interconvert specific emitters, respectively. The identification and chemical interconversion of these discretized emitters should significantly advance the understanding of solid-state chemistry and photophysics of hBN quantum emission.

Rational Design and Efficacy of Glucose‐Responsive Insulin Therapeutics and Insulin Delivery Systems by Computation Using Connected Human and Rodent Models

2026-02-01T07:32:49Z

Rational Design and Efficacy of Glucose‐Responsive Insulin Therapeutics and Insulin Delivery Systems by Computation Using Connected Human and Rodent Models Yang, Sungyun; Yang, Jing Fan; Gong, Xun; Weiss, Michael A; Strano, Michael S Glucose‐responsive insulins (GRIs) use plasma glucose levels in a diabetic patient to activate a specifically designed insulin analogue to a more potent state in real time. Alternatively, some GRI concepts use glucose‐mediated release or injection of insulin into the bloodstream. GRIs hold promise to exhibit much improved pharmacological control of the plasma glucose concentration, particularly for the problem of therapeutically induced hypoglycemia. Several innovative GRI schemes are introduced into the literature, but there remains a dearth of quantitative analysis to aid the development and optimization of these constructs into effective therapeutics. This work evaluates several classes of GRIs that are proposed using a pharmacokinetic model as previously described, PAMERAH, simulating the glucoregulatory system of humans and rodents. GRI concepts are grouped into three mechanistic classes: 1) intrinsic GRIs, 2) glucose‐responsive particles, and 3) glucose‐responsive devices. Each class is analyzed for optimal designs that maintain glucose levels within the euglycemic range. These derived GRI parameter spaces are then compared between rodents and humans, providing the differences in clinical translation success for each candidate. This work demonstrates a computational framework to evaluate the potential clinical translatability of existing glucose‐responsive systems, providing a useful approach for future GRI development.

Wearable sensors for monitoring marine environments and their inhabitants

2026-02-01T07:32:54Z

Wearable sensors for monitoring marine environments and their inhabitants Kaidarova, Altynay; Geraldi, Nathan R; Wilson, Rory P; Kosel, Jürgen; Meekan, Mark G; Eguíluz, Víctor M; Hussain, Muhammad Mustafa; Shamim, Atif; Liao, Hanguang; Srivastava, Mani; Saha, Swapnil Sayan; Strano, Michael S; Zhang, Xiangliang; Ooi, Boon S; Holton, Mark; Hopkins, Lloyd W; Jin, Xiaojia; Gong, Xun; Quintana, Flavio; Tovasarov, Adylkhan; Tasmagambetova, Assel; Duarte, Carlos M Human societies depend on marine ecosystems, but their degradation continues. Toward mitigating this decline, new and more effective ways to precisely measure the status and condition of marine environments are needed alongside existing rebuilding strategies. Here, we provide an overview of how sensors and wearable technology developed for humans could be adapted to improve marine monitoring. We describe barriers that have slowed the transition of this technology from land to sea, update on the developments in sensors to advance ocean observation and advocate for more widespread use of wearables on marine organisms in the wild and in aquaculture. We propose that large-scale use of wearables could facilitate the concept of an ‘internet of marine life’ that might contribute to a more robust and effective observation system for the oceans and commercial aquaculture operations. These observations may aid in rationalizing strategies toward conservation and restoration of marine communities and habitats.

Operational Fuel Inefficiency in Cruise Flight: A Worldwide Geospatial Analysis

2026-02-01T03:01:19Z

Operational Fuel Inefficiency in Cruise Flight: A Worldwide Geospatial Analysis Trávník, Marek; Hansman, R. John

Low-Cost Deep Learning for Building Detection with Application to Informal Urban Planning

2026-03-08T03:39:48Z

Low-Cost Deep Learning for Building Detection with Application to Informal Urban Planning González, Lucas; Toutouh, Jamal; Nesmachnow, Sergio This article studies the application of deep neural networks for automatic building detection in aerial RGB images. Special focus is put on accuracy robustness in both well-structured and poorly planned urban scenarios, which pose significant challenges due to occlusions, irregular building layouts, and limited contextual cues. The applied methodology considers several CNNs using only RBG images as input, and both validation and transfer capabilities are studied. U-Net-based models achieve the highest single-model accuracy, with an Intersection over Union (𝐼⁢𝑜⁢𝑈) of 0.9101. A soft-voting ensemble of the best U-Net models further increases performance, reaching a best ensemble 𝐼⁢𝑜⁢𝑈 of 0.9665, improving over state-of-the-art building detection methods on standard benchmarks. The approach demonstrates strong generalization using only RGB imagery, supporting scalable, low-cost applications in urban planning and geospatial analysis.

When the Psyche and the 'Net Collide: Sources of and potential methods for preventing Bad Behavior Online

2026-02-11T15:35:02Z

When the Psyche and the 'Net Collide: Sources of and potential methods for preventing Bad Behavior Online Wedeman, Sara; Clark, David D With the emergence of social networking has come a range of harmful and malicious behaviors online, including disinformation, cyberbullying, and sextortion among others. These behaviors arise from a number of causes, including the incentives of the providers of the social networking platforms and the technical affordances of those platforms, which in some cases facilitate these abuses. This report sheds light on the causes and possible mitigations of these behaviors through the lens of behavioral psychology. Results from psychological research suggest that these abuses play on specific human attributes. To design effective mitigations, it is crucial that these human attributes be understood. This report draws on literature from psychology research to outline the important human behavioral attributes, relates these to some of the important affordances found in social networking applications, and suggests possible approaches that can damp the bad behavior we observe online. This is the final report for an NSF-sponsored study of psychology literature related to the online experience, and the drivers and possible mitigations of bad behavior online.

Systems of Visualization for Musical Futures

2026-01-30T03:24:59Z

Systems of Visualization for Musical Futures Naseck, Perry This thesis investigates how large-scale visual systems can communicate the presence, agency, and foresight of improvising musical agents–human and AI–during live performance. We propose a framework for manifesting AI collaborators on stage through five principles: musical transparency, live improvisational reactivity, demonstrated virtuosity, communication for collaboration, and visual fit. Two public performances operationalize these ideas: an addressable-light sculpture that renders harmonic space, and a stage-sized kinetic sculpture built from novel, low-cost Generic Pan Tilt fixtures that visualize the AI’s planned “musical futures.” The latter combines a real-time, MIDI-conditioned, Transformer-based hand-motion model with deterministic, pattern-based mappings that signal states such as resting and regeneration. Audience surveys indicate that viewers perceived links between musical turns and kinetic gestures while requesting clearer explanatory cues. We document the open-source hardware, firmware, and control protocols of the Generic Pan Tilt platform and reflect on design tradeoffs for accessibility, reliability, and expressivity. Finally, we outline a real-time analysis toolchain–motif detection, parallelism, and continuous energy/tension estimators–that emits OSC triggers for lighting, media, kinetic, and spatial-audio systems, enabling reactive shows beyond timecode. Together, these systems advance performable visualizations of human-improvised and AI-driven musical futures.

Design Rules for LLM-Generated Code: A RealWorld Case Study

2026-01-30T03:24:56Z

Design Rules for LLM-Generated Code: A RealWorld Case Study Lawrence, Jennifer M. This thesis conducts a case study exploring the interaction between software design, extensibility, and LLM code generation. The central problem we investigate is whether LLMs violate software design principles in ways that introduce bugs and ultimately hinder extensibility. We examine several repositories belonging to the RealWorld collection, a project that demonstrates combinations of frameworks, database, and programming languages for building full stack web apps modeled on an existing social media application. We create a concept-based implementation of the RealWorld API. Concept Design defines software systems in terms of the abstract purposes and relationships of self-contained units of functionality. It enforces stringent design standards and aims to aid humans better understand complex software behavior. To test code extensibility, we develop three phases of new functionality to be added to the RealWorld API. Each phase is intended to mimic real-world software development, adding functionality that is commonly found in social media platforms while increasing nuance and complexity. The code for these extensions is generated by an AI agent, then reviewed by a human coder who classifies and fixes any bugs. In this study, we examine how LLMs interact with software paradigms like Concept Design, the kinds of design violations they produce, and whether these violations correlate with bugs that impede extensibility.

Cognify: An On-Device, AI-powered Learning Assistant

2026-01-30T03:24:55Z

Cognify: An On-Device, AI-powered Learning Assistant Huang, Siyong Large Language Models (LLMs) have proven highly effective for a wide range of natural language processing tasks, but their size and compute requirements often restrict their use to powerful cloud-based infrastructures. In recent years, significant progress has been made in shrinking LLMs while maintaining performance levels comparable to much larger models. We are approaching the point where the capabilities of massive, multi-billion parameter models can be realistically replicated on consumer-grade devices. This thesis builds upon that foundation by developing an AI-powered note-taking application that runs entirely offline, using only the compute resources available on a personal laptop. The application is designed to listen to lectures alongside the student and provide support in real-time—through transcription, notes generation, and enabling context-aware search. Achieving this level of interactivity locally introduces challenges in reducing end-to-end latency, which this project addresses through both model-level optimizations and the design of efficient prompting and inference algorithms. A demo of the app can be found on Youtube.

Performance Analysis of the Apple AMX Matrix Accelerator

2026-01-30T03:24:55Z

Performance Analysis of the Apple AMX Matrix Accelerator Zhou, Jonathan Apple Silicon integrates a dedicated Apple Matrix Coprocessor (AMX) that executes outer-product style computations with high throughput, but its public programming model remains largely hidden behind the Accelerate framework. This thesis turns AMX into a more predictable and practical target by combining (i) empirical throughput characterization, (ii) a case study on AMX specific matrix multiplication (GEMM) design, and (iii) an interpretable rule-based latency model that predicts cycle counts for short AMX instruction sequences. First, microbenchmarks quantify AMX load/store and compute limits across matrix and vector modes and data types. We analyze throughput in both GFLOPS and AMX instructions per cycle, and also observe output register based throughput limitations. Second, we develop an in-place GEMM that uses masked outer products and strategically overlapping tiles to avoid scratch buffers used by Accelerate, outperforming Accelerate while preserving simplicity. Third, we introduce a compact latency model that decomposes cycles into per-instruction BaseTime, symmetric SwitchLatency for instruction changes, and instruction FullLatency (data dependency) terms. Fitted with non-negative coordinate descent on length-2 loops and validated on length-3 sequences via a lightweight loop simulation, the model obtains reasonably high accuracy while remaining helpful for those trying to understand the architecture.

Weak Identification and Network Measurement Error in Peer Effects Estimation

2026-01-30T03:06:54Z

Weak Identification and Network Measurement Error in Peer Effects Estimation Wang, William Wei The growing availability of social network data has enabled a surge of research on social interactions. In particular, peer effects, once considered unidentifiable, have now been shown to be identified given knowledge of the network structure. Despite this positive result, questions remain about the existence and nature of peer effects, due to concerns about identification strength and the reliability of network data. This work investigates two key threats to the estimation of peer effects: weak identification and network measurement error. We show that weak instrument problems arise in moderately dense networks due to rapid averaging, leading to slow convergence rates even when estimators remain consistent. On the measurement error side, we show that additive edge weight errors can be mitigated in such networks due to the same averaging phenomena, but the error remains a relevant threat to consistency in sparser networks. We further demonstrate that when both issues are present, the resulting estimators exhibit non-vanishing bias, suggesting that the combined effect of weak instruments and measurement error can be more severe than either problem in isolation. Overall, our results aim to clarify how these non-standard estimation challenges impact our ability to study peer effects using network data.

Seeing Beyond Limits with Physics-Informed Priors

2026-01-30T03:06:34Z

Seeing Beyond Limits with Physics-Informed Priors Liu, Yang Conventional imaging systems are limited by dimensionality and visibility: standard sensors capture only two-dimensional data, while light diffuses or scatters across surfaces and through complex media. This dissertation reformulates imaging as an interplay of optical encoding and neural decoding. It models forward physical processes and iteratively refines them using deep denoisers. By embedding physics-informed priors into this optimization, it aims to surpass conventional limits in dimensionality and visibility. First, I develop Privacy Dual Imaging using an ambient light sensor. This approach tackles both dimensionality and visibility challenges when imaging with a single-point, non-imaging component on smart devices. Inspired by 1984’s “Big Brother” telescreen, I demonstrate how subtle light intensity fluctuations can reveal unseen image information; however, the goal is to highlight privacy concerns, not exploit them. It addresses two visibility limits—pixel-less and lens-less imaging—by using the screen as a spatial modulator and exploiting involuntary motion to create a virtual pinhole effect. A quantized, physics-informed prior improves reconstruction from heavily quantized sensor measurements. Second, I propose Snapshot Compressive Imaging (SCI) augmented with deep plug-and-play physics-informed priors to overcome the dimensionality limit of 2D sensors. SCI compressively encodes multiple temporal, spectral, or angular frames into a single measurement. A deep plug-and-play prior algorithm introduces high-dimensional priors learned from images and videos into the iterative reconstruction process, improving fidelity, speed, and flexibility. Experiments show notable gains in reconstruction quality and efficiency across different SCI datasets, including largeformat 4K UHD scenarios. Third, I introduce Rank-Reduced physics-informed priors, showing that large pretrained AI models—especially diffusion models—can act as general visual priors across both dimensionality and visibility challenges. A relax-then-tighten strategy handles ill-conditioning by applying truncated singular value decomposition to reduce rank deficiencies, followed by a Stable Diffusion refiner (SDEdit) plug-and-play prior that constrains reconstructions to valid image spaces. Simulations and passive non-line-of-sight imaging experiments verify the approach’s stability and effectiveness. Physics-informed priors promise to extend the boundaries of imaging, enabling us to see beyond current dimensionality and visibility limits and to unlock new applications from macro-scale to micro-scale observations.

Optimizing Large Language Models from a Data SystemsPerspective

2026-01-30T03:24:54Z

Optimizing Large Language Models from a Data SystemsPerspective Chen, Peter Baile Strong retrieval and reasoning capabilities are essential for large language models (LLMs) to effectively handle a broad spectrum of downstream tasks, such as open-domain question answering and solving math or science problems. While current LLM-based frameworks achieve strong performance on complex retrieval and reasoning tasks, they do so at a high computational cost. Additionally, they often lack structured, systematic problem-solving strategies, leading to unexpected failures. In particular, these models typically operate in an iterative, online, and isolated fashion—failing to exploit relationships across data sources, opportunities for offline computation, and the benefits of reusability—resulting in less-than-optimal outcomes. In contrast, traditional data management systems are engineered for both efficiency and accuracy, with careful coordination across all stages of the query pipeline. Inspired by these principles, this work proposes novel approaches to improve LLMbased retrieval and reasoning by incorporating optimization techniques from data systems. Our evaluation across a range of knowledge- and reasoning-intensive datasets demonstrates significant gains in both accuracy and computational efficiency.

Foundational Abstractions for Quantum Programming

2026-01-30T03:06:30Z

Foundational Abstractions for Quantum Programming Yuan, Charles Bringing the promise of quantum computation into reality requires not only building a quantum computer but also correctly programming it to run a quantum algorithm. To obtain asymptotic advantage over classical algorithms for applications including simulation, search, and optimization, quantum algorithms rely on the ability of data in quantum superposition to exhibit phenomena such as interference and entanglement. In turn, an implementation of the algorithm as a program must correctly orchestrate these phenomena in the states of qubits. Otherwise, it would yield incorrect outputs or lose quantum computational advantage. Given a quantum algorithm, what are the challenges and costs of realizing it as a program that can run on a physical quantum computer? In this thesis, I answer this question by showing how the basic abstractions of programming upon which many quantum algorithms rely – such as data structures and control flow – can fail to work correctly or efficiently on a quantum computer. I then demonstrate how we can leverage insights from research in programming languages to re-invent the software stack – including abstractions, libraries, and compilers – to meet the demands of quantum algorithms. This approach holds out a promise of expressive and efficient tools to program a quantum computer and thereby practically realize its computational advantage.

Fabrication of Superconducting Reflectionless Filters for Quantum Microwave Circuits

2026-01-30T03:24:51Z

Fabrication of Superconducting Reflectionless Filters for Quantum Microwave Circuits Bui, Eric The performance and scalability of superconducting quantum circuits depends critically on the microwave environment. Minimizing signal reflections and suppressing thermal noise are essential for achieving high-fidelity readout and preserving qubit coherence. A significant challenge arises from the use of conventional cryogenic components such as isolators and circulators, which exhibit nonideal out-of-band reflection characteristics. Reflections degrade impedance matching and limit the performance of broadband quantum limited amplifiers. Superconducting implementations of reflectionless microwave filters offer a promising solution to mitigate these issues. The focus of this work is the fabrication and cryogenic characterization of reflectionless filters compatible with superconducting qubit fabrication flows. Devices were implemented on high resistivity silicon substrates using aluminum ground planes, integrated nichrome resistors, and crossovers formed with SiO2 interlayer dielectric. Cryogenic measurements at 20 mK demonstrate high return loss, confirming the viability of these filters for co-fabrication with traveling-wave parametric amplifiers (TWPAs) and circuit quantum electrodynamics (cQED) architectures. The filters exhibit low insertion loss in the passband to maintain quantum measurement efficiency and provide broadband reflection suppression across frequencies relevant to superconducting qubits, offering a scalable way to manage microwave noise in superconducting quantum processors.

CONDOR: Clinical Ontology-aware Networked Data Organization and Retrieval

2026-01-30T03:24:23Z

CONDOR: Clinical Ontology-aware Networked Data Organization and Retrieval Dongo Aguirre, Gyalpo Melchisedeck Until now, state-of-the-art research into AI-driven clinical workflows has been confined to proprietary, closed-source systems from vendors like Epic and Oracle, or private experiments like Stanford’s ChatEHR, creating a critical barrier to academic innovation. This thesis introduces CONDOR, the first fully open-source and replicable research environment designed to simulate an agentic, conversational AI interacting with a high-fidelity Electronic Health Record (EHR). By integrating an open-source, FHIR-native EHR (Medplum) with a complex, realistic public clinical dataset (MIMIC-IV FHIR), CONDOR provides a foundational testbed that has been previously unavailable to the research community. The framework’s primary contribution is a novel alignment and evaluation methodology that adapts the principles of SelfCite to the clinical domain. We propose a ‘ClinicalConfidence‘ score to quantify the trustworthiness of generated statements and programmatically generate a high-quality preference dataset for alignment using Simple Preference Optimization (SimPO). We compare a standard vector-based Retrieval-Augmented Generation (RAG) baseline against a more advanced GraphRAG architecture that leverages a two-tiered knowledge graph of patient data and medical ontologies. Our results demonstrate that the full CONDOR system, combining GraphRAG with SimPO alignment, significantly improves citation quality and verifiability, establishing a new open-source benchmark for the development of safe and reliable clinical AI.

Multi-Stage LLM Reasoning for Automated Detection and Classification of High-Impact Misinformation

2026-01-30T03:24:50Z

Multi-Stage LLM Reasoning for Automated Detection and Classification of High-Impact Misinformation Nair, Anushka Manchanda As of 2025, social platforms have become a primary news source, magnifying the reach of misleading content [1]. Exposure to misinformation has been linked to shifts in public attitudes and behavior, including vaccine uptake [2] and voting behaviors [3]. However, current misinformation detection approaches can often focus on a narrow definition of misinformation: factual claims that can be clearly judged as true or false. However, recent research suggests the problem lies elsewhere: overt falsehoods (“vaccines contain microchips”) can carry little harm, while technically accurate but decontextualized narratives can be more influential. Allen et al. (2024) [4] found that factually accurate ”vaccine-skeptical” content had a much greater impact on vaccine hesitancy than misinformation flagged by fact-checkers. These narratives can work by omitting information, misleading framing, or cherry-picked evidence, forms of manipulation that can elude traditional fact-checking. Though professional fact-checkers are often able to recognize these tactics and the broader context of information, they cannot keep pace with the volume of online content. This thesis designs a Large Language Model (LLM) based pipeline meant to partner with, rather than replace, human fact checkers. The system decomposes content into its explicit and implicit claims, rhetorical tactics, and the “missing context” questions it raises; retrieves evidence from fact-check databases and reliable sources; and synthesizes grounded explanations while assigning calibrated harm scores to guide triage. Evaluated on fact-checked tweets, the pipeline matched expert judgments in 92.6% of cases where experts agreed, and flagged for review posts where experts disagreed, a gray zone requiring human judgment. The system’s explanations ranked higher than crowdsourced Community Notes in helpfulness, clarity, and trustworthiness when assessed by an LLM, and harm evaluations aligned with human reviewers in 87.5% of cases, enabling prioritization of content with greatest potential impact. Despite constraints of sample size and processing latency, the results demonstrate the feasibility of a human–AI workflow that treats disagreement as a signal and directs scarce attention towards high-impact misinformation that current automated systems can miss.

Learning Simple Chemical Heuristics to Model and Discover Materials

2026-01-30T03:07:04Z

Learning Simple Chemical Heuristics to Model and Discover Materials Ma, Andrew Computational approaches have long played an important role in the field of materials science, driving both the scientific study of materials’ fundamental properties and the design of materials for technological applications. Currently, mainstream methods in computational materials science typically rely on either first-principles calculations or deep learning models. In this thesis, we take a different direction by developing remarkably simple data-driven models for predicting fundamental properties of materials, including electronic topology, metallicity, and band gap. These models take the form of highly interpretable chemical heuristics. A key finding of this work is the surprising result that electronic topology diagnosis – often regarded as a highly complex task – can, in fact, be performed heuristically using a simple and intuitive model. We further integrate this model into a workflow for discovering new topological materials. Altogether, this work revisits the classic idea of chemical heuristics through a modern data-driven lens, shedding new light on fundamental problems in materials science.

Spatial Emission and Polarization Control in Integrated Photonics for Optical-Trapping and Trapped-Ion Systems

2026-01-30T03:24:49Z

Spatial Emission and Polarization Control in Integrated Photonics for Optical-Trapping and Trapped-Ion Systems Sneh, Tal Recent advances in silicon photonics have yielded impressive results in fields including biophotonic optical tweezers and trapped-ion quantum systems. However, the majority of these demonstrations, while offering advantages in size, cost, and dense integration, lag behind their bulk-optic counterparts, limited by a lack of critical advanced functionality such as spatial control of light in the near field or polarization control at visible wavelengths. This thesis addresses this gap by designing and experimentally demonstrating the first, to the best of our knowledge, cell experiments using single-beam integrated optical tweezers, chip-based 3D printers, and integrated polarization rotators and splitters at blue wavelengths. First, we demonstrate optical trapping and tweezing of microspheres using a nearfield-focusing integrated optical phased array, at a standoff distance over two orders of magnitude larger than prior integrated demonstrations. We then use this system to perform the first cell experiments using single-beam integrated optical tweezers. Second, we use a tunable integrated optical phased array operating at red wavelengths to print designs in a visible-light-curing resin, demonstrating the first chip-based 3D printer. Third, we design and experimentally demonstrate the first integrated polarization rotators and splitters operating at blue wavelengths, enabling polarization control on chip for sophisticated integrated manipulation of trapped-ion and neutral-atom quantum systems. Finally, we develop key polarization-diverse integrated-photonics devices and utilize them to implement a variety of integrated-photonics-based polarization-gradient-cooling systems, culminating in the first demonstration of polarization-gradient cooling of a trapped ion by an integrated-photonics-based system.

Language Modeling from Visually Grounded Speech

2026-01-30T03:06:52Z

Language Modeling from Visually Grounded Speech Lai, Cheng-I Jeff Recent advancements in spoken language processing have significantly reduced automatic speech recognition (ASR) error rates, driven by large-scale supervised training on paired speech–text data and, more recently, self-supervised pre-training on unpaired speech and audio. These methods have facilitated robust transfer learning across diverse speech and audio tasks. However, fully leveraging multimodal inputs, particularly visual context, remains underexplored. This thesis addresses this gap by developing novel language modeling techniques directly from visually grounded speech. We first introduce the Audio-Visual Neural Syntax Learner (AV-NSL), an unsupervised parser that recovers constituency trees directly from raw speech paired with images, demonstrating how visual context effectively bootstraps grammar induction without textual supervision. Next, we investigate Audio-Visual Word Discovery for Speech Translation, using the Fisher Spanish–English corpus to train a series of speech-to-speech translation models based on pseudo-word units discovered via audio-visual grounding. This study highlights that simplistic acoustic tokens and limited training data degrade re-synthesis and translation quality, underscoring two crucial missing ingredients: richer semantic tokens and large-scale training. Guided by these insights, we present Audio-Visual Gemma (AV-Gemma), a family of multimodal foundation models that condition jointly on images and learned semantic speech tokens. At scale, AV-Gemma generates visually coherent spoken captions and transfers robustly to tasks such as video-to-speech generation and spoken visual question answering, significantly advancing multimodal spoken-language processing.

ALPACA: An Algorithmic Pipeline for Automated Contour Annotation of Carnatic Music: A Dynamic Programming Framework for Pitch Segmentation and Note Transcription

2026-01-30T03:24:45Z

ALPACA: An Algorithmic Pipeline for Automated Contour Annotation of Carnatic Music: A Dynamic Programming Framework for Pitch Segmentation and Note Transcription Parthasarathi, Sruthi In recent years, a wide range of computational techniques have been developed to extract information from recorded performances of Western music. However, these methods often achieve limited success when applied to non-Western musical traditions. Carnatic music, in particular, poses unique challenges due to the absence of a standardized notation system and the lack of a consistent mapping between frequency bands and note categories. This project introduces a dynamic programming–based transcription framework, incorporating novel methods for label estimation, contour segmentation, and related subtasks, and establishes the foundations for end-to-end automatic transcription of this art form.

Modeling Diverse Treatment Policies from Observational Health Data

2026-01-30T03:24:53Z

Modeling Diverse Treatment Policies from Observational Health Data Ejilemele, Abe Learning policies for real world tasks often requires modeling human behavior, especially in domains like healthcare and driving. In these settings, skills are learned from expert human demonstrations, but such data are typically multimodal, violating the common single expert assumption. We study sequential clinical treatment decision making in the offline imitation learning setting, where environment interaction is prohibited, reflecting the challenges of experimentation in safety critical domains. Existing methods for multi expert offline imitation learning often restrict the latent space, underspecify its structure, or omit objective terms that prevent latent collapse and encourage behavior discovery. We propose a fully offline approach that addresses these shortcomings and improves learning from multi expert demonstrations through modifications to the formulation of the latent approximate posterior and the model architecture. We suggest that our method is more robust to real world settings where the true number of demonstrators may not be known. We also incorporate an occupancy matching term into our objective that injects awareness of the rollout distribution over trajectories into our behavior cloning objective. We evaluate our method against baselines on both simulated multi expert demonstrations from an extended S-CVSim and real world demonstrations from MIMIC. Our approach achieves consistently higher next step action prediction and behavior discovery performance. While ground truth expert policies are unavailable for MIMIC, visual analysis shows our method uncovers clinically meaningful variations in expert strategies, reflecting treatment population diversity.

Scalable Assembly of General Objects

2026-01-30T03:06:32Z

Scalable Assembly of General Objects Tian, Yunsheng In this thesis, I present a scalable system towards fully automated and flexible robotic assembly that generalizes over diverse geometries and complex structures. Most real-world objects are assemblies composed of multiple parts. Assembly presents significant challenges for robots to execute long-horizon, contact-rich manipulation with both reliability and generalization. However, most manufacturing facilities today still rely heavily on manually programmed assembly lines, which require significant labor, time, and setup costs yet offer no flexibility to object variations. My proposed system synergizes global multi-step planning with local reactive learning-based control to enable generalizable and precise assembly. Such an integrated paradigm effectively leverages the best of both worlds, accomplishing results that neither planning nor learning could achieve alone. For planning, I leverage guidance from physical simulation and learned feasibility networks to efficiently search for part sequences, precise motions, and stable grasps for dual-arm robots over long horizons. For learningbased control, I train robust policies via reinforcement learning for submillimeter-level insertion across different part geometries, assembly paths, and grasp poses. I introduce and open-source the largest-scale assembly dataset to date and demonstrate my system’s generalization on thousands of simulated assemblies as well as through end-to-end real robot experiments. By integrating planning and learning, I showcase the first system to achieve complete and generalizable real-world multi-part assembly without domain knowledge or human demonstrations. Although the system plans and learns purely in simulation, it transfers zero-shot to the real world and achieves 80% successful steps. Finally, I will share insights that further scale up robotic assembly and opportunities to extend to general manipulation, and discuss future directions to equip general-purpose robots with multi-step, precise manipulation capabilities.

Towards a Modular Superconducting Quantum Processor using Chiral Waveguide Quantum Electrodynamics

2026-01-30T03:24:46Z

Towards a Modular Superconducting Quantum Processor using Chiral Waveguide Quantum Electrodynamics Yankelevich, Beatriz As the field of superconducting quantum computing advances, networking qubits within a single system becomes essential for building modular processors. Modularity allows the system to circumvent scalability constraints and enable architectures and computational schemes that exploit non-local connectivity to enhance processing capabilities. This work proposes non-local entanglement generation methods based on the theory of chiral quantum waveguide dynamics, which is the quantum-optical framework that describes systems of atoms coupled non-reciprocally to a continuum of modes. We leverage these effects to design a chiral communication module composed of multiple superconducting qubits, capable of both directional single photon routing and the realization of chiral, driven-dissipative entanglement protocols.

Fine-tuning Boltz for Antibody-Antigen Binding Prediction

2026-01-30T03:24:32Z

Fine-tuning Boltz for Antibody-Antigen Binding Prediction Kim, Ji Won Accurate prediction of antibody-antigen binding is a central challenge in computational immunology. Its direct implication for therapeutic antibody design and vaccine development has made it one of the most rapidly growing fields. Recent advances in protein language models and structure prediction have provided new tools for modeling, yet these approaches often fall short in capturing the fine-grained features that drive binding specificity in antibody and antigens. This thesis evaluates multiple strategies for improving predictive performance. First, we investigate a custom multiple sequence alignment (MSA) experiment. Standard Boltz-2 training relies on MSAs from broad protein databases, which capture global diversity but under-represent lineage-specific constraints. To address this, we constructed antibody-specific MSAs to test whether restricting the search space to antibody repertoires improves model learning. Unfortunately, gains in downstream binding prediction were limited, suggesting that further work needs to be done in training models for specific databases in the first place. Our second line of investigation focused on fine-tuning Boltz-2, a generative structural foundation model, using curated antibody–antigen data. By leveraging Boltz-2’s internal sequence embeddings, we trained a predictive model for binding affinity. This approach yielded stronger ROC performance compared to baseline models, achieving a validation AUROC of 0.645, demonstrating the advantages of structural generative priors for antibody–antigen binding prediction.

Deterministic Circuit Range Avoidance is (Likely) Intractable

2026-01-30T03:24:47Z

Deterministic Circuit Range Avoidance is (Likely) Intractable Ilango, Rahul Circuit Range Avoidance (denoted Avoid) is a computational problem where, given a Boolean circuit with more output bits than input bits, one must output a string outside of the range of the circuit. A simple counting argument implies that such a string must always exist and also guarantees that outputting a uniformly random string is correct with good probability. A natural question is whether this can be derandomized: does there exist an efficient deterministic algorithm for Avoid? We give the first evidence that deterministically solving Avoid is intractable. We show that there is no polynomial-time algorithm for Avoid under plausible assumptions in complexity theory and cryptography. Specifically, our assumptions are that NP ≠ coNP and that subexponentially-secure indistinguishability obfuscation exists.

Learning to Tackle Task Variations in Control - A Transportation Context

2026-01-30T03:06:23Z

Learning to Tackle Task Variations in Control - A Transportation Context Jayawardana, Vindula Muthushan Real-world control tasks are messy and often exhibit task variations. Practical solutions to these problems must exhibit generalization across task variations. For example, in the task of controlling traffic signals, control strategies must adapt to different intersection topologies (the variations), each with distinct dynamics. In this thesis, we consider the challenge of coping with task variations in the context of transportation problems, specifically in roadway interventions where many such variations are both common and imperative to handle. We develop machine learning techniques to address three key challenges: 1) quantify the impact of task variations in control, 2) model them to align with the real world, and 3) optimize in the presence of them. To this end, we begin with a large-scale case study of cooperative eco-driving and illustrate how explicitly modeling task variations can surface otherwise overlooked insights. Building on this, we argue for the necessity of formally incorporating task variations into problem specifications, emphasizing that task underspecification due to loosely defined task variations can severely impair decision-making. We then introduce a contextual reinforcement learning algorithm capable of leveraging the structure of task variations to generalize effectively in cooperative eco-driving with autonomous vehicles. We also present IntersectionZoo, a benchmark designed to promote the development of learning algorithms that generalize by exploiting task variation structures, thus standardizing progress in the field. Last, we explore task variation modeling through a generative modeling lens, using human driver behavior modeling as a case study. Overall, this thesis lays the groundwork for robust control methods by leveraging machine learning to tackle task variations, specifically in roadway intervention designs.

An Empirical Evaluation of LLMs for the Assessment of Subjective Qualities

2026-01-30T03:24:42Z

An Empirical Evaluation of LLMs for the Assessment of Subjective Qualities Ranade, Esha Large Language Models (LLMs) have achieved remarkable success in natural language processing tasks and are increasingly being used for language generation. Significant advancements in this field have unlocked capabilities that enable their adoption in sophisticated roles, including acting as evaluators or "judges" of text for various attributes such as factuality, relevance, fluency, and reasoning quality. However, their understanding and ability to assess subjective attributes, such as the level of formality in a piece of writing, and produce content matching these subjective attributes remains unclear and underexplored. This research develops a methodology to study how LLMs evaluate subjective attributes. It has three primary contributions: (i) a reproducible user study to generate human-annotated labels for different attributes, (ii) an analysis of the extent to which different LLMs provide subjective labels aligned with human annotators, and (iii) an analysis of the extent to which LLMs generate content aligned with specified intended subjective labels, relative to humans. The user study and the analyses have been conducted both with and without a reference scale. The scale itself, the survey design, and the evaluation questions have all undergone multiple rounds of iteration informed by study tester feedback to improve clarity, consistency, and reliability for the final study. Comparisons between human-generated ratings and LLM-generated ratings for both human-generated content and LLM-generated content reveal the extent to which LLMs align with human judgment, providing insights into their capabilities and limitations. While humans typically do better in their roles, LLMs are able to attain reliably high levels of success in producing and judging text, despite tending to err on the more-formal side. Both groups’ performance increases significantly with the aid of a formalized reference scale. Across the suite of models tested, OpenAI’s GPT family leads overall performance, with Anthropic’s Claude and Meta’s LLaMA series showing notable strengths in specific formality ranges. Although this work focuses on the formality attribute of text, the methodology developed can be used to evaluate other subjective qualities of text, such as conciseness, usefulness, or persuasiveness. Ultimately, these findings may guide future efforts to fine-tune LLMs to produce text that more precisely matches the desired stylistic or ethical standards.

Accelerating Burst Parallelism of SigmaOS processes with CRIU

2026-01-30T03:24:29Z

Accelerating Burst Parallelism of SigmaOS processes with CRIU Tang, Frederick σOS is a multi-tenant cloud operating system designed to integrate the agility of serverless environments with the interactivity of microservices. A goal of achieving this integration is the ability to start new instances of server processes quickly. However, σOS only handles σcontainer initialization, and does not assist with runtime and app initialization costs. One approach to overcome this challenge is to checkpoint processes using Checkpoint/Restore in Userspace (CRIU). CRIU is a linux toolset which can start new server instances by restoring them from a saved checkpointed state, avoiding the full cost of reinitialization and setup. This thesis introduces σCRIU, which adapts CRIU for burst-parallel spawning of microservices in σOS. σCRIU implements a number of optimizations: compressing checkpointed proc metadata to reduce network communication costs, implementing demand-paging using a lazy page service, and caching kernel metadatadata to reduce CRIU’s restore operation latency. These optimizations allow σCRIU to start new microservices on remote machines quickly while still making use of CRIU’s existing proven checkpoint and restore technology.

Modern methods for causal inference and missing data

2026-01-30T03:06:38Z

Modern methods for causal inference and missing data Xia, Eric The proliferation of data-driven approaches in a wide array of settings is one of the defining characteristic of the modern era. With this rise, there has been much focus on using data to answer causal questions, e.g. whether A causes a change in B. Furthermore aspects of data collection has given rise to datasets that are often quite messy, sometimes missing important entries. These are both problems that are incredibly relevant to practitioners in a variety of disciplines, including policy-makers looking to make critical decisions that can influence lives of many. On the surface these problems seem quite distinct, yet the literature has highlighted deep connections between these two settings. Indeed, many methods for addressing one question can often be repurposed to address the other. These two settings are quite classical and approaches to address the are still quite so, but there has been great interest recently to develop techniques and algorithms to address them that harness modern developments in statistics and machine learning. This thesis contributes to the literature by providing new methods as well as novel understandings of existing ones.

Visually Accurate Database-Enabled Reconstructions of Scenes (VADERS)

2026-01-30T03:24:39Z

Visually Accurate Database-Enabled Reconstructions of Scenes (VADERS) Gosalia, Mehek This work introduces a novel pipeline for scene reconstruction that jointly prioritizes semantic accuracy and visual fidelity, addressing a gap in current approaches. Prior pipelines often emphasize either semantic analysis or photorealistic rendering, but rarely both. This method combines scene analysis, segmentation, and retexturing to yield reconstructions that preserve structural semantics, while convincingly reflecting the visual qualities of the original image. The motivation lies in the limitations of existing systems. Existing databaseassisted approaches depend on proprietary datasets that restrict stylistic diversity or using in-the-wild assets. This constrains expressiveness and often produces results that are visually misaligned. Conversely, pipelines optimized for visual realism neglect semantic correctness, generating outputs that may appear plausible but lack categorical or structural grounding. Our framework addresses this by first enforcing semantic accuracy via selecting database assets, then editing those assets to be stylistically faithful to the reference, producing reconstructions that are both interpretable and expressive. We begin with database-assisted scene analysis, using an open-source asset database containing chairs, lamps, sofas, tables, and benches. Input images are depth-mapped, segmented, and parsed into object masks, which are matched to database assets based on semantic labels and visual correspondence. Each asset is broken into semantic segments and rescaled per-component using vision-language model predictions to match the reference object better. Finally the asset is retextured based on the image mask of the reference object in the input image. Evaluation on six diverse scenes—both photographs and artworks—shows the pipeline produces semantically grounded, visually accurate reconstructions under non-research conditions. Future work will focus on expanding the asset database, reducing reliance on proprietary texturing, and releasing an open-source implementation to broaden accessibility.

Designing Planar Silicon Solar Cells for Singlet Fission Sensitization

2026-01-30T03:24:48Z

Designing Planar Silicon Solar Cells for Singlet Fission Sensitization Wang, Janet Z. Singlet fission (SF)-sensitized silicon (Si) solar cells offer a path towards surpassing the Shockley-Queisser efficiency limit for single-junction solar cells. However, realizing efficient charge transfer from the SF material to Si remains a significant challenge that requires careful interface engineering. Prior work showed that Si microwire cells sensitized with tetracene (Tc) and a zinc phthalocyanine (ZnPc) donor layer can boost photocurrent and external quantum efficiency (EQE). Planar devices are simpler to fabricate than microwire devices and reproduce the planar geometry of optical test samples to connect studies of the interface to device performance. This thesis integrates modeling and experimental approaches to guide the design of planar SF-sensitized Si solar cells. We developed a fabrication process for planar cells comparing varied oxide passivation layer growth conditions and surface treatments, Si(100) versus Si(111) orientation, and junctions formed by diffusion doping versus ion implantation. Complementary surface photovoltage (SPV) measurements on matching optical stacks show evidence of an illumination-induced transient positive charge density at the Tc/ZnPc/oxide/Si interface, consistent with increased field effect passivation. We find that SPV responses on AlOx/n-Si are dominated by substrate band bending; consequently, SiOx is the preferred passivation to suppress the background and isolate the SPV signals driven by the organics. A drift–diffusion model shows that the diffusion doping (exponential) emitters reduce surface recombination rates compared to ion implantation (Gaussian) emitters. We also show that a positive fixed charge density at the surface enhances short wavelength EQE, with the effect strongest for Gaussian emitters. Together, these results provide practical design rules for planar SF-sensitized Si cells and the study of charge transfer at organic-Si interfaces.

Microsecond Time Synchronization for Computing Fiber Networks

2026-01-30T03:24:43Z

Microsecond Time Synchronization for Computing Fiber Networks Li, Jenny Y. We present a microsecond-accurate time synchronization method and time localization system for a sensor network of spatially-separated, low-power Bluetooth nodes, with the goal of integrating this system into thermally-drawn computing fibers. Each node consists of an nRF54L15 SoC paired with an ICS-43434 digital I2S microphone, enabling synchronized audio data collection. Our design leverages Bluetooth LE connection events to synchronize local clocks with sub-10 µs accuracy across a multi-peripheral topology; we trigger precise, CPU-independent hardware events to timestamp audio samples. We demonstrate that timestamped I2S data stored in external SPI flash can be correlated across devices to extract TDoA measurements for localizing sound sources. Cross-correlation techniques allow us to estimate direction and position, with localization errors reduced from 4.17 m to 0.39 m through clock synchronization. This prototype provides a roadmap for embedding synchronized sensing and computation within fibers and smart textiles, with implications for on-body audio perception and distributed sensing in flexible electronics.

From String to Structure: Graph Threading for Physical Assembly

2026-01-30T03:24:37Z

From String to Structure: Graph Threading for Physical Assembly Lin, Rebecca Y. E. Many artistic and engineering applications—from beadwork to deployable structures—create intricate, and sometimes dynamic, designs by threading cord through tubular components. We model the underlying design challenge—threading tubes so that they achieve a target connectivity when the string is pulled taut—as graph threading. In this formulation, tubes and their junctions correspond to edges and vertices of a graph, and the goal is to find a closed walk that induces a connected graph at every vertex while avoiding U-turns. We study two optimization objectives motivated by fabrication and deployment: minimizing length to reduce material cost and assembly time, and minimizing turn to reduce frictional resistance during deployment. For the length metric, we present a polynomial-time algorithm via reduction to minimum-weight perfect matching, prove tight worst-case bounds on optimal threadings, and identify special cases with faster algorithms. For the turn metric, we characterize the complexity landscape, proving NP-hardness for graphs of maximum degree 4, tractability for degree 3, and giving exact and approximation algorithms for restricted variants, including rectangular grid graphs. Finally, we turn from theory to fabrication, proposing multi-configuration threading—a new approach for achieving multiple predetermined configurations within a single system. As in earlier chapters, framing the problem in graph-theoretical terms provides access to powerful problem-solving techniques, guiding both algorithmic analysis and physical design.

Steering Vision at Scale: From the Model Weights to Training Data

2026-01-30T03:06:51Z

Steering Vision at Scale: From the Model Weights to Training Data Materzyńska, Joanna We study the interpretability and controllability of multimodal and generative models, with a particular focus on text–image representation models and text-to-image diffusion systems. We begin by addressing limitations in CLIP’s multimodal embeddings, specifically the entanglement between visual and textual concepts within images. We demonstrate the consequences of this entanglement in both generative and discriminative tasks, and introduce a method for disentangling visual and textual representations. We showcase the utility of these disentangled embeddings in typographic attack resistance, improved image generation, and robust out-of-domain OCR detection. Building on this foundation, we explore methods to enhance the controllability of diffusion models. First, we tackle the challenge of unwanted concept generation. We introduce a technique to remove specific visual concepts using only their names, leveraging negative prompts and guidance to suppress target content without modifying training data or requiring model retraining. This approach enhances ethical alignment and enables greater user control in generative systems. We then turn to the complementary problem: incorporating new concepts. We present a few-shot motion customization technique for video generation models, which transfers motion patterns from a small set of examples to novel subjects. This method maintains the generalization capabilities of the base model while enabling consistent, subject-agnostic animation that preserves both identity and temporal coherence. To improve the fine-grained control of visual outputs, we propose a method for continuous manipulation of image attributes. This framework introduces smooth, intuitive controls, that allow for dynamic, continuous steering of generated images. Unlike prompt engineering or token-level interventions, our approach offers real-time adjustment without sacrificing output realism. Finally, we examine whether artistic styles in diffusion models require large-scale pretraining or can be learned in a lightweight, post-training manner. To this end, we train a base model on art-free data and introduce a compact adapter method that learns stylistic concepts from a small set of exemplar artworks. Our findings suggest that artistic domains can be integrated efficiently and ethically, without reliance on web-scale scraped datasets.

A Data Attribution-Based Approach to Model Diagnosis in LC-MS/MS Structure Prediction

2026-01-30T03:24:40Z

A Data Attribution-Based Approach to Model Diagnosis in LC-MS/MS Structure Prediction Khoo, Ling Min Serena Elucidating the structure of small molecules from complex mixtures using liquid chromatography tandem mass spectrometry (LC-MS/MS) is a challenging task with far-reaching implications in many areas such as drug discovery, environmental science and metabolism research. Yet, despite its importance and significant efforts to develop machine learning (ML) models for the task of elucidating the molecular structures of unknown compounds from LC-MS/MS spectra, the performance of these ML-based models remains limited. As a result, the performance of current ML-based models has been reported as insufficient for practical applications, thereby warranting a deeper investigation into their limitations to advance ML-based molecular structure elucidation from LC-MS/MS and enable their utility in real-world settings. Here, we leverage data attribution methods to systematically identify and validate hypotheses about the sources of generalization challenges that hinder current model performance. Our goal is to automatically uncover insights into the failure modes of existing ML models for LC-MS/MS, thereby laying the foundation for developing more robust and accurate models.

Modeling Dynamic Objects in Scenes with Generative Particle Systems

2026-01-30T03:24:31Z

Modeling Dynamic Objects in Scenes with Generative Particle Systems Li, Eric Humans readily interpret the motion of deformable and rigid bodies, even when encountering unfamiliar objects with minimal shape or texture cues. In such cases, motion serves as a critical signal for recognition and understanding. Inspired by this ability, we propose a generative model that represents 3D matter as small Gaussians (“particles”) drawn from clusters capturing groups of coherently moving matter. We develop an e!cient inference algorithm based on parallelized block Gibbs sampling to recover stable particle motion and rigid groupings. Our model provides a tractable, object-centric generalization of as-rigidas-possible (ARAP) regularizers used in motion tracking. To assess alignment with human perceptual judgments, we test our approach on random dot kinematograms—sparse motion displays in which dot trajectories convey latent object structure, often used to probe visual understanding of motion and grouping. In this setting, our approach captures human-like responses, including graded patterns of uncertainty across ambiguous conditions. Applied to naturalistic RGB videos, it infers dense particle representations that track object motion and deformation over time. These results demonstrate that our model enables persistent latent scene structure suitable for object-level reasoning.

The Arm Qubit for Faster, Higher Fidelity Readout and Gates

2026-01-30T03:24:44Z

The Arm Qubit for Faster, Higher Fidelity Readout and Gates Kline, Jeremy B. Currently, superconducting qubit processors are bottlenecked by errors during two-qubit gates, readout, and idle time. All three error contributions could be reduced if we improved the speed of operations (without introducing additional leakage errors) compared to the qubit lifetime. Readout and two-qubit gates are multimode interactions and therefore are limited by the coupling strength between the modes. In this thesis, we introduce a two-mode superconducting qubit which uses one mode to facilitate strong coupling to other modes of the quantum processor and one mode to store data with high coherence. Simulations show that this architecture could enable order-of-magnitude reductions in error during readout and two-qubit gates.

Clustering Algorithms for Component Placement in Printed Circuit Boards

2026-01-30T03:24:41Z

Clustering Algorithms for Component Placement in Printed Circuit Boards Petrusenko, Vlada In 2024, approximately 12 billion printed circuit boards (PCBs) were manufactured globally [1], with the trend increasing gradually, and the majority of PCB layouts still being completed manually. The manual design process amounts to millions of hours of tedium that can be eased with automation. One of the biggest challenges is that the complex Printed Circuit Board designs typically have hundreds, sometimes thousands of components and even more net connections between them. This makes both manual and automated placement very time-consuming. As a way to improve placement performance, in this thesis, we constructed a custom weighted undirected graph representation of components and nets for any board that would encode physical and electrical constraints. Additionally, we integrated the Louvain and Leiden clustering algorithms for component clustering in PCB placement. We also showed comparative metrics with the spectral clustering algorithm applied to unweighted graph representations, which is the prior state of this project, but it has no knowledge of electrical and physical constraints associated with PCB designs and would thus produce results that require more manual correction. This new clustering approach was able to generate more optimal clustering and reduced average runtime by 51.05%, decreased estimated length of routing by 7.72%, and improved component association score by 12.8%.

How Data Drives ML Models Performance

2026-01-30T03:06:27Z

How Data Drives ML Models Performance Khaddaj, Alaa Data has been been playing an increasingly more important role in the machine learning (ML) pipeline. This thesis deepens the understanding of the effect of the data on model performance and reliability. First, we study how choice of training data affects model performance. We consider a transfer learning setting and present a framework for selecting from a large pool of data a pretraining subset that improves model performance on downstream tasks. Our approach, however, requires training multiple target models which becomes prohibitively expensive at large-scale. To that end, we explore using smaller—and cheaper—proxy models to approximate large model behavior and select the pretraining data using that cheaper model. We show the effectiveness of this approach in two dataset selection settings: language modeling and imitation learning. Second, we explore the role of data in model reliability and consider two different threat models: backdoor attacks and malicious data editing. In this first threat model, an adversary injects a few doctered samples into the training set to control model predictions at inference time. We study the effect of these malicious samples on model behavior and then propose a framework for detecting and removing them from the training data. In the second threat model, an adversary leverages generative models such as diffusion models to maliciously modify personal data and generate harmful digital content. We focus on image editing and investigate how we can imperceptibly modify personal images to mitigate editing using diffusion models and raise and the cost of hamrful content generation. Overall, this thesis contributes to the understanding of the role of the data in driving model behavior. Through these efforts, we aim to provide mechanisms for (i) training models that perform better and (ii) are more reliable when deployed in the real world.

Efficient Learning and Computation of Linear Correlated Equilibrium in General Convex Games

2026-01-27T04:50:46Z

Efficient Learning and Computation of Linear Correlated Equilibrium in General Convex Games Daskalakis, Constantinos; Farina, Gabriele; Fishelson, Maxwell; Pipis, Charilaos; Schneider, Jon We propose efficient no-regret learning dynamics and ellipsoid-based methods for computing linear correlated equilibria—a relaxation of correlated equilibria and a strengthening of coarse correlated equilibria—in general convex games. These are games where the number of pure strategies is potentially exponential in the natural representation of the game, such as extensive-form games. Our work identifies linear correlated equilibria as the tightest known notion of equilibrium that is computable in polynomial time and is efficiently learnable for general convex games. Our results are enabled by a generalization of the seminal framework of Gordon et al. for Φ-regret minimization, providing extensions to this framework that can be used even when the set of deviations Φ is intractable to separate/optimize over. Our polynomial-time algorithms are similarly enabled by extending the Ellipsoid-Against-Hope approach of Papadimitriou and Roughgarden and its generalization to games of non-polynomial type proposed by Farina and Pipis. We provide an extension to these approaches when we do not have access to the separation oracles required by these works for the dual player. Constantinos Daskalakis, Gabriele Farina, Maxwell Fishelson, Charilaos Pipis, and Jon Schneider. 2025. Efficient Learning and Computation of Linear Correlated Equilibrium in General Convex Games. In Proceedings of the 57th Annual ACM Symposium on Theory of Computing (STOC '25). Association for Computing Machinery, New York, NY, USA, 542–553.

When Connectivity Is Hard, Random Walks Are Easy with Non-determinism

2026-01-27T04:50:34Z

When Connectivity Is Hard, Random Walks Are Easy with Non-determinism Doron, Dean; Pyne, Edward; Tell, Roei; Williams, R. Ryan Two fundamental problems on directed graphs are to decide s-t connectivity, and to estimate the behavior of random walks. Currently, there is no known algorithm for s-t connectivity running in polynomial time and no(1) space, and no known algorithm for estimating the n-step random walk matrix running in non-deterministic logspace. We show that for every directed graph, at least one of these problems is solvable in time and space that significantly improve on the respective state-of-the-art. In particular, there is a pair of algorithms A1 and A2 such that for every graph G, either: A1(G) outputs the transitive closure of G in polynomial time and polylogarithmic space. A2(G) outputs an approximation of the n-step random walk matrix of G in non-deterministic logspace. As one application, we show surprisingly tight win-win results for space-bounded complexity. For example, for certain parameter regimes, either Savitch’s theorem can be non-trivially sped up, or randomized space can be almost completely derandomized. We also apply our techniques to significantly weaken the assumptions required to derandomize space-bounded computation, and to make non-deterministic space-bounded computation unambiguous. Specifically, we deduce such conclusions from lower bounds against uniform circuits of polynomial size, which is an exponential improvement on the required hardness in previous works (Doron–Pyne–Tell STOC 2024, Li–Pyne–Tell FOCS 2024). We further show similar results for minimal-memory derandomization (Doron–Tell CCC 2024). To prove these results, we substantially improve the array of technical tools introduced in recent years for studying hardness-vs.-randomness for bounded-space computation. In particular, we develop derandomized distinguish-to-predict transformations for new types of distinguishers (corresponding to compositions of PRGs with weak distinguishers), we construct a derandomized logspace reconstruction procedure for the Shaltiel–Umans generator (JACM 2005) that can compress hard truth-tables to polylogarithmic size, and we design a version of the Chen–Tell generator (FOCS 2021) that is particularly suitable for the space-bounded setting. STOC ’25, Prague, Czechia

QMA vs QCMA and Pseudorandomness

2026-01-27T04:50:44Z

QMA vs QCMA and Pseudorandomness Liu, Jiahui; Mutreja, Saachi; Yuen, Henry We study a longstanding question of Aaronson and Kuperberg on whether there exists a classical oracle separating QMA from QCMA. Settling this question in either direction would yield insight into the power of quantum proofs over classical proofs. We show that such an oracle exists if a certain quantum pseudorandomness conjecture holds. Roughly speaking, the conjecture posits that quantum algorithms cannot, by making few queries, distinguish between the uniform distribution over permutations versus permutations drawn from so-called “dense” distributions. Our result can be viewed as establishing a “win-win” scenario: either there is a classical oracle separation of QMA from QCMA, or there is quantum advantage in distinguishing pseudorandom distributions on permutations. STOC ’25, Prague, Czechia

Semantics of Integrating and Differentiating Singularities

2026-01-27T04:50:47Z

Semantics of Integrating and Differentiating Singularities Michel, Jesse; Lee, Wonyeol; Yang, Hongseok A singular function is a partial function such that at one or more points, the left and/or right limit diverge (e.g., the function 1/x). Since programming languages typically support division, programs may denote singular functions. Although on its own, a singularity may be considered a bug, introducing a division-by-zero error, singular integrals—a version of the integral that is well-defined when the integrand is a singular function and the domain of integration contains a singularity—arise in science and engineering, including in physics, aerodynamics, mechanical engineering, and computer graphics. In this paper, we present the first semantics of a programming language for singular integration. Our differentiable programming language, SingularFlow, supports the evaluation and differentiation of singular integrals. We formally define the denotational semantics of SingularFlow, deriving all the necessary mathematical machinery so that this work is rigorous and self-contained. We then define an operational semantics for SingularFlow that estimates integrals and their derivatives using Monte Carlo samples, and show that the operational semantics is a well-behaved estimator for the denotational semantics. We implement SingularFlow in JAX and evaluate the implementation on a suite of benchmarks that perform the finite Hilbert transform, an integral transform related to the Fourier transform, which arises in domains such as physics and electrical engineering. We then use SingularFlow to approximate the solutions to four singular integral equations—equations where the unknown function is in the integrand of a singular integral—arising in aerodynamics and mechanical engineering.

SoS Certificates for Sparse Singular Values and Their Applications: Robust Statistics, Subspace Distortion, and More

2026-01-27T04:50:50Z

SoS Certificates for Sparse Singular Values and Their Applications: Robust Statistics, Subspace Distortion, and More Diakonikolas, Ilias; Hopkins, Samuel B.; Pensia, Ankit; Tiegel, Stefan We study sparse singular value certificates for random rectangular matrices. If M is a d × n matrix with independent Gaussian entries, we give a new family of polynomial-time algorithms which can certify upper bounds on the maximum of ||M u||, where u is a unit vector with at most η n nonzero entries for a given η ∈ (0,1). This basic algorithmic primitive lies at the heart of a wide range of problems across algorithmic statistics and theoretical computer science, including robust mean and covariance estimation, certification of distortion of random subspaces of n, certification of the 2 → p norm of a random matrix, and sparse principal component analysis. Our algorithms certify a bound which is asymptotically smaller than the naive one, given by the maximum singular value of M, for nearly the widest-possible range of n,d, and η. Efficiently certifying such a bound for a range of n,d and η which is larger by any polynomial factor than what is achieved by our algorithm would violate lower bounds in the statistical query and low-degree polynomials models. Our certification algorithm makes essential use of the Sum-of-Squares hierarchy. To prove the correctness of our algorithm, we develop a new combinatorial connection between the graph matrix approach to analyze random matrices with dependent entries, and the Efron-Stein decomposition of functions of independent random variables. As applications of our certification algorithm, we obtain new efficient algorithms for a wide range of well-studied algorithmic tasks. In algorithmic robust statistics, we obtain new algorithms for robust mean and covariance estimation with tradeoffs between breakdown point and sample complexity, which are nearly matched by statistical query and low-degree polynomial lower bounds (that we establish). We also obtain new polynomial-time guarantees for certification of ℓ1/ℓ2 distortion of random subspaces of n (also with nearly matching lower bounds), sparse principal component analysis, and certification of the 2→ p norm of a random matrix. STOC ’25, Prague, Czechia

Weak Recovery, Hypothesis Testing, and Mutual Information in Stochastic Block Models and Planted Factor Graphs

2026-01-27T04:51:10Z

Weak Recovery, Hypothesis Testing, and Mutual Information in Stochastic Block Models and Planted Factor Graphs Mossel, Elchanan; Sly, Allan; Sohn, Youngtak The stochastic block model is a canonical model of communities in random graphs. It was introduced in the social sciences and statistics as a model of communities, and in theoretical computer science as an average case model for graph partitioning problems under the name of the “planted partition model.” Given a sparse stochastic block model, the two standard inference tasks are: (i) Weak recovery: can we estimate the communities with non-trivial overlap with the true communities? (ii) Detection/Hypothesis testing: can we distinguish if the sample was drawn from the block model or from a random graph with no community structure with probability tending to 1 as the graph size tends to infinity? In this work, we show that for sparse stochastic block models, the two inference tasks are equivalent except at a critical point. That is, weak recovery is information theoretically possible if and only if detection is possible. We thus find a strong connection between these two notions of inference for the model. We further prove that when detection is impossible, an explicit hypothesis test based on low-degree polynomials in the adjacency matrix of the observed graph achieves the optimal statistical power. This low-degree test is efficient as opposed to the likelihood ratio test, which is not known to be efficient. Moreover, we prove that the asymptotic mutual information between the observed network and the community structure exhibits a phase transition at the weak recovery threshold. Our results are proven in much broader settings including the hypergraph stochastic block models and general planted factor graphs. In these settings, we prove that the impossibility of weak recovery implies contiguity and provide a condition that guarantees the equivalence of weak recovery and detection. STOC ’25, Prague, Czechia

Near-Optimal Time-Sparsity Trade-Offs for Solving Noisy Linear Equations

2026-01-27T04:50:49Z

Near-Optimal Time-Sparsity Trade-Offs for Solving Noisy Linear Equations Bangachev, Kiril; Bresler, Guy; Tiegel, Stefan; Vaikuntanathan, Vinod We present a polynomial-time reduction from solving noisy linear equations over in dimension Θ(klogn/(logk,logq,loglogn)) with a uniformly random coefficient matrix to noisy linear equations over in dimension n where each row of the coefficient matrix has uniformly random support of size k. This allows us to deduce the hardness of sparse problems from their dense counterparts. In particular, we derive hardness results in the following canonical settings: • Assuming the ℓ-dimensional (dense) learning with errors () problem over a polynomial-size field takes time 2Ω(ℓ), k-sparse in dimension n takes time nΩ(k/(logk · (logk + loglogn))) . • Assuming the ℓ-dimensional (dense) learning parity with noise () problem over ℤ/2ℤ takes time 2Ω(ℓ/logℓ), k-sparse in dimension n takes time nΩ(k/(logk · (logk + loglogn)2)) . These running time lower bounds are nearly tight as both sparse problems can be solved in time nO(k), given sufficiently many samples. Our reduction allows us to derive several consequences in cryptography and the computational complexity of statistical problems. In addition, as a new application, we give a reduction from k-sparse LWE to noisy tensor completion. Concretely, composing the two reductions implies that order-k rank-2k−1 noisy tensor completion in ℝn⊗ k takes time nΩ(k/ logk · (logk + loglogn)), assuming the exponential hardness of standard worst-case lattice problems. STOC ’25, Prague, Czechia

Stochastic Matching via In-n-Out Local Computation Algorithms

2026-01-27T04:50:43Z

Stochastic Matching via In-n-Out Local Computation Algorithms Azarmehr, Amir; Behnezhad, Soheil; Ghafari, Alma; Rubinfeld, Ronitt Consider the following stochastic matching problem. We are given a known graph G=(V, E). An unknown subgraph Gp = (V, Ep) is realized where Ep includes every edge of E independently with some probability p ∈ (0, 1]. The goal is to query a sparse subgraph H of G, such that the realized edges in H include an approximate maximum matching of Gp. This problem has been studied extensively over the last decade due to its applications in kidney exchange, online dating, and online labor markets. For any fixed є > 0, [BDH STOC’20] showed that any graph G has a subgraph H with (1/p) = (1/p)(log(1/p)) maximum degree, achieving a (1−є)-approximation. A major open question is the best approximation achievable with (1/p)-degree subgraphs. A long line of work has progressively improved the approximation in the (1/p)-degree regime from .5 [BDH+ EC’15] to .501 [AKL EC’17], .656 [BHFR SODA’19], .666 [AB SOSA’19], .731 [BBD SODA’22] (bipartite graphs), and most recently to .68 [DS ’24]. In this work, we show that a (1/p)-degree subgraph can obtain a (1−є)-approximation for any desirably small fixed є > 0, achieving the best of both worlds. Beyond its quantitative improvement, a key conceptual contribution of our work is to connect local computation algorithms (LCAs) to the stochastic matching problem for the first time. While prior work on LCAs mainly focuses on their out-queries (the number of vertices probed to produce the output of a given vertex), our analysis also bounds the in-queries (the number of vertices that probe a given vertex). We prove that the outputs of LCAs with bounded in- and out-queries (in-n-out LCAs for short) have limited correlation, a property that our analysis crucially relies on and might find applications beyond stochastic matching STOC ’25, Prague, Czechia

Colloidal State Machines as Smart Tracers for Chemical Reactor Analysis

2026-01-24T03:10:54Z

Colloidal State Machines as Smart Tracers for Chemical Reactor Analysis Zhang, Ge; Yang, Jing Fan; Yang, Sungyun; Brooks, Allan M; Koman, Volodymyr B; Gong, Xun; Strano, Michael S A widely utilized tool in reactor analysis is passive tracers that report the residence time distribution, allowing estimation of the conversion and other properties of the system. Recently, advances in microrobotics have introduced powered and functional entities with sizes comparable to some traditional tracers. This has motivated the concept of Smart Tracers that could record the local chemical concentrations, temperature, or other conditions as they progress through reactors. Herein, the design constraints and advantages of Smart Tracers by simulating their operation in a laminar flow reactor model conducting chemical reactions of various orders are analyzed. It is noted that far fewer particles are necessary to completely map even the most complex concentration gradients compared with their conventional counterparts. Design criteria explored herein include sampling frequency, memory storage capacity, and ensemble number necessary to achieve the required accuracy to inform a reactor model. Cases of severe particle diffusion and sensor noise appear to bind the functional upper limit of such probes and require consideration for future design. The results of the study provide a starting framework for applying the new technology of microrobotics to the broad and impactful set of problems classified as chemical reactor analysis.

Mitigation of ventilation air methane (VAM) using novel methanotrophic coating materials: a technical analysis

2026-01-24T03:11:06Z

Mitigation of ventilation air methane (VAM) using novel methanotrophic coating materials: a technical analysis Lundberg, Daniel James; Kim, Jimin; Parviz, Dorsa; Strano, Michael S Ventilation air methane (VAM) is a potent greenhouse gas source originating from geological wells, current and extinct mineshafts and other terrestrial conduits venting methane to the atmosphere, contributing to global methane emissions and disproportionate warming potential. Herein, we introduce the concept of the methanotrophic material as an engineering solution. Such materials should be capable of converting methane at ambient temperatures and pressures to a binder product, capturing and permanently sequestering the methane while simultaneously restricting its further emission. While such materials are currently under research development, this goal is supported and facilities by the mathematical framework, introduced and used herein, to evaluate the ability to convert methane, using currently published activity data. We include a case study of the conversion of a characteristic stream of VAM (0.6% methane in air, 1.7 × 108 l hr−1 equivalent to 100 000 standard cubic feet per minute). We show that when appropriately designed, such systems require a surface coverage of less than 1000 m of mine tunnel length (equivalent to 20 000 m2 areal coverage) in order to reduce the methane emission from this stream by over 99%. Finally, we highlight formaldehyde as a reactive intermediate of methane oxidation which may itself be incorporated into these coating materials. As a component of binders and polymers already used ubiquitously in commercial products, this intermediate ultimately allows these systems to sequester the carbon from methane in a stable and solid form. The results presented here are easily extended to the treatment of other methane streams—either more concentrated or dilute—and the results herein will guide the design and development of a new class of carbon-negative materials.

Synergistic multi-source ambient RF and thermal energy harvester for green IoT applications

2026-01-24T03:11:05Z

Synergistic multi-source ambient RF and thermal energy harvester for green IoT applications Bakytbekov, Azamat; Nguyen, Thang Q; Zhang, Ge; Strano, Michael S; Salama, Khaled N; Shamim, Atif In a future green Internet of Things (IoT) reality, billions of devices of the IoT infrastructure should be self-powered. Harvesting ambient energy to power IoT devices is an attractive solution that can extend battery life or can completely replace batteries. Considering the global applications of IoT, ubiquitous and continuous availability is an important requirement for ambient energy sources. Radio frequency (RF) energy from mobile phone towers and thermal energy from diurnal cycle temperature fluctuations are good candidates. In this study, we present a synergistic multi-source energy harvester (MSEH) comprising an RF energy harvester (RFEH) and a thermal energy harvester (TEH) integrated through a dual-function component, heatsink antenna. Both harvesters collect ambient energy 24 h a day and are not location specific. The TEH, which is in the shape of a box, collects energy using heatsinks on its sidewalls. The same heatsinks are optimized to also serve as receiving antennas of the RFEH, which collects energy from the GSM900, GSM1800, and 3G bands. Due to the synergistic integration, radiation efficiency of the antenna doubled from 40% to 80% which resulted in ∼ 10% increase in power conversion efficiency of the RFEH. Similarly, the average power of the TEH without heatsinks 120 μ W is doubled to 240 μ W for TEH with heatsinks. Field tests have shown that the outputs of the TEH and RFEH have increased 4 and 3 times compared to the independent TEH and RFEH respectively. A temperature and humidity sensor based IoT node has been successfully powered through this energy harvesting system. Overall, the MSEH can collect 3680 μ W h of energy per day which is sufficient to obtain the sensors data with a time interval of 3.5 s.

Chromatic covalent organic frameworks enabling in-vivo chemical tomography

2026-01-24T03:11:03Z

Chromatic covalent organic frameworks enabling in-vivo chemical tomography Wang, Song; Han, Yangyang; Reddy, Vaishnavi Amarr; Ang, Mervin Chun-Yi; Sánchez-Velázquez, Gabriel; Saju, Jolly Madathiparambil; Cao, Yunteng; Khong, Duc Thinh; Jayapal, Praveen Kumar; Cheerlavancha, Raju; Loh, Suh In; Singh, Gajendra Pratap; Urano, Daisuke; Rajani, Sarojam; Marelli, Benedetto; Strano, Michael S Covalent organic frameworks designed as chromatic sensors offer opportunities to probe biological interfaces, particularly when combined with biocompatible matrices. Particularly compelling is the prospect of chemical tomography – or the 3D spatial mapping of chemical detail within the complex environment of living systems. Herein, we demonstrate a chromic Covalent Organic Framework (COF) integrated within silk fibroin (SF) microneedles that probe plant vasculature, sense the alkalization of vascular fluid as a biomarker for drought stress, and provide a 3D in-vivo mapping of chemical gradients using smartphone technology. A series of Schiff base COFs with tunable pKa ranging from 5.6 to 7.6 enable conical, optically transparent SF microneedles with COF coatings of 120 to 950 nm to probe vascular fluid and the surrounding tissues of tobacco and tomato plants. The conical design allows for 3D mapping of the chemical environment (such as pH) at standoff distances from the plant, enabling in-vivo chemical tomography. Chromatic COF sensors of this type will enable multidimensional chemical mapping of previously inaccessible and complex environments.

Decoding early stress signaling waves in living plants using nanosensor multiplexing

2026-01-24T03:11:01Z

Decoding early stress signaling waves in living plants using nanosensor multiplexing Ang, Mervin Chun-Yi; Saju, Jolly Madathiparambil; Porter, Thomas K; Mohaideen, Sayyid; Sarangapani, Sreelatha; Khong, Duc Thinh; Wang, Song; Cui, Jianqiao; Loh, Suh In; Singh, Gajendra Pratap; Chua, Nam-Hai; Strano, Michael S; Sarojam, Rajani Increased exposure to environmental stresses due to climate change have adversely affected plant growth and productivity. Upon stress, plants activate a signaling cascade, involving multiple molecules like H2O2, and plant hormones such as salicylic acid (SA) leading to resistance or stress adaptation. However, the temporal ordering and composition of the resulting cascade remains largely unknown. In this study we developed a nanosensor for SA and multiplexed it with H2O2 nanosensor for simultaneous monitoring of stress-induced H2O2 and SA signals when Brassica rapa subsp. Chinensis (Pak choi) plants were subjected to distinct stress treatments, namely light, heat, pathogen stress and mechanical wounding. Nanosensors reported distinct dynamics and temporal wave characteristics of H2O2 and SA generation for each stress. Based on these temporal insights, we have formulated a biochemical kinetic model that suggests the early H2O2 waveform encodes information specific to each stress type. These results demonstrate that sensor multiplexing can reveal stress signaling mechanisms in plants, aiding in developing climate-resilient crops and pre-symptomatic stress diagnoses.

Polymeric Nanocarriers Autonomously Cross the Plant Cell Wall and Enable Protein Delivery for Stress Sensing

2026-01-24T03:10:59Z

Polymeric Nanocarriers Autonomously Cross the Plant Cell Wall and Enable Protein Delivery for Stress Sensing Zhang, Yilin; Cao, Yunteng; Jiang, Wenzhi; Ma, Qingquan; Shin, Jinwoo; Sun, Hui; Cui, Jianqiao; Chen, Yongsheng; Giraldo, Juan Pablo; Strano, Michael S; Lowry, Gregory V; Sheen, Jen; Marelli, Benedetto Delivery of proteins in plant cells can facilitate the design of desired functions by modulation of biological processes and plant traits but is currently limited by narrow host range, tissue damage, and poor scalability. Physical barriers in plants, including cell walls and membranes, limit protein delivery to desired plant tissues. Herein, a cationic high aspect ratio polymeric nanocarriers (PNCs) platform is developed to enable efficient protein delivery to plants. The cationic nature of PNCs binds proteins through electrostatic. The ability to precisely design PNCs’ size and aspect ratio allowed us to find a cutoff of ≈14 nm in the cell wall, below which cationic PNCs can autonomously overcome the barrier and carry their cargo into plant cells. To exploit these findings, a reduction‐oxidation sensitive green fluorescent protein (roGFP) is deployed as a stress sensor protein cargo in a model plant Nicotiana benthamiana and common crop plants, including tomato and maize. In vivo imaging of PNC‐roGFP enabled optical monitoring of plant response to wounding, biotic, and heat stressors. These results show that PNCs can be precisely designed below the size exclusion limit of cell walls to overcome current limitations in protein delivery to plants and facilitate species‐independent plant engineering.

Analysis of Glucose Responsive Glucagon Therapeutics using Computational Models of the Glucoregulatory System

2026-03-08T03:39:38Z

Analysis of Glucose Responsive Glucagon Therapeutics using Computational Models of the Glucoregulatory System Alizadehmojarad, Ali A; Yang, Sungyun; Gong, Xun; Strano, Michael S Glucose‐responsive glucagon (GRG) therapeutics are a promising technology for reducing the risk of severe hypoglycemia as a complication of diabetes mellitus. Herein, the performance of candidate GRGs in the literature by modeling the kinetics of activation and connecting them as input into physiological glucoregulatory models is evaluated and projected the two distinct GRG designs, experimental results reported in Wu et al. (GRG‐I) and Webber et al. (GRG‐II) is considered. Both are evaluated using a multi‐compartmental glucoregulatory model (IMPACT) and used to compare in‐vivo experimental data of therapeutic performance in rats and mice. For GRG‐I and GRG‐II, the total integrated glucose material balances are overestimated by 41.5% ± 14% and underestimated by 24.8% ± 16% compared to in‐vivo time‐course data, respectively. These large differences to the relatively simple computational descriptions of glucagon dynamics in the model, which underscores the urgent need for improved glucagon models is attributed. Additionally, therapeutic insulin and glucagon infusion pumps are modeled for type 1 diabetes mellitus (T1DM) human subjects to extend the results to additional datasets. These observations suggest that both the representative physiological and non‐physiological models considered in this work require additional refinement to successfully describe clinical data that involve simultaneous, coupled insulin, glucose, and glucagon dynamics.

A Microrobotic Design for the Spontaneous Tracing of Isochemical Contours in the Environment

2026-03-08T03:39:42Z

A Microrobotic Design for the Spontaneous Tracing of Isochemical Contours in the Environment Brooks, A Merritt; Yang, Sungyun; Kang, Byung Ha; Strano, Michael S Microrobotic platforms hold significant potential to advance a variety of fields, from medicine to environmental sensing. Herein, minimally functional robotic entities modeled on readily achievable state-of-the-art features in a modern lab or cleanroom are computationally simulated. Inspired by Dou and Bishop (Phys Rev Res. 2019;1(3):1–5), it is shown that the simple combination of unidirectional steering connected to a single environmental (chemical) sensor along with constant propulsion gives rise to highly complex functions of significant utility. Such systems can trace the contours orthogonal to arbitrary chemical gradients in the environment. Also, pairs of such robots that are additionally capable of emitting the same chemical signal are shown to exhibit coupled relative motion. When the pair has unidirectional steering in opposite directions within the 2D plane (i.e., counter-rotating), they move in parallel trajectories to each other. Alternatively, when steering is in the same direction (corotation), the two move in the same epicyclical trajectory. In this way, the chirality of the unidirectional steering produces two distinct emergent phenomena. The behavior is understood as a ratchet mechanism that exploits the differential in the radii of curvature corresponding to different spatial locations. Applications to environmental detection, remediation, and monitoring are discussed.

Electrokinetic Motion of Neurotransmitter Ions through a 1.01 nm Diameter Single-Walled Carbon Nanotube

2026-03-08T03:39:40Z

Electrokinetic Motion of Neurotransmitter Ions through a 1.01 nm Diameter Single-Walled Carbon Nanotube Ellison, Mark D; Allen, Jacqueline; Bonfiglio, Michael; Seeburger, Matthew; Setenet, Jean; DiGinto, Biagio; Bonanny, Harrison; Russell, Aaliyah; Baird, David; Davis, Liana; McCarthy, Ella; Manley, Alyson; Blatt, Sarah; Lippe, David; Ragone, Daniel; Dyer, Brock; Osgood, Jillian; Strano, Michael S The transport of cations of the neurotransmitters acetylcholine, choline, and dopamine through a 1.01 nm-diameter, 1.1 mm-long single-walled carbon nanotube (SWNT) has been studied for the first time. As a comparison, sodium and aniline ion transport was also investigated. All of these ions exhibited significantly enhanced electrophoretic mobilities over bulk transport. The electrophoretic mobilities of acetylcholine, choline, and sodium were found to depend on pH, specifically increasing as pH decreases. This result is explained by hydrogen ions saturating the surface charges of the SWNT. Conversely, dopamine and aniline have mobilities that do not depend on pH. This difference is attributed to the benzene ring and the size of these ions. An analysis of the time required for an ion to traverse the nanotube shows that the ions adsorb to and desorb from the walls as they pass through the tube. Acetylcholine, choline, and sodium show desorption rate constants that decrease with increasing pH, whereas dopamine and aniline have rate constants that remain constant over different pH values. This is consistent with the relationship between adsorption and desorption rate constants and mobility from an adsorption/desorption kinetic model.

Advancements in Plant Diagnostic and Sensing Technologies

2026-03-08T03:39:42Z

Advancements in Plant Diagnostic and Sensing Technologies Krishnamoorthi, Shalini; Koh, Sally Shuxian; Ang, Mervin Chun‐Yi; Teo, Mark Ju Teng; Jie, Randall Ang; Dinish, US; Strano, Michael S; Urano, Daisuke Recent advancements in plant sensing technologies have significantly improved agricultural productivity while reducing resource inputs, resulting in higher yields by enabling early disease detection, precise diagnostics, and optimized fertilizer and pesticide applications. Each adopted technology offers unique advantages suitable for various farm operations, breeding programs, and laboratory research. This review article first summarizes key target traits, endogenous structures, and metabolites that serve as focal points for plant diagnostic and sensing technologies. Next, conventional plant sensing technologies based on light reflectance and fluorescence, which rely on foliar phytopigments and fluorophores such as chlorophylls are discussed. These methods, along with advanced analytical strategies incorporating machine learning, enable accurate stress detection and classification beyond general assessments of plant health and stress status. Advanced optical techniques such as Fourier transform infrared spectroscopy (FT‐IR) and Raman spectroscopy, which allow specific measurements of various plant metabolites and structural components are then highlighted. Furthermore, the design and applications of nanotechnology chemical sensors capable of highly sensitive and selective detection of specific phytochemicals, including phytohormones and signaling second messengers, which regulate physiological and developmental processes at micro‐ to sub‐micromolar concentrations are introduced. By selecting appropriate sensing methodologies, agricultural production, and relevant research activities can be significantly improved.

The Jacobi Factoring Circuit: Quantum Factoring with Near-Linear Gates and Sublinear Space and Depth

2026-03-08T03:22:28Z

The Jacobi Factoring Circuit: Quantum Factoring with Near-Linear Gates and Sublinear Space and Depth Kahanamoku-Meyer, Gregory D.; Ragavan, Seyoon; Vaikuntanathan, Vinod; Van Kirk, Katherine We present a compact quantum circuit for factoring a large class of integers, including some whose classical hardness is expected to be equivalent to RSA (but not including RSA integers themselves). Most notably, we factor n-bit integers of the form P2 Q with logQ = Θ(na) for a ∈ (2/3, 1) in space and depth sublinear in n (specifically, O(logQ)) using O(n) quantum gates; for these integers, no known classical algorithms exploit the relatively small size of Q to run asymptotically faster than general-purpose factoring algorithms. To our knowledge, this is the first polynomial-time circuit to achieve sublinear qubit count for a classically-hard factoring problem. We thus believe that factoring such numbers has potential to be the most concretely efficient classically-verifiable proof of quantumness currently known. Our circuit builds on the quantum algorithm for squarefree decomposition discovered by Li, Peng, Du, and Suter (Nature Scientific Reports 2012), which relies on computing the Jacobi symbol in quantum superposition. The technical core of our contribution is a new space-efficient quantum algorithm to compute the Jacobi symbol of A mod B, in the regime where B is classical and much larger than A. Our circuit for computing the Jacobi symbol generalizes to related problems such as computing the greatest common divisor and modular inverses, and thus could be of independent interest. STOC ’25, Prague, Czechia

Classical Commitments to Quantum States

2026-03-08T03:22:25Z

Classical Commitments to Quantum States Gunn, Sam; Tauman Kalai, Yael; Natarajan, Anand; Vill?nyi, ?gi We define the notion of a classical commitment scheme to quantum states, which allows a quantum prover to compute a classical commitment to a quantum state, and later open each qubit of the state in either the standard or the Hadamard basis. Our notion is a strengthening of the measurement protocol from Mahadev (STOC 2018). We construct such a commitment scheme from the post-quantum Learning With Errors (LWE) assumption, and more generally from any noisy trapdoor claw-free function family that has the distributional strong adaptive hardcore bit property (a property that we define in this work). Our scheme is succinct in the sense that the running time of the verifier in the commitment phase depends only on the security parameter (independent of the size of the committed state), and its running time in the opening phase grows only with the number of qubits that are being opened (and the security parameter). As a corollary we obtain a classical succinct argument system for QMA under the post-quantum LWE assumption. Previously, this was only known assuming post-quantum secure indistinguishability obfuscation. As an additional corollary we obtain a generic way of converting any X/Z quantum PCP into a succinct argument system under the quantum hardness of LWE. STOC ’25, Prague, Czechia

Symmetric Perceptrons, Number Partitioning and Lattices

2026-03-08T03:22:22Z

Symmetric Perceptrons, Number Partitioning and Lattices Vafa, Neekon; Vaikuntanathan, Vinod The symmetric binary perceptron (SBPκ) problem with parameter κ : ℝ≥1 → [0,1] is an average-case search problem defined as follows: given a random Gaussian matrix A ∼ N(0,1)n × m as input where m ≥ n, output a vector x ∈ {−1,1}m such that || A x ||∞ ≤ κ(m/n) · √m . The number partitioning problem (NPPκ) corresponds to the special case of setting n=1. There is considerable evidence that both problems exhibit large computational-statistical gaps. In this work, we show (nearly) tight average-case hardness for these problems, assuming the worst-case hardness of standard approximate shortest vector problems on lattices. • For SBPκ, statistically, solutions exist with κ(x) = 2−Θ(x) (Aubin, Perkins and Zdeborová, Journal of Physics 2019). For large n, the best that efficient algorithms have been able to achieve is a far cry from the statistical bound, namely κ(x) = Θ(1/√x) (Bansal and Spencer, Random Structures and Algorithms 2020). The problem has been extensively studied in the TCS and statistics communities, and Gamarnik, Kızıldağ, Perkins and Xu (FOCS 2022) conjecture that Bansal-Spencer is tight: namely, κ(x) = Θ(1/√x) is the optimal value achieved by computationally efficient algorithms. We prove their conjecture assuming the worst-case hardness of approximating the shortest vector problem on lattices. • For NPPκ, statistically, solutions exist with κ(m) = Θ(2−m) (Karmarkar, Karp, Lueker and Odlyzko, Journal of Applied Probability 1986). Karmarkar and Karp’s classical differencing algorithm achieves κ(m) = 2−O(log2 m) . We prove that Karmarkar-Karp is nearly tight: namely, no polynomial-time algorithm can achieve κ(m) = 2−Ω(log3 m), once again assuming the worst-case subexponential hardness of approximating the shortest vector problem on lattices to within a subexponential factor. Our hardness results are versatile, and hold with respect to different distributions of the matrix A (e.g., i.i.d. uniform entries from [0,1]) and weaker requirements on the solution vector x. STOC ’25, Prague, Czechia

DNF Learning via Locally Mixing Random Walks

2026-03-08T03:22:21Z

DNF Learning via Locally Mixing Random Walks Alman, Josh; Nadimpalli, Shivam; Patel, Shyamal; Servedio, Rocco A. We give two results on PAC learning DNF formulas using membership queries in the challenging “distribution-free” learning framework, where learning algorithms must succeed for an arbitrary and unknown distribution over {0,1}n. (1) We first give a quasi-polynomial time “list-decoding” algorithm for learning a single term of an unknown DNF formula. More precisely, for any target s-term DNF formula f = T1 ∨ ⋯ ∨ Ts over {0,1}n and any unknown distribution D over {0,1}n, our algorithm, which uses membership queries and random examples from D, runs in quasipoly(n,s) time and outputs a list L of candidate terms such that with high probability some term Ti of f belongs to L. (2) We then use result (1) to give a quasipoly(n,s)-time algorithm, in the distribution-free PAC learning model with membership queries, for learning the class of size-s DNFs in which all terms have the same size. Our algorithm learns using a DNF hypothesis. The key tool used to establish result (1) is a new result on “locally mixing random walks,” which, roughly speaking, shows that a random walk on a graph that is covered by a small number of expanders has a non-negligible probability of mixing quickly in a subset of these expanders. STOC ’25, Prague, Czechia

Near Optimal Constant Inapproximability under ETH for Fundamental Problems in Parameterized Complexity

2026-03-08T03:22:48Z

Near Optimal Constant Inapproximability under ETH for Fundamental Problems in Parameterized Complexity Bafna, Mitali; Karthik C. S.; Minzer, Dor We prove that under the Exponential Time Hypothesis (ETH), for every ε > 0, there exists a constant C > 0 such that no algorithm running in time nk / logC k can determine whether a given 2-CSP instance with k variables, O(k) constraints, and alphabet size n, is perfectly satisfiable or if every assignment satisfies at most an ε fraction of the constraints. By known reductions in the literature, the above result implies near-optimal conditional lower bounds for approximating a host of parameterized problems, such as the k-Clique problem, k-Max-Coverage problem, k-Unique Set Cover problem, k-Median and k-Means problems, parameterized variants of the Nearest Codeword problem, Minimum Distance of a Code problem, Closest Vector problem, and Shortest Vector problem. We also establish a densification theorem for the parameterized 2-CSP problem, showing that the aforementioned conditional lower bound for sparse 2-CSPs also holds when the constraint graph is a complete graph. From this densification, we conclude that assuming ETH, there is no algorithm running in time n√k / logC k that approximates the k-Directed Steiner Network problem and the k-Strongly Connected Steiner Subgraph problem to some constant factors. Mitali Bafna, Karthik C. S., and Dor Minzer. 2025. Near Optimal Constant Inapproximability under ETH for Fundamental Problems in Parameterized Complexity. In Proceedings of the 57th Annual ACM Symposium on Theory of Computing (STOC '25). Association for Computing Machinery, New York, NY, USA, 2118–2129.

Oblivious Defense in ML Models: Backdoor Removal without Detection

2026-03-08T03:22:47Z

Oblivious Defense in ML Models: Backdoor Removal without Detection Goldwasser, Shafi; Shafer, Jonathan; Vafa, Neekon; Vaikuntanathan, Vinod As society grows more reliant on machine learning, ensuring the security of machine learning systems against sophisticated attacks becomes a pressing concern. A recent result of Goldwasser, Kim, Vaikuntanathan, and Zamir (FOCS ’22) shows that an adversary can plant undetectable backdoors in machine learning models, allowing the adversary to covertly control the model’s behavior. Backdoors can be planted in such a way that the backdoored machine learning model is computationally indistinguishable from an honest model without backdoors. In this paper, we present strategies for defending against backdoors in ML models, even if they are undetectable. The key observation is that it is sometimes possible to provably mitigate or even remove backdoors without needing to detect them, using techniques inspired by the notion of random self-reducibility. This depends on properties of the ground-truth labels (chosen by nature), and not of the proposed ML model (which may be chosen by an attacker). We give formal definitions for secure backdoor mitigation, and proceed to show two types of results. First, we show a “global mitigation” technique, which removes all backdoors from a machine learning model under the assumption that the ground-truth labels are close to a Fourier-heavy function. Second, we consider distributions where the ground-truth labels are close to a linear or polynomial function in ℝn. Here, we show “local mitigation” techniques, which remove backdoors with high probability for every input of interest, and are computationally cheaper than global mitigation. All of our constructions are black-box, so our techniques work without needing access to the model’s representation (i.e., its code or parameters). Along the way we prove a simple result for robust mean estimation. STOC ’25, Prague, Czechia

Faster Rates for No-Regret Learning in General Games via Cautious Optimism

2026-03-08T03:22:27Z

Faster Rates for No-Regret Learning in General Games via Cautious Optimism Soleymani, Ashkan; Piliouras, Georgios; Farina, Gabriele We establish the first uncoupled learning algorithm that attains O(n log2 d logT) per-player regret in multi-player general-sum games, where n is the number of players, d is the number of actions available to each player, and T is the number of repetitions of the game. Our results exponentially improve the dependence on d compared to the O(n  d logT) regret attainable by Log-Regularized Lifted Optimistic FTRL introduced by Farina, Anagnostides, Luo, Lee, Kroer, and Sandholm [2022], and also reduce the dependence on the number of iterations T from log4 T to logT compared to Optimistic Hedge, the previously well-studied algorithm with O(n logd log4 T) regret shown by Daskalakis, Fishelson, and Golowich [2021]. Our algorithm is obtained by combining the classic Optimistic Multiplicative Weights Update (OMWU) with an adaptive, non-monotonic learning rate that paces the learning process of the players, making them more cautious when their regret becomes too negative. STOC ’25, Prague, Czechia

Explicit Two-Sided Vertex Expanders beyond the Spectral Barrier

2026-03-08T03:22:27Z

Explicit Two-Sided Vertex Expanders beyond the Spectral Barrier Hsieh, Jun-Ting; Lin, Ting-Chun; Mohanty, Sidhanth; O'Donnell, Ryan; Zhang, Rachel Yun We construct the first explicit two-sided vertex expanders that bypass the spectral barrier. Previously, the strongest known explicit vertex expanders were given by d-regular Ramanujan graphs, whose spectral properties imply that every small subset of vertices S has at least 0.5d|S| distinct neighbors. However, it is possible to construct Ramanujan graphs containing a small set S with no more than 0.5d|S| neighbors. In fact, no explicit construction was known to break the 0.5 d-barrier. In this work, we give an explicit construction of an infinite family of d-regular graphs (for large enough d) where every small set expands by a factor of ≈ 0.6d. More generally, for large enough d1,d2, we give an infinite family of (d1,d2)-biregular graphs where small sets on the left expand by a factor of ≈ 0.6d1, and small sets on the right expand by a factor of ≈ 0.6d2. In fact, our construction satisfies an even stronger property: small sets on the left and right have unique-neighbor expansion 0.6d1 and 0.6d2 respectively. Our construction follows the tripartite line product framework of Hsieh et. al., and instantiates it using the face-vertex incidence of the 4-dimensional Ramanujan clique complex as its base component. As a key part of our analysis, we derive new bounds on the triangle density of small sets in the Ramanujan clique complex. STOC ’25, Prague, Czechia

All-Pairs Shortest Paths with Few Weights per Node

2026-03-08T03:22:36Z

All-Pairs Shortest Paths with Few Weights per Node Abboud, Amir; Fischer, Nick; Jin, Ce; Williams, Virginia Vassilevska; Xi, Zoe We study the central All-Pairs Shortest Paths (APSP) problem under the restriction that there are at most d distinct weights on the outgoing edges from every node. For d=n this is the classical (unrestricted) APSP problem that is hypothesized to require cubic time n3−o(1), and at the other extreme, for d=1, it is equivalent to the Node-Weighted APSP problem. We present new algorithms that achieve the following results: * Node-Weighted APSP can be solved in time Õ(n(3+ω)/2) = Õ(n2.686), improving on the 15-year-old subcubic bounds Õ(n(9+ω)/4) = Õ(n2.843) [Chan; STOC ’07] and Õ(n2.830) [Yuster; SODA ’09]. This positively resolves the question of whether Node-Weighted APSP is an ”intermediate” problem in the sense of having complexity n2.5+o(1) if ω=2, in which case it also matches an n2.5−o(1) conditional lower bound. * For up to d ≤ n3−ω−є distinct weights per node (where є > 0), the problem can be solved in subcubic time O(n3−f(є)) (where f(є) > 0). In particular, assuming that ω = 2, we can tolerate any sublinear number of distinct weights per node d ≤ n1−є, whereas previous work [Yuster; SODA ’09] could only handle d ≤ n1/2−є in subcubic time. This promotes our understanding of the APSP hypothesis showing that the hardest instances must exhaust a linear number of weights per node. With the current bounds on ω, we achieve a subcubic algorithm for d ≤ n0.628 whereas previously a subcubic running time could only be achieved for d ≤ n0.384. Our result also applies to the All-Pairs Exact Triangle problem, thus generalizing a result of Chan and Lewenstein on “Clustered 3SUM” from arrays to matrices. Notably, our technique constitutes a rare application of additive combinatorics in graph algorithms. We complement our algorithmic results with simple hardness reductions extending the n2.5−o(1) conditional lower bound for Node-Weighted APSP to undirected graphs. Interestingly, under fine-grained assumptions, the complexity in the undirected case jumps from O(nω) for d=1 to n2.5−o(1) for d ≥ 2. STOC ’25, Prague, Czechia

Weak Poincaré Inequalities, Simulated Annealing, and Sampling from Spherical Spin Glasses

2026-03-08T03:22:30Z

Weak Poincaré Inequalities, Simulated Annealing, and Sampling from Spherical Spin Glasses Huang, Brice; Mohanty, Sidhanth; Rajaraman, Amit; Wu, David X. There has been a recent surge of powerful tools to show rapid mixing of Markov chains, via functional inequalities such as Poincaré inequalities. In many situations, Markov chains fail to mix rapidly from a worst-case initialization, yet are expected to approximately sample from a random initialization. For example, this occurs if the target distribution has metastable states, small clusters accounting for a vanishing fraction of the mass that are essentially disconnected from the bulk of the measure. Under such conditions, a Poincaré inequality cannot hold, necessitating new tools to prove sampling guarantees. We develop a framework to analyze simulated annealing, based on establishing so-called weak Poincaré inequalities. These inequalities imply mixing from a suitably warm start, and simulated annealing provides a way to chain such warm starts together into a sampling algorithm. We further identify a local-to-global principle to prove weak Poincaré inequalities, mirroring the spectral independence and localization schemes frameworks for analyzing mixing times of Markov chains. As our main application, we prove that simulated annealing samples from the Gibbs measure of a spherical spin glass for inverse temperatures up to a natural threshold, matching recent algorithms based on algorithmic stochastic localization. This provides the first Markov chain sampling guarantee that holds beyond the uniqueness threshold for spherical spin glasses, where mixing from a worst-case initialization is provably slow due to the presence of metastable states. As an ingredient in our proof, we prove bounds on the operator norm of the covariance matrix of spherical spin glasses in the full replica-symmetric regime. Additionally, we resolve a question related to sampling using data-based initializations. The full version of this paper can be found on arXiv (arXiv ID: 2411.09075). STOC ’25, Prague, Czechia

Bypassing the Noisy Parity Barrier: Learning Higher-Order Markov Random Fields from Dynamics

2026-03-08T03:22:17Z

Bypassing the Noisy Parity Barrier: Learning Higher-Order Markov Random Fields from Dynamics Gaitonde, Jason; Moitra, Ankur; Mossel, Elchanan We consider the problem of learning graphical models, also known as Markov random fields (MRFs) from temporally correlated samples. As in many traditional statistical settings, fundamental results in the area all assume independent samples from the distribution. However, these samples generally will not directly correspond to more realistic observations from nature, which instead evolve according to some stochastic process. From the computational lens, even generating a single sample from the true MRF distribution is intractable unless NP=RP, and moreover, any algorithm to learn from i.i.d. samples requires prohibitive runtime due to hardness reductions to the parity with noise problem. These computational barriers for sampling and learning from the i.i.d. setting severely lessen the utility of these breakthrough results for this important task; however, dropping this assumption typically only introduces further algorithmic and statistical complexities. In this work, we surprisingly demonstrate that the direct trajectory data from a natural evolution of the MRF overcomes the fundamental computational lower bounds to efficient learning. In particular, we show that given a trajectory with Ok(n) site updates of an order k MRF from the Glauber dynamics, a well-studied, natural stochastic process on graphical models, there is an algorithm that recovers the graph and the parameters in Ok(n2) time. By contrast, all prior algorithms for learning order k MRFs inherently suffer from nΘ(k) runtime even in sparse instances due to the reductions to sparse parity with noise. Our results thus surprisingly show that this more realistic, but intuitively less tractable, model for MRFs actually leads to efficiency far beyond what is known and believed to be true in the traditional i.i.d. case. STOC ’25, Prague, Czechia

Simulating Time with Square-Root Space

2026-03-08T03:22:41Z

Simulating Time with Square-Root Space Williams, R. Ryan We show that for all functions t(n) ≥ n, every multitape Turing machine running in time t can be simulated in space only O(√t logt). This is a substantial improvement over Hopcroft, Paul, and Valiant’s simulation of time t in O(t/logt) space from 50 years ago [FOCS 1975, JACM 1977]. Among other results, our simulation implies that bounded fan-in circuits of size s can be evaluated on any input in only √s · poly(logs) space, and that there are explicit problems solvable in O(n) space which require at least n2−ε time on every multitape Turing machine for all ε > 0, thereby making a little progress on the P versus PSPACE problem. Our simulation reduces the problem of simulating time-bounded multitape Turing machines to a series of implicitly-defined Tree Evaluation instances with nice parameters, leveraging the remarkable space-efficient algorithm for Tree Evaluation recently found by Cook and Mertz [STOC 2024]. STOC ’25, Prague, Czechia

Model Stealing for Any Low-Rank Language Model

2026-03-08T03:39:26Z

Model Stealing for Any Low-Rank Language Model Liu, Allen; Moitra, Ankur Model stealing, where a learner tries to recover an unknown model via carefully chosen queries, is a critical problem in machine learning, as it threatens the security of proprietary models and the privacy of data they are trained on. In recent years, there has been particular interest in stealing large language models (LLMs). In this paper, we aim to build a theoretical understanding of stealing language models by studying a simple and mathematically tractable setting. We study model stealing for Hidden Markov Models (HMMs), and more generally low-rank language models. We assume that the learner works in the conditional query model, introduced by Kakade, Krishnamurthy, Mahajan and Zhang. Our main result is an efficient algorithm in the conditional query model, for learning any low-rank distribution. In other words, our algorithm succeeds at stealing any language model whose output distribution is low-rank. This improves upon the previous result which also requires the unknown distribution to have high “fidelity” – a property that holds only in restricted cases. There are two key insights behind our algorithm: First, we represent the conditional distributions at each timestep by constructing barycentric spanners among a collection of vectors of exponentially large dimension. Second, for sampling from our representation, we iteratively solve a sequence of convex optimization problems that involve projection in relative entropy to prevent compounding of errors over the length of the sequence. This is an interesting example where, at least theoretically, allowing a machine learning model to solve more complex problems at inference time can lead to drastic improvements in its performance. STOC ’25, Prague, Czechia

Maximum Circuit Lower Bounds for Exponential-Time Arthur Merlin

2026-03-08T03:22:32Z

Maximum Circuit Lower Bounds for Exponential-Time Arthur Merlin Chen, Lijie; Li, Jiatu; Liang, Jingxun We show that the complexity class of exponential-time Arthur Merlin with sub-exponential advice (AMEXP/2nε) requires circuit complexity at least 2n/n. Previously, the best known such near-maximum lower bounds were for symmetric exponential time by Chen, Hirahara, and Ren (STOC’24) and Li (STOC’24), or randomized exponential time with MCSP oracle and sub-exponential advice by Hirahara, Lu, and Ren (CCC’23). Our result is proved by combining the recent iterative win-win paradigm of Chen, Lu, Oliveira, Ren, and Santhanam (FOCS’23) together with the uniform hardness-vs-randomness connection for Arthur-Merlin protocols by Shaltiel-Umans (STOC’07) and van Melkebeek-Sdroievski (CCC’23). We also provide a conceptually different proof using a novel ”critical win-win” argument that extends a technique of Lu, Oliveira, and Santhanam (STOC’21). Indeed, our circuit lower bound is a corollary of a new explicit construction for properties in coAM. We show that for every dense property P ∈ coAM, there is a quasi-polynomial-time Arthur-Merlin protocol with short advice such that the following holds for infinitely many n: There exists a canonical string wn ∈ P ∩ {0,1}n so that (1) there is a strategy of Merlin such that Arthur outputs wn with probability 1 and (2) for any strategy of Merlin, with probability 2/3, Arthur outputs either wn or a failure symbol ⊥. As a direct consequence of this new explicit construction, our circuit lower bound also generalizes to circuits with an AM ∩ coAM oracle. To our knowledge, this is the first unconditional lower bound against a strong non-uniform class using a hard language that is only ”quantitatively harder”. STOC ’25, Prague, Czechia

List-Decoding Capacity Implies Capacity on the 𝑞-ary Symmetric Channel

2026-03-08T03:22:33Z

List-Decoding Capacity Implies Capacity on the 𝑞-ary Symmetric Channel Pernice, Francisco; Sprumont, Oscar; Wootters, Mary It is known that the Shannon capacity of the q-ary symmetric channel (qSC) is the same as the list-decoding capacity of an adversarial channel, raising the question of whether there is a formal (and black-box) connection between the two. We show that there is: Any linear code C⊆ Fqn that has superconstant minimum distance and achieves list-decoding capacity also achieves capacity on the qSC. STOC ’25, Prague, Czechia

A molecularly impermeable polymer from two-dimensional polyaramids

2026-03-08T03:39:23Z

A molecularly impermeable polymer from two-dimensional polyaramids Ritt, Cody L; Quien, Michelle; Wei, Zitang; Gress, Hagen; Dronadula, Mohan T; Altmisdort, Kaan; Nguyen, Huong Giang T; Zangmeister, Christopher D; Tu, Yu-Ming; Garimella, Sanjay S; Amirabadi, Shahab; Gadaloff, Michael; Hu, Weiguo; Aluru, Narayana R; Ekinci, Kamil L; Bunch, J Scott; Strano, Michael S All polymers exhibit gas permeability through the free volume of entangled polymer chains1, 2–3. By contrast, two-dimensional (2D) materials including graphene stack densely and can exhibit molecular impermeability4, 5–6. Solution-synthesized 2D polymers that exhibit the latter by poly-condensation have been a longstanding goal. Herein, we demonstrate self-supporting, spin-coated 2D polyaramid nanofilms that exhibit nitrogen permeability below 3.1 × 10−9 Barrer, nearly four orders of magnitude lower than every class of existing polymer, and similar for other gases tested (helium, argon, oxygen, methane and sulfur hexafluoride). Optical interference during the pressurization of nanofilm-coated microwells allows measurement of mechanosensitive rim opening and sealing, creating gas-filled bulges that are stable exceeding three years. This discovery enables 2D polymer resonators with high resonance frequencies (about 8 MHz) and quality factors up to 537, similar to graphene. A 60-nm coating of air-sensitive perovskites reduces the lattice degradation rate 14-fold with an oxygen permeability of 3.3 × 10−8 Barrer. Molecularly impermeable polymers promise the next generation of barriers that are synthetically processable, chemically amenable and maximize molecular rejection with minimal material, ultimately advancing sustainability goals.

Understanding and Overcoming Optimization Barriers in Non-convex and Non-smooth Machine Learning

2026-01-21T03:25:03Z

Understanding and Overcoming Optimization Barriers in Non-convex and Non-smooth Machine Learning Gatmiry, Khashayar At their core, our machine learning systems are trained by solving an optimization problem, where the goal is to minimize a predefined objective function by adjusting model parameters based on the data. Despite the wealth of structure and prior knowledge present in the data and feedback, our training methods remain relatively simple and independent of this structure. In spite of, or perhaps because of, this simplicity, these methods are often lacking in theoretical guarantees. To design machine learning algorithms that are less data-hungry while ensuring theoretical guarantees on both computational efficiency and output validity, it is essential to better understand and leverage the rich structure within the learning setup and the data distribution, e.g. by altering the geometry of the solution space or adjusting the objective function to induce a more effective learning procedure. This approach moves beyond classical algorithm design, which focuses primarily on handling worst-case instances. This thesis investigates the optimization landscape of central learning problems and develops geometric and analytic schemes adapted to their structure, leading to algorithms with superior computational and statistical performance. In addition, it seeks to advance our mathematical understanding of the principles underlying the success of deep learning.

Seeing the Forest Through the Trees: Knowledge Retrieval for Streamlining Particle Physics Analysis

2026-01-21T04:07:55Z

Seeing the Forest Through the Trees: Knowledge Retrieval for Streamlining Particle Physics Analysis McGreivy, James C. Generative Large Language Models (LLMs) are a promising approach to structuring knowledge contained within otherwise unmanageable corpora of research literature produced by large-scale and long-running scientific collaborations. Within experimental particle physics, such structured knowledge bases could expedite methodological and editorial review. Complementarily, within the broader scientific community, generative LLM systems grounded in published work could make for reliable companions allowing non-experts to analyze openaccess data. Techniques such as Retrieval Augmented Generation (RAG) rely on semantically matching localized text chunks, but struggle to maintain coherent context when relevant information spans multiple segments, leading to a fragmented representation devoid of global cross-document information. In this work I utilize the hierarchical organization of experimental physics articles to build a tree representation of the corpus, and present the SciTreeRAG system which leverages this structure with the aim of constructing contexts more focused and contextually rich than a standard RAG. Additionally, I develop methods for using LLMs to transform the unstructured corpus into a structured knowledge graph representation. I then implement SciGraphRAG, a retrieval system that leverages this knowledge graph to access global cross-document relationships eluding standard RAG, with the goal of encapsulating domain-specific connections and expertise. I demonstrate proof-of-concept implementations of both systems using the corpus of the LHCb experiment at CERN.

Development of Low-Cost In Situ Gas Sensors for Oceanographic Applications

2026-01-21T04:07:47Z

Development of Low-Cost In Situ Gas Sensors for Oceanographic Applications Gower, Elizabeth Ann Anthropogenic activity has increased atmospheric carbon dioxide (CO₂) levels, disrupting the global carbon cycle and driving widespread environmental change. The ocean acts as a major sink. Accurate and scalable in situ monitoring of oceanic carbon chemistry is vital for understanding the impacts of climate change and informing marine carbon dioxide removal (mCDR) strategies. Many existing in situ instruments for marine applications are constrained by their size, cost, power requirements, or reliance on consumable reagents. Developing low-cost, compact, low-power, and accurate in situ sensors would significantly enhance the spatiotemporal resolution of oceanographic data and enable widespread monitoring of dissolved gases throughout the ocean. This, in turn, would deepen our understanding of how, where, and when changes are occurring within the marine carbon cycle. Two key variables essential for studying this cycle are the partial pressure of carbon dioxide (pCO₂) and dissolved inorganic carbon (DIC). This thesis presents the development of two sensors, one for in situ pCO₂ measurement and another for novel DIC quantification, both designed to be affordable, reliable, and scalable tools for advancing our understanding of ocean chemistry and the global carbon system. First, the development, calibration, and open-ocean deployment of a miniaturized Dissolved Multi-Gas Sensor (DMGS) that measures pCO₂ and partial pressure of oxygen (pO₂) is presented. The sensor was integrated into a custom-built surface drifter designed to entangle with Sargassum mats and send data autonomously. The drifter utilized commercial off-theshelf (COTS) components and cost roughly $1000 to build. After lab testing, a drifter was deployed in the Great Atlantic Sargassum Belt (GASB) and collected data for 22-days. In addition to gas data, the drifter tracked temperature, light intensity, humidity, pressure, and location sending measurements via an Iridium satellite. The resulting data captured dynamic changes in localized gas concentrations, temperature, and light levels that highlighted photosynthetic and respiratory activity within Sargassum patches. These drifters demonstrate the value of in situ data to investigate marine biogeochemical processes that contribute to the marine carbon cycle, especially in areas with high biologic activity. Next, this thesis presents the iterative development of a novel DIC sensor with potential for future in situ applications. Initial prototypes tested the feasibility of using a COTS CO2 sensor in both static and flow-through configurations, however sensor saturation issues prompted a shift to a pressure-based detection method. Multiple test setups were evaluated for pressure stability and sensor sensitivity, culminating in a bottle-based flow system that demonstrated the potential for reagent-minimized, pressure-based DIC quantification. With the final setup, a COTS pressure sensor that sat behind a gas permeable membrane was found to repeatably and accurately quantify DIC from acidified seawater. This approach of quantifying DIC via pressure change is novel in the field of gas sensing and maintains a low-cost, accessible design. Together, the sensors developed in this thesis expand the toolkit for marine carbon monitoring and provide a foundation for affordable, distributed sensing networks. These technologies enable higher-resolution insights into ocean biogeochemistry and support critical monitoring, reporting, and verification (MRV) frameworks needed to evaluate the effectiveness of mCDR techniques. Continued refinement of these low-cost platforms could play a key role in understanding and mitigating anthropogenic impacts on marine systems.

Applying Reference Class Forecasting to Multifamily Investments: Identifying and Capturing Operational Alpha through the Outside View

2026-01-21T04:07:55Z

Applying Reference Class Forecasting to Multifamily Investments: Identifying and Capturing Operational Alpha through the Outside View Firouzian, Fardean This thesis applies Reference Class Forecasting (RCF) to multifamily real estate underwriting as a means of countering optimism bias, strategic misrepresentation, and other distortions embedded in the traditional “inside view.” Adapted from its proven application in infrastructure and corporate capital budgeting, RCF anchors projections in the actual performance distributions of comparable assets rather than in deal-specific narratives. The research centers on the development of the “Comp Warehouse,” a structured repository of property-level financials organized by market, asset class, vintage, and unit scale. By benchmarking assumptions against statistically valid reference classes, the approach enforces empirical discipline and highlights opportunities for “operational alpha”—the marginal increase in net operating income (NOI) achieved when underperforming assets converge on median peer performance. A South Florida case study demonstrates the method’s utility in an acquisition context. Analysis of 48 assets across Melbourne, Miami, Fort Lauderdale, and West Palm Beach shows that while rent levels cluster tightly around market medians, operating expenses vary widely, producing large dispersion in realized NOI. Applying the framework to a 191-unit Class A property in Fort Lauderdale illustrates how RCF can ground underwriting assumptions by distinguishing between defensible revenue-driven growth strategies and less plausible expense-reduction projections proposed in a bidding scenario. Recognizing constraints of both scale and frequency, this thesis also explores artificial intelligence as a tool for automating the ingestion and standardization of operating statements and rent rolls. Properly deployed in a human-in-the-loop framework, AI can reduce data friction, expand sample sizes, and sharpen forecasting precision. The contribution of this thesis is twofold: it demonstrates the feasibility of applying RCF to the multifamily sector—an asset class whose relative standardization, liquidity, and data availability make it especially conducive to outside-view benchmarking—and it situates the methodology within a technology-native architecture designed to scale empirical discipline, enhance underwriting rigor, and systematically capture operational alpha.

Concretely-Efficient Multi-Key Homomorphic Secret Sharing and Applications

2026-01-21T04:07:53Z

Concretely-Efficient Multi-Key Homomorphic Secret Sharing and Applications He, Kaiwen Homomorphic secret sharing (HSS) is a powerful cryptographic primitive that enables efficient, low-communication secure computation without the use of fully homomorphic encryption. Public-key HSS is a well-known variant that supports inputs from multiple parties, but all parties must agree on a joint public key before any party can encode their inputs, requiring extra rounds of communication in applications. Recently, Couteau et al. (EUROCRYPT 2025) constructed multi-key HSS (MKHSS)—a new primitive which allows parties to encode their inputs under independent keys—under the DCR assumption. MKHSS assumes only a reusable common reference string, without the need for prior interactions between parties or a public-key infrastructure. In this paper, we construct and implement the first concretely-efficient MKHSS scheme under the same assumptions used by Couteau et al. Using an algorithmic insight that reduces the largest modulus in Couteau et al. from N⁴ to N², our optimized implementation can homomorphically multiply inputs in 5.0 milliseconds—while an implementation of Couteau et al. requires 224.6 milliseconds—thereby achieving a 45× speedup. A powerful application of MKHSS is to realize attribute-based non-interactive key exchange (ANIKE), which generalizes password-based key exchange (PAKE) to arbitrary attribute policies. ANIKE is currently only known from MKHSS. We use our implementation to evaluate the first concretely-efficient ANIKE schemes for a range of practically useful policies. Using our implementation, two parties can perform a geolocation-based key exchange in 1.65 seconds and a fuzzy PAKE on an 8-word passphrase in 7.59 seconds for realistic parameters, on a single core. Compared to using Couteau et al., which requires 62.5 and 253 seconds, we achieve 38× and 33× speedups, respectively.

Reciprocity and Normality in the Scattering Matrix of Disordered Media

2026-01-21T04:07:46Z

Reciprocity and Normality in the Scattering Matrix of Disordered Media Bharadwaj, Shreyas K. The scattering matrix formalism provides a practical characterization of wave transport in linear, source-free systems by relating a set of operationally defined input and output spatial channels. The matrix is structured as a block operator, with diagonal blocks encoding same-side reflection matrices (RMs) and off-diagonal blocks encoding transmission matrices (TMs) in opposing propagation directions. Under Helmholtz reciprocity, symmetry relations are imposed: RMs are symmetric, and forward and reverse TMs are mathematical transposes of each other. These relations were employed as constraints to correct system-induced aberrations in measured scattering matrices of complex optical media via a matrix-based gradient descent procedure. Resulting phase corrections corresponded closely with classical aberration modes without heuristic parameterizations, suggesting that these modes naturally arise to restore reciprocity-induced symmetry. Vectorial TMs were measured for single- and double-pass propagation through step-index MMFs and scattering samples, with corrected phase terms showing agreement across sample types. Furthermore, matrix normality was introduced as a descriptor of stable modal transport. Normal matrices admit unitary diagonalization, reflecting orthogonal eigenchannels and spectrally coherent propagation. Near-normal behavior was observed in fiber TMs, while RMs of scattering slabs remained strongly non-normal, as quantified by a normalized Henrici departure. Sufficient conditions for normality were identified in terms of the system Green’s function and its bi-compression onto the measurement basis. A complementary dispersion experiment investigated two regimes: nearly-normal MMFs, where the Wigner–Smith time-delay operator was jointly diagonalizable and supported accurate first-order spectral models; and mechanically compressed fibers, where loss of normality produced noncommuting operators and collapse of model fidelity. These results suggest that normality captures well-behaved modal transport, underpinning the validity of parametric models and other operator-based analyses of disordered media. Together, reciprocity and normality impose complementary constraints on wave transport: reciprocity governs global symmetry, while normality captures internal coherence of modal propagation. Relevance is noted for matrix-based imaging, inverse scattering theory, and non-Hermitian wave physics, where symmetry and modal stability remain central.

Mesh Differentiable Rendering for Real-World Scenes

2026-01-21T04:07:54Z

Mesh Differentiable Rendering for Real-World Scenes Charatan, David Differentiable rendering has established itself as an effective tool for 3D reconstruction and novel view synthesis. Most state-of-the-art differentiable rendering methods use purpose-built renderers to optimize specialized, nonstandard 3D representations. However, most downstream applications of differentiable rendering rely on 3D meshes, which are near-universally supported due to their suitability for a wide range of rendering, simulation, and 3D modeling workflows. While prior methods have explored using 3D meshes directly within gradient-based optimization, they have been limited to object-centric scenes and cannot reconstruct real-world, unbounded scenes. This work addresses this shortcoming via a differentiable rendering formulation that combines an off-the-shelf, non-differentiable triangle rasterizer with a 3D representation that consists of nested mesh shells. During every forward pass, these shells are extracted from an underlying signed distance field. Then, the shells are independently rasterized and the resulting images are alpha-composited using opacities derived from the shells' per-vertex signed distance values. Notably, the shells' vertex positions are updated only via the underlying signed distance field, not via backpropagation through the rasterizer itself. This makes our method compatible with off-the-shelf, non-differentiable triangle rasterizers. To the best of our knowledge, our method is the first differentiable mesh rendering method that scales to unbounded, real-world 3D scenes, where it produces high-quality novel view synthesis results whose quality approaches the quality of state-of-the-art, non-mesh-based methods. Our method's performance is also competitive with state-of-the-art surface rendering methods on object-centric scenes. Ultimately, our method suggests that it may be possible to solve the differentiable rendering problem using tools from the conventional graphics toolbox rather than relying on specialized renderers.

Task-Aware Spatial and Temporal Aggregation for Capacity Expansion Planning

2026-01-21T04:07:49Z

Task-Aware Spatial and Temporal Aggregation for Capacity Expansion Planning Duguey, Gabriel As we plan tomorrow’s electricity system, we face fundamental questions: where should new power plants go, which technologies deserve investment, and how much transmission is enough? These decisions are the domain of Capacity Expansion Planning (CEP), a class of optimization models that guide long-term infrastructure investments in power systems. To be realistic, CEP models must capture fine-grained spatial and temporal variations because demand varies by city and climate, while wind and solar output depend on weather patterns that shift hour by hour and location by location. But representing the system with thousands of time steps and hundreds of nodes makes the optimization problem computationally too large to solve. This thesis addresses the core question: how can spatial and temporal aggregation in CEP models be designed to preserve planning-relevant patterns that drive investment decisions? Existing approaches often treat aggregation as a neutral preprocessing step, relying on heuristics like political boundaries or geographic proximity. In contrast, we propose a task-aware pipeline that treats aggregation as an integral modeling decision, explicitly aligned with planning objectives. The approach builds a composite similarity metric that blends diverse planning-relevant signals, including, but not limited to, duration curves, ramping behavior, and spatial correlation, and uses k-medoids clustering to define spatial zones. Temporal aggregation is then applied to daily system-wide profiles, selecting representative days that maintain cross-zonal interactions. The result is a reduced spatio-temporal dataset fed into a CEP model. The resulting investment decisions are re-evaluated at full resolution to evaluate their feasibility and real cost. Experiments on a New England case study show the pipeline consistently outperforms common baselines like political boundaries, geographic proximity, or capacity factor statistics. Among 50 feature weightings, the best design reduces system cost by 13% compared to heuristics. Correlation-based features drive the best results, while raw amplitude and geographic location often degrade performance when used alone.

Earth as Equity Partner: A Revolutionary Approach to Ecological Conservation through Housing Development

2026-01-21T04:07:45Z

Earth as Equity Partner: A Revolutionary Approach to Ecological Conservation through Housing Development McDonough, Kate Duddington Farm is a 312-acre site north of Baltimore, Maryland. A stream restoration project was completed at the location nearly a decade ago in concert with the State of Maryland, the Manor Conservancy, Ecotone, and landowners Harry and Tara McDonough. The project was conducted with some success, however due to a lack of State oversight and long-term management provisions, the ecology has since declined. The following proposal outlines a new model for long-term land restoration and conservation, whereby land conservation and restoration are financed not solely through short term grants and fragile easements, but through the thoughtful use of modest real estate interventions. A small cluster of homes is developed on one portion of the site. The act increases the value of the land, generates equity, and establishes a permanent conservation fund. The design protects habitat and invites people into a deeper relationship with the natural world. The plan offers scalability in taking the land value capture and applying it to future land conservation projects, compounding returns and projecting a model to preserve hundreds of thousands of acres of critical land across the United States. This model highlights Indigenous ecological knowledge (TEK) and traditional practices of engaging with the land, highlighting a deeper understanding of how humans and nature can coexist in mutually healthy ways. The model is designed at a time when watersheds, national parks, and old-growth forests are faced with the greatest threat to global ecology. Duddington Farm is used as a retrospective case, but the broader goal is to create a regenerative framework for conservation-based development across critical watershed regions.

The Causal Effects of Mandatory Quarterly Earnings Guidance on Corporate Information Environment and Corporate Short-Termism

2026-01-21T03:24:57Z

The Causal Effects of Mandatory Quarterly Earnings Guidance on Corporate Information Environment and Corporate Short-Termism Wang, Yuting I examine the causal effects of mandatory quarterly earnings guidance using a regulatory mandate in China that required a subset of listed firms to issue bundled quarterly earnings guidance from 2007 to 2018. A difference-in-differences analysis shows that when these firms are no longer required to issue such guidance, their corporate information environment deteriorates, evidenced by reduced analyst coverage, fewer site visits, and lower price timeliness, meaning that stock prices incorporate less information about current and future earnings. However, these firms increase R&D and SG&A spending, consistent with alleviated managerial myopia as short-term market pressure eases. These findings highlight the dual-edged nature of the mandatory quarterly earnings guidance and offer insights for both practitioners and policymakers.

Automated Finite Elements

2026-01-21T04:07:51Z

Automated Finite Elements Collin, Teodoro Fields Finite element methods (FEMs) are a powerful and ubiquitous tool for solving engineering problems. Experimenting with different finite elements can improve the quality and efficiency of solutions. Furthermore, in some cases, the wrong (but nonetheless most common) choice of finite element will produce solutions which converge to the wrong answer regardless of mesh resolution. However, in practice, the choice of finite element is not explored due to the complexity of re-deriving and re-implementing finite element methods. Trying a new finite element is challenging because practitioners must manually deduce formulas to use these elements and they must implement these formulas within the context of a potentially complex system. We address this problem by introducing ElementForge, a finite element system that is parametric over the literate mathematical specification of a finite element in a domain-specific language (DSL). The ElementForge compiler reasons about tensor spaces, tensors, and tensor bases from first principles to derive implementations of finite elements. The ElementForge compiler is able to automatically derive implementations of finite elements previously only derived by hand. Further, ElementForge minimally couples several key mathematical concepts, mainly tensor fields, mesh topologies, sparse tensors, and assembled finite element operators, to produce a complete finite element system that is parametric over the choice of element. Consequently, the elements derived by the compiler can be applied parametrically to new meshes, PDEs, and boundary conditions. We evaluate our system by implementing several simulations with different finite elements, demonstrating that our system can explore tradeoffs in generality, accuracy, speed, and representational complexity. For example, we are able to implement the Morley, Bell, Argyris, and Hermite like elements with less than 50 lines of code and use them all in a single simulation.

Bespoke Threat Models: Achieving Realistic Privacy Guarantees for Deployed Protocols

2026-01-21T03:24:48Z

Bespoke Threat Models: Achieving Realistic Privacy Guarantees for Deployed Protocols Hogan, Kyle This thesis focuses on the question of what degree of privacy is achievable in the real world for long-running applications. We explore this question in two main settings: private advertising and anonymous communication. In doing so, we consider constraints each application may have in practice and what adversarial model is realistic for the context in which the application will be deployed. For real world applications, achieving perfect privacy — especially against a worst case adversary — can be impossible. That is, perfect privacy, while achievable in theory, may in practice require assumptions that conflict with usability, deployability, or utility requirements. This presents a challenge as privacy-preserving technologies can, necessarily, only provide privacy for the people who use them. Because of this, designing around user experience is critical, even if doing so requires compromises in the theoretical degree of privacy a system can provide or the strength of adversaries considered in its threat model. In the space of private advertising, we first propose a novel protocol, AdVeil, that eliminates leakage of user data beyond that revealed by the input/output of the ads ecosystem as a whole. We then provide a minimal modeling of the functionality of digital advertising which we use to prove that, even for systems like AdVeil with minimal leakage, the advertising metrics released at the end of the protocol are sufficient to leak information about end users to advertisers when combined with their audience targeting criteria. In the space of anonymous communication, we propose ShorTor, a new routing protocol for Tor that utilizes techniques popular with content distribution networks (CDNs) to reduce latency while maintaining Tor’s existing anonymity guarantees. We evaluate this protocol using a dataset of over 400,000 latency measurements we collected between the 1,000 most popular Tor relays.

Operationalizing Reliable Machine Learning: From Data Collection to Model Presentation

2026-01-21T03:25:00Z

Operationalizing Reliable Machine Learning: From Data Collection to Model Presentation Balagopalan, Aparna Automated systems driven by machine learning (ML) have made exciting progress across a spectrum of applications. Despite such progress, encoded biases and other failure modes may create barriers to the real-world utility and reliability of such systems. For example, nonrandom data missingness, biased algorithmic optimization objectives, or model presentation strategies that incorrectly impact user trust can all cause models to fail in practice. In this thesis, guided by such observations and prior work on pipeline-awareness in machine learning, we aim to operationalize reliable ML. Under this goal, we propose a framework consisting of the following three components: responsible data collection, robust algorithm development, and fair model presentation. We first conduct two case studies to advance responsible data collection. We investigate whether standard procedures for acquiring data can be repurposed when training models to mimic human judgments about norm violations. We also demonstrate patterns of delayed demographic data reporting within a longitudinal healthcare dataset and show that timevarying missingness due to such delays can distort disparity assessments. Second, we introduce two novel algorithms to improve reliability: a method that leverages representations from vision-language models to filter noisy training data, and a method to produce fair rankings that account for properties of search queries. Finally, since the presentation design of predictions impacts trust in model consumers, we propose metrics to quantify the fairness of post-hoc explainability techniques. Thus, with this thesis, we re-evaluate measurements throughout the machine learning pipeline and contribute to the broader goal of reliable machine learning.

Anti-phage defense as a driver of molecular innovation

2026-01-21T03:24:56Z

Anti-phage defense as a driver of molecular innovation Doering, Christopher Ross Bacteriophages, or phages for short, pose a near-constant threat to the bacteria they infect. Billions of years of conflict has been a catalyzing force for the creation of bacterial defense systems and corresponding phage evasion strategies. To counter phage predation, bacteria have developed a vast diversity of enzyme chemistries and molecular sensing mechanisms whose study has produced new biotechnological tools and insights into our own immune systems. In this work, I have investigated anti-phage defense mechanisms at multiple scales using a combination of genetic, biochemical, and bioinformatic approaches. First, I characterized the mechanism of action of the anti-phage defense system CmdTAC, a toxin-antitoxin-chaperone system that recognizes a viral structural protein to activate a novel mRNA ADP-ribosyltransferase, thereby halting infection. Next, I examined the diversity and distribution of anti-phage mechanisms encoded by E. coli lysogenic phages – phages capable of integrating into and lying dormant within their bacterial hosts. This analysis uncovered overlooked classes of lysogenic phages harboring novel candidate defense systems, including one newly validated system with no detectable homology to previously known mechanisms. Together, this work broadens our understanding of bacterial immune systems, expands the pool of known enzyme chemistries, and highlights areas where continued study can reveal additional mechanisms of anti-phage defense.

Shaping Function Through Space: The Role of Spatial Organization in Microbial Communities

2026-01-21T03:24:46Z

Shaping Function Through Space: The Role of Spatial Organization in Microbial Communities Toneatti Vercelli, Gabriel Spatial organization plays a critical role in microbial community function, influencing how cells exchange metabolites, coordinate behavior, and compete for resources. This thesis investigates the consequences of spatial structure in natural microbial systems and introduces a novel method to engineer these systems with high precision and scalability. First, we examine the colonization of chitin particles by marine bacteria, a model for particulate organic matter degradation. Using high-throughput phenotyping of natural isolates, we show that vitamin cross-feeding is essential for successful colonization of chitin-particles by many auxotrophic strains. We then model two distinct vitamin cross-feeding mechanisms: lysis and secretion. Using a resource-explicit modeling approach, we leverage metabolic-flux and physiological measurements to predict the colonization success of auxotrophic cross-feeders in this spatially structured environment. Second, we introduce a new chemical method for engineering microbial cell surfaces that enables covalent attachment of molecules such as enzymes and DNA strands to the cell surface. We show that this surface functionalization procedure leads to the acquisition of new phenotypes like antibiotic resistance and programmable adhesion. Altogether, this work reinforces the importance of spatial organization for microbial community function and introduces a new technique to harness this community feature and turn it into a design principle for synthetic microbial systems.

LumiModeling: A Gaussian Splatting-Based Tool for Recreating Dynamic Material and Lighting Interaction in Architecture

2026-01-21T04:07:43Z

LumiModeling: A Gaussian Splatting-Based Tool for Recreating Dynamic Material and Lighting Interaction in Architecture Cao, Biru This thesis presents LumiModeling, a real-time visualization tool based on Gaussian Splatting (GS) that simulates the dynamic interplay between materiality and lighting in architectural environments. While conventional design workflows rely on geometric modeling and photorealistic rendering, they often abstract complex material behaviors and fall short in capturing light-material interactions. In contrast, GS enables the reconstruction of high-fidelity 3D models from 2D image sets, representing viewdependent effects such as reflection, transparency, and surface roughness. A comparative analysis using real-world data from the MIT Stata Center and the Met Warehouse demonstrates GS’s advantages over mesh-based photogrammetry, particularly in rendering reflective and transparent materials. This work extends existing GS capabilities by implementing a relightable pipeline based on the existing model Relightable3DGaussian (Gao et al., 2023), in which each Gaussian point is augmented with physical parameters, including BRDF, surface normals, and incident lighting. The Stata Center dataset is used to test the relighting of GS. A user study involving architecture professionals reveals that perceptual focus shifts from geometry to materiality and lighting as visual realism increases. The findings highlight the potential of relightable GS in architectural visualization and anticipate its integration into future design workflows.

Urban Data Memory: Using Generative AI to Structure and Visualize Zoning Data for Urban Planning Evaluation

2026-01-21T04:07:52Z

Urban Data Memory: Using Generative AI to Structure and Visualize Zoning Data for Urban Planning Evaluation Kupershmidt, Adi Urban planners face significant challenges in systematically and quantitatively evaluating past planning practices, stemming, among other reasons, from the scarcity of accessible structured data. The period from a plan’s initiation to implementation can span generations; recorded data from the planning processes are often deemed obsolete for addressing present concerns by the time of post-occupancy evaluation. This research examines whether generative AI can help bridge this gap and under what conditions - highlighting both challenges and opportunities - by introducing a system that responsively transforms qualitative zoning data into structured, queryable formats to support the quantitative analysis of planning practices. A database of ~150 approved semi-structured urban plans under Tel Aviv municipality’s local jurisdiction supports this project's case study. The system relies on proprietary LLMs (ChatGPT, Claude), streamlining a natural language query input through 3 agentic tasks: (1) RAG (Retrieval Augmented Generation) based querying, generating free-text answers from all plans, (2) structuring the answers to a valid JSON, and (3) visualizing structured data. Key findings indicate an 85.45% precision of the system, as evaluated through an end-to-end assessment of 11 representative queries, each validated against 40 manually labeled plans. The tool provides actionable insights, enabling queries such as trends in sheltered bicycle parking approvals or the status of affordable housing planning over the past decade. This research underlines the significance of flexibly structuring non- and semi-structured data for urban science. It addresses the growing gap between static legacy data collection and real-time policymaking, democratizing access to planning information and fostering informed decision-making practices. Integrating cutting-edge AI-driven tools contributes to the current discourse on AI applications for city management and planning by providing a replicable model for more cities and planning datasets to build upon and improve.

Developing Methods for Enhanced Measurement of DNA Single-Strand Breaks and Somatic Variants

2026-01-21T03:24:52Z

Developing Methods for Enhanced Measurement of DNA Single-Strand Breaks and Somatic Variants Elacqua, Juniper J. Maintenance and repair of DNA are essential for proper cellular functioning and preventing the emergence of disease states. As cells divide, mutations accumulate in the genome which contributes to aging phenotypes and can result in genetic diseases such as cancer. The rate at which a cell develops mutations can be accelerated through exposure to genotoxic agents that introduce lesions which, if left unrepaired, prevent accurate replication of the genome. As such, it is crucial to understand the ways in which DNA becomes damaged, how cells respond to various types of damage, and how this damage contributes to mutagenesis and the development of genetic disease. These fields of study have been greatly advanced by improvements in DNA sequencing technologies, and here we present two sequencing-based methods that aim to enable deeper study of DNA damage, repair, and mutagenesis. First, we demonstrate DENT-seq, a method that identifies single-strand breaks with single-nucleotide resolution. Single-strand breaks are the most common form of DNA damage, occurring at rates of ~10,000 per cell per day, but have to date been understudied due to lack of an unbiased, high-resolution method for their detection. Second, we improve upon lineage sequencing, a previously reported method that uniquely measures somatic single nucleotide variants in dividing cells to achieve high specificity/sensitivity as well as the ability to temporally resolve variants and to relate sequenced genotypes to optically observed cellular phenotypes. Despite the high-quality data and unique capabilities offered by this method, it has so far been underused due to a need for complex, microfluidic-based cell collection. We demonstrate novel protocols for performing lineage sequencing that enable easy adoption of the method without the need for highly specialized equipment or expertise. In addition, we expand the repertoire of mutations measurable with the technique to include indels and variants that arise specifically in response to a genotoxic treatment. The methods we show can be applied to reveal novel findings regarding the causes and consequences of DNA damage and mutagenesis that underly numerous genetic diseases.Maintenance and repair of DNA are essential for proper cellular functioning and preventing the emergence of disease states. As cells divide, mutations accumulate in the genome which contributes to aging phenotypes and can result in genetic diseases such as cancer. The rate at which a cell develops mutations can be accelerated through exposure to genotoxic agents that introduce lesions which, if left unrepaired, prevent accurate replication of the genome. As such, it is crucial to understand the ways in which DNA becomes damaged, how cells respond to various types of damage, and how this damage contributes to mutagenesis and the development of genetic disease. These fields of study have been greatly advanced by improvements in DNA sequencing technologies, and here we present two sequencing-based methods that aim to enable deeper study of DNA damage, repair, and mutagenesis. First, we demonstrate DENT-seq, a method that identifies single-strand breaks with single-nucleotide resolution. Single-strand breaks are the most common form of DNA damage, occurring at rates of ~10,000 per cell per day, but have to date been understudied due to lack of an unbiased, high-resolution method for their detection. Second, we improve upon lineage sequencing, a previously reported method that uniquely measures somatic single nucleotide variants in dividing cells to achieve high specificity/sensitivity as well as the ability to temporally resolve variants and to relate sequenced genotypes to optically observed cellular phenotypes. Despite the high-quality data and unique capabilities offered by this method, it has so far been underused due to a need for complex, microfluidic-based cell collection. We demonstrate novel protocols for performing lineage sequencing that enable easy adoption of the method without the need for highly specialized equipment or expertise. In addition, we expand the repertoire of mutations measurable with the technique to include indels and variants that arise specifically in response to a genotoxic treatment. The methods we show can be applied to reveal novel findings regarding the causes and consequences of DNA damage and mutagenesis that underly numerous genetic diseases.

Reconfigurable and Interference-Tolerant Receivers for Next Generation Wireless Systems

2026-01-21T03:24:43Z

Reconfigurable and Interference-Tolerant Receivers for Next Generation Wireless Systems Araei, Soroush An “all-in-one” radio, programmable across the sub-7 GHz spectrum, offers significant hardware efficiency for 5G systems. However, addressing strong interferers in this wide and congested spectrum remains a major design challenge. N-path filters offer a promising solution for efficiently suppressing interference, thanks to their clock-controlled reconfigurability and excellent linearity against in-band and adjacent-channel blockers. While widely adopted in modern receiver architectures, these switched-capacitor circuits remain inherently vulnerable to blockers at clock harmonics, due to their hard-switching nature. These blockers, common in 5G bands, pose a key bottleneck, delaying the realization of fully integrated multi-band, multi-mode radios. This dissertation introduces fully passive topologies to address this challenge. The first design leverages simultaneous charge sharing and capacitor stacking to implement harmonic rejection filtering. It operates entirely without active circuitry and exhibits exceptionally low loss. A second-generation technique, termed “harmonic reset switching”, builds on this approach by rejecting harmonic blockers directly at the driving point of the N-path filter, achieving superior performance with reduced circuit complexity. As a result, existing reconfigurable receiver topologies can be seamlessly transformed into harmonic blocker–resilient architectures. For example, a taped-out mixer-first receiver adopting this technique achieves a 100× improvement in third-harmonic blocker tolerance compared to state-of-the-art broadband receivers. This dissertation also proposes a reconfigurable receiver for IoT-class radios that is tolerant to both close-in and far-out blockers. A scalable clock bootstrapping technique is introduced to enhance linearity while maintaining both power and cost efficiency. All designs are validated through prototypes fabricated in advanced 22-nm and 45-nm silicon-on-insulator (SOI) technologies. By addressing this long-standing challenge, this work paves the way for fully reconfigurable, interference-resilient radios for 5G and beyond.

Video as the Language of Embodied Intelligence

2026-01-21T03:24:38Z

Video as the Language of Embodied Intelligence Chen, Boyuan Achieving general-purpose embodied intelligence remains a central challenge in artificial intelligence. While recent efforts have extended Large Language Models (LLMs) to robotics by incorporating additional modalities, these adaptations face critical limitations in perception, grounding, and control. For example, spatial reasoning—a simple yet indispensable capability for robots—reveals one of such shortcoming clearly: multimodal LLMs often fail even basic spatial perception tasks like estimating distances. This thesis begins by examining these failures through SpatialVLM, a system that augments vision-language models with 3D spatial reasoning. Although more effective in spatial estimation, this work reveals a deeper issue: the fundamental expressive limitations of language-only outputs in capturing sensorimotor dynamics. Based on these findings, the thesis advocates for a ground-up methodology for robot foundation models, starting with identifying an appropriate “language” for embodied AI, then architecting models and training regimes accordingly. We investigate video as the foundational language, integrated with model-based planning for decision-making. This new paradigm is instantiated through two core contributions. The first is Diffusion Forcing, a hybrid modeling framework that combines causal next-token prediction with full-sequence diffusion. This approach supports stable, coherent rollouts far beyond the training horizon and allows guided generation for decision-making tasks, bridging predictive modeling and planning. Building on Diffusion Forcing, we introduce the Diffusion Forcing Transformer (DFoT), a natural architectural extension designed for flexible video generation conditioned on variable-length histories. To further support long-horizon world-modeling, we propose History Guidance, a set of techniques that enhance sample fidelity, temporal consistency, and compositional generalization. Together, these methods enable robust modeling of visual dynamics across extended timeframes. Finally, we present a preliminary yet promising video foundation model for zero-shot robot motion planning, highlighting the potential of video as the foundational language of embodied intelligence.

Biologically Interpretable Representation Learning for Mechanistic Insights into Cancer Immunotherapy Resistance

2026-01-21T03:24:22Z

Biologically Interpretable Representation Learning for Mechanistic Insights into Cancer Immunotherapy Resistance Tariq, Ifrah Resistance to immune checkpoint inhibitors (ICIs) remains a critical barrier to effective cancer therapy, driven by complex, multi-scale interactions that current biomarkers often fail to capture. This dissertation introduces the Biologically Disentangled Variational Autoencoder (BDVAE)—an interpretable deep learning framework designed to uncover mechanistic drivers of ICI resistance through multi-omic data integration. Using RNA-seq and wholeexome sequencing data from 366 patients across melanoma, renal cell, urothelial, and gastric cancers, BDVAE learns low-dimensional latent representations that are both predictive of response and biologically meaningful. The model reveals distinct latent dimensions aligned with immune regulation, tumorintrinsic signaling, metabolism, and neuroimmune interactions. SHAP-based interpretation and pathway analysis highlight key resistance-associated programs, including immunosuppressive cytokine signaling, metabolic signaling, and neuroactive pathways such as calcium and cAMP signaling. Unsupervised clustering identifies three tumor subtypes—responder-dominant, non-responder-dominant, and an intermediate group—suggesting plastic or transitional immune states. Survival analyses confirm the clinical relevance of these clusters and expose heterogeneity within standard RECIST categories. Overall, this work presents a novel, interpretable framework for modeling ICI response, offering insights into resistance mechanisms and actionable paths for biomarker discovery, patient stratification, and therapeutic innovation in precision immuno-oncology.

Geometric interpretations of structural demand for the analysis and reduction of design complexity

2026-01-21T03:24:19Z

Geometric interpretations of structural demand for the analysis and reduction of design complexity Lee, Keith Janghyun This dissertation presents a computational framework to effectively interpret the distribution of structural demand that emerges from the design of large-scale structural systems, and develops methods for its quantification and manipulation. Structural demand is the required strength and geometry of individual building components that emerges from design as a result of global geometry, topology, and loading. Existing metrics of structural performance fail to consider how variations in demand at the component level can lead to designs that are theoretically efficient but difficult to construct. This has led to a rejection of low-carbon, high-performance design solutions in practice, or the need for extensive post-hoc rationalization, both under the presumption of untenable design complexity for conventional building practices. This dissertation argues that an explicit consideration of the distribution of induced structural demand can bridge this gap between design intent and construction feasibility. To achieve this, structural demand is interpreted as sets of geometric objects in n-dimensional feature spaces, where each dimension represents an independent component of demand, such as area, length, or stiffness. By directly visualizing the spatial distribution of demand, designers are presented with a richer context of non-physical structural design information, and can evaluate how decisions in structural form affect this distribution. Further, spatial interpretations of information allow for spatial metrics of similarity and variation to be defined, from which quantitative measures of design complexity are derived that account for the shape and distribution of demand. This framework, named \emph{Demand Space Analysis}, is explored in depth and applied to a range of structural scales, from the demand of truss elements and their connections, to the relationship between demand and fixed sets of capacity. Advancements in structural optimization are also presented, enabling more efficient and direct minimization of modern structural performance metrics, from which the relationship between design performance and demand complexity can be explored. Through case studies in each chapter, this dissertation demonstrates how geometric analysis of structural demand information can inform the designer of the implications of decisions on the perceived complexity of design, and provides tools for its quantification and reduction.

Behavioral Responses to Congestion Pricing in New York City: Mode Shift, Preference Change, and Effect Persistence

2026-01-21T04:07:41Z

Behavioral Responses to Congestion Pricing in New York City: Mode Shift, Preference Change, and Effect Persistence Shen, ChenAn This thesis examines the behavioral impacts of New York City’s congestion pricing policy on weekday peak-hour travel into the pricing zone. Using a two-stage Bayesian Multinomial Logit framework applied to monthly aggregate mobility data, the study disentangles underlying preference shifts from observed mode share changes in response to the toll. Stage 1 estimates population-level travel sensitivities to cost and time, while Stage 2 uses a hierarchical structure to capture heterogeneity across demographic segments defined by income, age, and gender. The analysis spans January–June 2025 and compares results to the same months in 2024 as a counterfactual scenario without pricing. Findings show that while the policy generated a sustained mode shift away from private automobiles toward public transit, preference adaptation varied by demographic group and evolved over time. Some cohorts reinforced the intended policy effects through reduced transit travel time sensitivity, while others exhibited partial reversal as cost sensitivity shifted. These dynamic patterns underscore the importance of evaluating both immediate and evolving behavioral responses when designing congestion pricing strategies and highlight the value of aggregate behavioral modeling for timely, data-driven policy assessment.

The Inhabited Arctic: Architecture, Time, and the Making of the Past in the Bering Strait (1760–1980)

2026-01-21T03:24:08Z

The Inhabited Arctic: Architecture, Time, and the Making of the Past in the Bering Strait (1760–1980) Springstubb, Phoebe Our view of antiquity is not objective. From the eighteenth century on, the same actors and institutions involved in colonizing the Arctic shaped understandings of its deep past. Commercial whalers erected outposts on the Arctic Ocean’s edges; miners stripped tundra; trading companies raised forts. The demands of these projects complicated the Western imperial fiction of an Arctic without a past. Grappling with Arctic terrain, foreigners were confronted by a landscape inhabited not only by people and animals but by time and temporal imaginations that long preceded European colonization. They encountered contemporary Indigenous settlements coexisting with ancestral houses, fossil animals, the ruins of earlier colonial ventures, and ancient routes of exchange. This dissertation, centered on the Bering Sea and its adjacent geographies of eastern Siberia and Arctic North America, tells the story of how imperial upheaval and the rooting of colonial projects in the ground sparked a deliberate historiographic project to write the Arctic’s deep past. At the heart of this project was a conflict of different cultural views of time. Who had the right to narrate history in these northernmost borderlands? In episodes spanning two centuries, from the Russian empire’s claim to the Bering Sea to the rise of modern decolonial movements, this dissertation traces the central role of diverse Native architectures and technologies. Iñupiaq houses built from great whale skeletons, Unangax watercraft hewn from circulating driftwood, and Chukchi ice cellars carved into permafrost were both prisms for temporal explanations and sites driving change. Russian colonial administrators, British geologists, US ethnographers, Orthodox priests, and Soviet engineers co-opted them to the lineal, geological, eschatological, and paleolithic time that scaffolded imperial projects. Simultaneously, these material practices were vital sites for reinvention and identity, where Native nations built futures out of rupture. Illuminating how the ecological and epistemic limits to empire-building spurred new theories of Arctic time, this project shows history-making to be a crucial tool different states adopted to justify and naturalize their possessions of Native lands. At stake was not static historical truth but how politically situated temporalities structured their present-day actions. The ethical dimensions of deep time, imagined from the Bering Strait’s modern lands and seas, empowered empire’s practical work. How the past was conceived in different intellectual traditions informed whether animals and plants were exploitable resources or ancestors giving their bodies to architecture. This project contends that how people understood themselves as being in time was a decisive fulcrum ordering collective beliefs in what was owed to a larger, nonhuman world. Taking time as an analytical lens, this dissertation identifies repeated efforts to cleave the Arctic’s human history from nature’s past. Used to justify a wide range of colonial hierarchies and violence in the long nineteenth century, it underlies a contemporary bias toward seeing the Arctic as a region of deep naturalism. Viewed as a place where an “extreme” climate dominates manifold other historicities, the past so circumscribed continues to shape future possibilities.

Asset Limits, Savings Behavior, and Welfare: Evidence from the SIPP and a Life-Cycle Model

2026-01-21T04:07:33Z

Asset Limits, Savings Behavior, and Welfare: Evidence from the SIPP and a Life-Cycle Model Gamble IV, James Monroe This paper examines how asset limits in means-tested welfare programs shape household saving behavior. I exploit cross-state variation in Temporary Assistance for Needy Families (TANF) asset limits by linking these limits to individual-level data from the Survey of Income and Program Participation (SIPP) and estimating ordinary least squares (OLS) regressions with state and year fixed effects. I find that a $1 increase in the liquid asset limit corresponds to a $0.75 decrease in non-housing wealth among single mothers without a high school diploma. This suggests that less stringent asset tests reduce incentives to save, consistent with models in which more generous public insurance lowers the need for precautionary saving. To interpret these findings, I develop a dynamic life-cycle model of saving under income and medical expense risk, calibrated to key moments from the Hubbard, Skinner, and Zeldes framework. The model embeds Medicaid-style transfer rules and a guaranteed consumption floor. Simulations indicate that a $7,000 consumption floor can reduce median assets by up to 20% among low-education households, reflecting a decrease in self-insurance as public support increases. I then extend the model to include Achieving a Better Life Experience (ABLE) accounts, which are tax-advantaged savings vehicles for individuals with disabilities exempt from means testing. Simulations indicate that ABLE eligibility increases early-life consumption by approximately $10,000 and reduces retirement savings, with account holders shifting more spending into their working years. Together, these results yield a direct mapping from policy levers, including asset-limit generosity, earnings disregards, childcare subsidies, and ABLE exemption rules, to predicted shifts in median household assets. This offers policymakers a practical tool to balance public insurance and private precautionary savings.

Building the 3D Genome from the Ground Up: Local Interactions Give Rise to Global Order

2026-01-21T03:23:55Z

Building the 3D Genome from the Ground Up: Local Interactions Give Rise to Global Order Athreya, Advait The three-dimensional organization of the genome within the nucleus plays a central role in determining gene regulation and establishing cellular identity, but the mechanisms by which local molecular interactions give rise to global chromatin architecture remain an active area of study. Interactions between nucleosomes—modulated by histone tail post-translational modifications, histone sequence variants, and the DNA sequence itself—are thought to be a major driver of this emergent structure. In this thesis, I address the question of how these intrinsic physicochemical properties of nucleosomes drive the formation of large-scale structures such as chromatin compartments. I develop a theoretical framework based on Flory-Huggins solution theory to derive pairwise internucleosome contact energies from the results of condense-seq, a novel experimental technique that measures the phase separation likelihood of native nucleosomes. I then use these derived energies to parameterize coarse-grained molecular dynamics simulations of chromatin at various resolutions, ranging from 25kb segments to simulate an entire chromosome, down to individual nucleosomes to simulate up to 10Mb genomic regions. These simulations demonstrate that the intrinsic nucleosome properties alone can capture a significant degree of A/B compartment formation observed in Hi-C experiments, despite the deliberate exclusion of all other factors such as loop extrusion and transcriptionfactor-mediated phenomena. This finding establishes that local nucleosome properties play a fundamental role in genome organization. To capture more detailed chromatin physics, I develop an extended chromatin force-field that incorporates anisotropic nucleosome stacking interactions and linker DNA properties using a novel approach for simulating reversible bond formation in molecular dynamics. This model reveals how nucleosome stacking strength, linker DNA geometry, and torsional stress collectively influence higher-order structures. Early results show that the linker-length-dependent DNA torsion contributes to nematic ordering of chromatin, consistent with experimental studies. Future development of this model will enable probing of discrete domain formation observed in imaging studies. Finally, I address a critical consideration for researchers in the chromatin organization field when analyzing Hi-C results. I compare two software tools — cooltools and dcHiC — highlighting the importance of careful parameter selection and analytical choices in designing workflows to ensure reproducible research. Taken together, this work establishes a quantitative, bottom-up modeling framework that directly links the local physicochemical properties of nucleosomes to the global principles governing three-dimensional genome organization. It provides a complementary approach to more data-driven top-down models that have made significant inroads but are challenging to interpret mechanistically. With further development, the work presented in this thesis will contribute towards predicting the structural consequences of specific epigenetic modifications and move us closer to understanding the molecular grammar of chromatin and its role in cellular function and disease.

On Subannual Variability in the Abyssal Ocean

2026-01-21T03:23:46Z

On Subannual Variability in the Abyssal Ocean Chen, Si Yuan The abyssal ocean is a critical yet understudied component of the climate system and is of growing economic interest. This thesis combines field observations and numerical modeling to advance our understanding of subannual variability in the abyssal ocean and its broader implications. First, hydrographic measurements from the Clarion-Clipperton Zone of the tropical Northeastern Pacific are used to characterize the structure and variability of the bottom mixed layer (BML) in a region targeted for deep-sea mining. The observations reveal a spatially and temporally variable BML with a mean thickness of ~250 m and influenced by interactions with mesoscale eddies and abyssal thermal fronts. A simplified model of sediment transport suggests that such variations in BML structure could significantly influence the dispersal of sediments resuspended by seabed mining activities. Second, idealized model experiments are conducted to explore the genesis of benthic storms – episodes of strong near-bottom flows and sediment entrainment – underneath an unstable, surface-intensified jet resembling the Gulf Stream east of Cape Hatteras. In these experiments, the baroclinic instability of the jet gives rise to deep cyclonic and anticyclonic eddies through eddy barotropization and to high levels of eddy kinetic energy at abyssal depths through the convergence of vertical eddy pressure fluxes. The near-bottom currents are comparable in magnitude to those observed during benthic storms, with vertical shears strong enough to produce BMLs up to O(100) m thick. Deep cyclonic eddies transport particles from near the bottom over the entire BML and could contribute to benthic nepheloid layers. The results suggest that the abyssal response to the intrinsic instability of surface-intensified currents could contribute significantly to subannual variability near the seafloor. Third, a model simulation of western North Atlantic circulation is performed to study the deep cyclones (DCs) observed beneath Gulf Stream meander troughs. The characteristics of the simulated DCs compare well with field observations. The negative pressure tendency during cyclogenesis arises from a small imbalance between the sea surface depression and the vertically-integrated increase in seawater density. Vortex stretching is the primary source of cyclonic vorticity, while vortex tilting is a non-negligible sink. The deep pressure tendency, vorticity fluxes, and ageostrophic flows are diagnosed, and their similarities and differences with mid-latitude synoptic cyclones in the atmosphere are discussed. Near-bottom currents in DCs dominate the basin-scale bottom energy dissipation and transport fluid over ≥1000 km horizontally and O(100) m vertically within 3~4 months, suggesting that they provide an efficient mechanism for tracer and material transport in the abyssal interior. Collectively, this thesis highlights the importance of transient, mesoscale processes in contributing to subannual variability in the abyssal ocean, particularly near the seafloor. The findings have broader relevance for monitoring the environmental impacts of human activities, including deep-sea mining and carbon sequestration. While further questions remain for future investigation, this work underscores the need for sustained in-situ observations in the abyssal ocean and calls for the implementation of high vertical resolution in numerical ocean circulation models.

Democratizing High-Performance DSL Development with the BuildIt Framework

2026-01-21T03:23:34Z

Democratizing High-Performance DSL Development with the BuildIt Framework Brahmakshatriya, Ajay Modern high-performance software from a variety of domains relies on hand-written and hand-optimized libraries to obtain the best performance. Besides general fine-grained operators that can be composed to write entire applications, these libraries also provide coarser-grained fused and hand-optimized operators that are much faster due to being optimized for a specific sequence of operations. However, as application needs keep growing, library writers are not able to keep up and have to make the tradeoff of either sacrificing performance or generality. Domain-specific languages or DSLs are able to break this tradeoff by automatically generating the best implementation for any arbitrary sequence of operations specified by the end user. However, DSL compilers suffer from a bigger challenge that they require a lot of compiler knowledge to implement parsers, IR, analysis and transformations, and code generation, which is outside the scope of a typical domain expert. To make compiler technology and the benefits of code-generation more accessible to domain experts, I propose the use of multi-stage programming to allow writers to write library-like code while also combining it to generate the most efficient implementation for any whole program. In this thesis, I discuss the design of different multi-stage programming systems, the benefits and drawbacks. Next, I propose Re-Execution Based Multi-Staging (REMS) that addresses a critical flaw in many imperative Multi-Staging systems - the side-effect leak problem. I introduce BuildIt, an implementation of REMS in one of the most popular languages for writing high-performance applications, C++ in a type-based, lightweight way without changing the compiler. I describe the internals of BuildIt and how it implements the key features of REMS. Furthermore, I describe a set of extensions implemented on top of BuildIt that facilitate the development of high-performance DSLs with ease. I show the application of BuildIt to create three DSLs - EasyGraphit, NetBlocks, and BREeze that target graph analytics, ad-hoc network protocol generation, and Regex matching. All these case studies show 10-100x reduction in the amount of effort required to implement these DSLs that perform on-par with or better than state-of-the-art compiler frameworks while targeting diverse architectures like CPUs and GPUS. Finally, I introduce D2X, a system that is designed to add extensible and contextual debugging support to DSL implementations without having to make any changes to off-the-shelf debuggers or mess with complex debugging formats. Next, I show how applying D2X to the BuildIt system greatly improves the debugging experience for all DSLs written with BuildIt.

Tailoring Li₄Ti₅O₁₂ Thin Film Carrier Kinetics Through Solid Solution Doping for Battery and Memristor Applications

2026-01-21T03:23:32Z

Tailoring Li₄Ti₅O₁₂ Thin Film Carrier Kinetics Through Solid Solution Doping for Battery and Memristor Applications Buzzell, Drew E. A Lithium titanate, Li₄Ti₅O₁₂ (LTO4), due to its zero-strain behavior during cycling, excellent chemical stability and cyclability, is a promising anode material for solid-state batteries (SSB) applications. As a thin film, its applications expand to integrated circuits, sensors, flexible batteries, IoT devices, and memristors. Across these, precise control of mixed Li⁺ ionic–electronic transport is vital. While dopants have been shown to improve electron conduction and Li⁺ diffusion in LTO4 powders, thin-film studies remain limited. To bridge this gap, we investigate solid solution dopants (Nb⁵⁺, V⁵⁺, Mg²⁺, Cu²⁺) and their effects on LTO4 thin-film kinetics and performance in batteries and memristors. Films doped with Mg, Cu, Nb, and V with a 0.2M dopant concentration were deposited on Nbdoped SrTiO₂ substrates. Cyclic voltammetry and impedance spectroscopy show that Mg, Nb, and V improve kinetic metrics, while Cu reduces diffusivity but boosts electronic conductivity. Through galvanostatic cycling-based capacity, rate capability, and stability measurements, we found that while all dopants displayed enhanced rate performance, the capacity improved only with Mg, Nb, and V. Furthermore, the Mg-doped film was found to have an unstable capacity leading to Nb- and V-doped thin-films as the best overall performing battery anodes. For memristors, current–voltage cycling measurements revealed that low concentrations (0.05 M) of Cu and Nb doped devices presented the largest improvements in cycle-to-cycle stability, switching ON-voltages, ON-OFF current ratios, and lower loss in peak current with increasing scan rate. With increasing dopant concentrations however, devices would see relative drops in performance. In summary, the inclusion of dopants in LTO4 at the right concentration level leads to improvements in both battery and memristor performance allowing for one material multi-functional systems.

Strain-resolved transcriptomics: exploring functional heterogeneity of the gut microbiota in health and disease

2026-01-21T04:08:09Z

Strain-resolved transcriptomics: exploring functional heterogeneity of the gut microbiota in health and disease Burgos Robles, Emanuel Felipe The gut microbiome plays a critical role in inflammatory bowel diseases (IBDs), yet current analyses treat bacterial species as functionally uniform, ignoring extensive strain-level diversity that may drive disease mechanisms. Here, we developed a strain-resolved metatranscriptomics framework to investigate how transcriptional activity varies across bacterial lineages and relates to IBD pathogenesis. Using paired metagenomics and metatranscriptomics data from 1,067 fecal samples (103 IBD and 335 non-IBD patients), we first constructed phylogenetic trees for over 250 bacterial species using the single nucleotide variants within essential housekeeping genes, enabling the identification of bacterial strains. Next, we devised a statistical approach to assign mRNA reads to these strains, leveraging the natural genetic variation that is present across them. My analysis revealed that closely related bacterial strains exhibit dramatically different transcriptional programs, with some strains enriched in IBD patients showing upregulation of genes involved in stress response, sugar metabolism pathways, and antimicrobial resistance. Notably, we identified transcriptionally active but genomically low-abundance taxa, highlighting the importance of measuring the transcriptional activities of strains beyond species composition. Lineage-aware differential expression analysis uncovered strain-specific adaptations to inflammatory environments. This strain-resolved approach provides a powerful framework for understanding microbial functional heterogeneity and identifying specific bacterial lineages that could potentially contribute to disease pathogenesis, potentially guiding more targeted microbiome-based therapeutic interventions.

The Role of CH–π Interactions in Protein-Carbohydrate Binding

2026-01-21T03:24:35Z

The Role of CH–π Interactions in Protein-Carbohydrate Binding Keys, Allison M. Protein-carbohydrate binding is essential for biological processes, including cellular recognition and immune signaling. Binding is driven by several types of non-covalent interactions: hydrogen bonding, metal ion coordination, and the less well-understood CH–π interactions. CH–π interactions are pervasive in protein-carbohydrate binding sites and have emerged as critical drivers of protein–carbohydrate recognition; however, the energetics of CH–π stacking interactions, their orientational landscapes, and their interplay with other non-covalent interactions have been unclear. In this thesis, I identified carbohydrate-aromatic CH–π stacking interactions from crystallographic structures in the Protein Data Bank. I performed quantum mechanical calculations to quantify interaction energies and found that CH–π stacking interactions can be more favorable than hydrogen bonds. Using atomistic simulations, I also demonstrated that CH–π stacking interactions are necessary for human galectin-3 binding to lactose. To assess the orientational landscape of CH–π stacking interactions, I evaluated the orientations of CH–π stacking interactions formed by β-D-galactose and found that numerous orientations are highly favorable. I then identified carbon atom distances that define an orientational landscape for these interactions. To assess the interplay between non-covalent interactions in protein-carbohydrate binding sites, I used CH–π distance features to bias metadynamics simulations of a curated set of protein–β-D-galactoside complexes. From these simulations, I found that while bound carbohydrates sample many CH–π stacking orientations, the hydrogen bonds in the protein binding site drive the optimal orientation of each ligand. Longer carbohydrate ligands with more hydrogen bonding constraints have more specific orientational dependence, while ligands in binding sites with a reduced number of hydrogen bonds occupy a broader range of orientations. Unlike hydrogen bonds, CH–π stacking interactions confer orientational flexibility: enzymes can exploit multiple CH–π stacking interactions to facilitate the translocation of polysaccharide substrates. Extending this analysis to other carbohydrates, I showed that carbohydrate stereochemistry drives the orientational preferences of CH–π stacking interactions; however, there is also a tradeoff between the presence of hydrogen bonds to charged amino acids and the CH–π interaction strength for each carbohydrate. Overall, this thesis demonstrates that CH–π interactions are favorable and confer high orientational flexibility and that hydrogen bonds act in concert with CH–π interactions to stabilize protein-carbohydrate binding. Tuning the number and positions of these interactions through protein engineering should alter protein selectivity and ligand movement in protein binding sites.

Quantum Networking using Waveguide Quantum Electrodynamics

2026-01-21T03:23:30Z

Quantum Networking using Waveguide Quantum Electrodynamics Almanakly, Aziza The architectural principle of modularity enables the construction of complex systems from simpler components, each responsible for a particular function. The quantum computer is an intricate system comprising fragile, error-prone parts known as qubits. Entanglement distribution across a network of non-local processing modules facilitates robust and extensible quantum computation. In modular quantum architectures, photons are natural quantum information carriers which propagate through interconnects between processing nodes. In this thesis, we engineer a quantum interconnect between superconducting modules underpinned by the physics of waveguide Quantum Electrodynamics (wQED). First, we realize a multi-qubit module that exploits quantum interference to emit microwave photons into a waveguide with a specified propagation direction. Next, we construct the quantum interconnect by coupling two modules to a common waveguide and demonstrate directional (chiral) photon emission and absorption. Finally, using this chiral quantum interconnect, we generate remote entanglement, establishing a key resource for distributed quantum computation in an all-to-all network architecture.

Large Language Models and Quantifying the Regulatory Expenses of Affordable Housing: A Thorough Examination Utilizing Generative Assessment

2026-01-21T04:08:08Z

Large Language Models and Quantifying the Regulatory Expenses of Affordable Housing: A Thorough Examination Utilizing Generative Assessment Xu, Bangjie This thesis presents an innovative methodology using Large Language Model-based methods to extract and quantify housing regulations from municipal zoning codes, making possible the most comprehensive examination of regulatory costs at the municipal level across California to date. A multi-staged extraction framework is devised that delivers 85-95% accuracy in the identification and standardization of complex regulatory requirements from legal documents. Applying this methodology to over twenty California cities over the period 2015-2025, it is estimated that regulatory constraints raise the cost of developing a housing unit by roughly between 5% to 10% (or $50,000 and $100,000+) per housing unit, with the most acute constraints in the state’s coastal metros. This method is used to find that factors such as regulation costs limit housing supply elasticity from 1.24 in low-regulation jurisdictions to 0.08 in high-regulation areas. The LLM-based framework allows us to conduct analyses at an unprecedented scale and granularity and to reveal, for example, that the relaxation of regulation by streamlining policies like the Los Angeles Transit Oriented Communities program boosts housing production in eligible zoned areas by 43%. This study makes significant contributions to the restructuring of California’s housing regulation system in response to the affordability crisis, and its methodology presents a replicable tool for regulatory analysis in other policy domains.

Essays in Macro-Finance

2026-01-21T03:24:44Z

Essays in Macro-Finance Batista, Quentin In Chapter 1 (joint with J.R. Scott), we revisit the high-frequency and narrative approaches to estimating the effects of monetary policy shocks. We find that state-of-the-art estimates using both approaches are biased: high-frequency estimates due to nonlinear predictability and narrative estimates due to regularization. To correct for the bias in these approaches, we propose a new estimation procedure called LP-DML that combines ideas from double/debiased machine learning with the local projections framework. We find that LP-DML results in significantly smaller effects of monetary policy on macroeconomic outcomes. In Chapter 2 (joint with Taisuke Nakata and Takeki Sunakawa), we study the following question: how a central bank credibly implement a ”lower-for-longer” strategy? To answer this question, we analyze a series of optimal sustainable policy problems—indexed by the duration of reputational loss—in a sticky-price model with an effective lower bound (ELB) constraint on nominal interest rates. We find that, even when the central bank lacks commitment, the central bank can still credibly keep the policy rate at the ELB for an extended period though not as extended as under the optimal commitment policy—and meaningfully mitigate the adverse effects of the ELB constraint on economic activity. In Chapter 3, I examine the impact of central bank real estate purchases on financial markets, focusing on the Bank of Japan’s (BoJ) intervention in the Real Estate Investment Trust (REIT) market. Using a regression discontinuity design that exploits a discontinuity in the BoJ’s policy rule, I find that a typical intervention — amounting to about 0.014% of market capitalization — leads to an increase of 0.1% to 0.2% of REIT prices in the hours following the intervention. However, at longer horizons, the interventions do not have a significant effect on REIT prices. These findings suggest that the BoJ did not achieve the program’s intended objective of significantly reducing the risk premium on real estate assets.

Toward an Integrative Study of Human-AI Interaction

2026-01-21T03:24:10Z

Toward an Integrative Study of Human-AI Interaction Alsobay, Mohammed As artificial intelligence (AI) systems are increasingly embedded in the workflows of individuals and groups, designers and researchers of human-AI interaction (HAI) navigate a vast design space of possible configurations, making decisions that span algorithmic parameters, interface choice, and interaction protocols. This thesis develops an integrative approach that examines how design factors combine and interact to determine the outcomes of human-AI collaboration. Chapter 1 synthesizes prior HAI research into a coherent design space framework encompassing algorithms, interfaces, users, and task settings, motivating a research program for systematic exploration of interdependencies between these factors. Chapters 2 and 3 turn to group-AI interaction through large-scale behavioral experiments. Chapter 2 investigates how social information---both direct conversation and peer behavior indicators---affects individual reliance on algorithmic decision support. The study reveals that while social information modulates the effects of performance feedback and model explanations on reliance, it does not improve predictive accuracy, illuminating critical tensions between social mechanisms and system design. Chapter 3 examines large language models as facilitators of group deliberation in hidden profile tasks. While LLM facilitation increased information sharing volume, density, and breadth, it did not improve decision quality, highlighting fundamental challenges in group-AI system design beyond information aggregation. Chapter 4 advances an integrative approach to HAI research, emphasizing shared design spaces, systematic exploration strategies, and predictive models that generalize across contexts. The chapter provides methodological guidance and a tractable roadmap for advancing this integrative research agenda, laying the foundation for a more context-aware science of human-AI collaboration.

Leveraging design to build with less: Evaluating the embodied carbon reduction potential of architectural design across scales

2026-01-21T03:24:40Z

Leveraging design to build with less: Evaluating the embodied carbon reduction potential of architectural design across scales Feickert, Kiley Reducing embodied carbon (EC) in structural systems -- the most significant contributor to EC in a building -- is urgent to address the simultaneous need to reduce global warming and increase urban density. Much of the policy and research to date to reduce EC has focused on material-scale interventions or substitutions. However, EC depends on both: 1) the carbon intensity of the processes used to manufacture construction materials, and 2) the volume of raw materials required. Architects have significant agency to reduce the volume of structural materials in a building (and the resulting emissions) since the required quantity depends on design decisions architects make, including column spacing, structural typology, massing, etc. To date, most methods used to estimate EC during early-stage design do not: 1) integrate with architects’ existing design workflows, 2) evaluate multiple material systems simultaneously, and/or 3) include structural analysis to estimate material quantities. This functionality is critical so that designers can understand which decisions EC is sensitive to and evaluate design and EC tradeoffs before significant carbon is locked in. To address this problem, this dissertation presents a method towards transparent estimation of structural material quantities, intending to inform architectural design and policy, or other emerging EC standards. This method is used to contribute an analysis of the effectiveness of emerging U.S. EC policies, which focus on different scales of intervention, at the building scale. These policies are evaluated in isolation and in combination with strategic design levers that take advantage of structural mechanics to reduce material quantities for various building configurations and material systems. It finds that the most prominent policy approach, “Buy Clean” materials, only reduces EC by ~9% and ~16% for steel and concrete systems, respectively, compared to strategic design choices that have the potential to yield savings of up to ~79%. This dissertation also identifies building massing as a key lever in the EC outcomes of structural systems and proposes a method to quantify the impact of massing using automated structural design and analysis. It finds that in some situations, cantilevered massing typologies can be materialized for no carbon penalty if efficient configurations are used. Indeed, if inefficient configurations are used, they can incur a significant carbon penalty (2.4x) compared to normative massing. The presented results highlight the potential of design to reduce demand-side EC across scales.

Generalizable Robot Manipulation through Unified Perception, Policy Learning, and Planning

2026-01-21T03:23:49Z

Generalizable Robot Manipulation through Unified Perception, Policy Learning, and Planning Fang, Xiaolin Advancing robotic manipulation to achieve generalization across diverse goals, environments, and embodiments is a critical challenge in robotics research. While the availability of data and large-scale training has brought exciting progress in robotics manipulation, current methods often struggle with generalizing to unseen, unstructured environments and solving long-horizon tasks. In this thesis, I will present my work in robot learning and planning that enables multi-step manipulation in partially observable environments, towards general-purpose embodied agents. Specifically, I will talk about my work in 1) constructing a modular framework that estimates affordances with learned perception models with task-and-motion-planning (TAMP) for object rearrangement in unstructured scenes, 2) learning generative diffusion models of robot skills, which can be composed to solve unseen combination of environmental constraints through infeference-time optimization, 3) leveraging large vision-language models (VLMs) in building task-oriented visual abstractions, allowing skills to generalize across different environments with only 5 to 10 demonstrations. Together, these approaches contribute to the generality and scalability of embodied agents towards solving real-world manipulation in unstructured environments.

Towards High-Dimensional Generalization in Neural Networks

2026-01-21T03:24:21Z

Towards High-Dimensional Generalization in Neural Networks Boopathy, Akhilan Neural networks excel in a wide range of applications due to their ability to generalize beyond training data. However, their performance degrades on high-dimensional tasks without large-scale data, a challenge known as the curse of dimensionality. This thesis addresses this limitation by pursuing three key objectives aimed at understanding and improving neural network generalization. 1. We aim to investigate the scaling laws underlying generalization in neural networks including double descent, a phenomenon in which as a model’s capacity or training data is increased, the test error temporarily increases at a certain point before continuing to decrease. In particular, we will have two goals: 1) a better understanding of when double descent can and cannot be empirically observed and 2) a better understanding of scaling laws with respect to training time. 2. Inductive bias refers to the set of assumptions a learning algorithm makes to predict outputs on inputs it has not encountered. We propose quantifying the amount of inductive bias required for a model to generalize well with a fixed amount of training data. By developing methods to measure inductive bias, we can assess how much information model designers need to incorporate into neural networks to improve their generalizability. This quantification can guide the design of harder tasks that better test a model’s generalization. 3. Finally, we aim to develop new methods to enhance neural network generalization, particularly focusing on reducing the exponential number of training samples required for high-dimensional tasks. This involves creating algorithms and architectures that can learn effectively from limited data by incorporating stronger inductive biases. In particular, we will focus on two inductive biases in particular: 1) learning features of the training loss landscape correlated with generalization and 2) using modular neural network architectures. We expect that these techniques can improve generalization, particularly in high-dimensional tasks. Together, these contributions aim to deepen our theoretical understanding and develop practical tools for enabling neural networks to generalize effectively from limited data.

Remote Control: Art, Technology, and the Politics of Distance (1966-1972)

2026-01-21T03:23:42Z

Remote Control: Art, Technology, and the Politics of Distance (1966-1972) Wexelblatt, Nina Rrose Platforms carrying dancers across a stage, doors sliding open as if by magic, and simultaneous Happenings in Berlin and Buenos Aires: remote control promised thrills as postwar artists experimented with technologies of distance. Focused on the half-decade between 1966 and 1972, this thesis intervenes in the history of art and technology to argue that a desire to activate the supposedly empty space between artist, art object, and audience effected a new fixation on the nature of that distanced interval, leading artists to incorporate actual remote control technologies into their work. This impulse grew from an unorthodox reading of the work of modernist painters, particularly Jackson Pollock. Where a generation of critics had canonized “presentness” and medium specificity, a younger cohort read the work differently, finding in it permission to embrace remoteness, intermedia experimentation, and political messaging. Artists including Robert Rauschenberg, Allan Kaprow, Marta Minujín, Wolf Vostell, and Carolee Schneemann, among others, undertook radical experiments with remote systems, often in collaboration with engineers. Theirs was not a technocratically neutral position; this thesis demonstrates that these artists consciously cast the “remoteness” enabled by new technologies as a charged concept, just as controlled distance emerged to define military and industrial relations on domestic, urban, and geopolitical scales. Remote control enabled artists to incorporate, not reject, the expanding frames of reference taking place outside of the sanctioned spaces of the art studio or gallery, from automation to satellite communications to warfare. Artists’ uses of remote technologies intentionally surfaced questions about critical power relations, tying the stakes of their work to debates about the future of U.S. social and economic control and development. In doing so, it also crystallized a newly diffuse, participatory artistic subject: the controller. The introduction theorizes “remote control” in historical and historiographic context. A second chapter follows Automation House (1970-1972), a Manhattan art space that combined labor mediation and media art to experiment with the American postindustrial labor economy to come. A third chapter centers on Three Country Happening (1966), which took place in New York, Buenos Aires, and Berlin, supposedly mediated by satellite—foiled by the uneven development of the Cold War-era satellite system itself. A fourth chapter delves into Snows (1967), a multimedia performance in protest of the war in Vietnam, which incorporated audience-controlled feedback sensors. A concluding discussion traces the ongoing nature of remote control as it implicates artists and audiences alike in a network of shared responsibility.

Reimagining Public Land: Municipal Land Use to Ease Housing Crisis in Boston

2026-01-21T04:08:06Z

Reimagining Public Land: Municipal Land Use to Ease Housing Crisis in Boston Murphy, Ryan Boston is in the midst of a severe housing crisis, driven by decades of underproduction, rising construction costs, restrictive zoning, and an inelastic real estate market that has resulted in persistent affordability challenges. This thesis explores the untapped potential of city-owned land as a powerful tool to increase housing supply and affordability in Boston. Using Boston’s 2022 Citywide Land Audit and detailed development assumptions, the analysis estimates that between 19,000 and 31,000 new housing units could be constructed across city-controlled parcels, including between 3,200 and 6,100 affordable units under the current Inclusionary Development Policy. The research draws on case studies from peer cities such as Chicago and Atlanta where municipal land has been successfully leveraged through transparent disposition processes, fast-tracked entitlements, and flexible affordability models. It argues for a policy shift in Boston toward a more streamlined, market-aware, and scalable land release strategy that prioritizes speed, cross-subsidization, and financial feasibility. Key recommendations include expanding the Welcome Home, Boston program to include mixed-income and rental housing, implementing predictable RFP cycles, offering tax abatements, and expediting the entitlement process.

Modular Zipping for Transformable and Dynamic Systems

2026-01-21T04:08:01Z

Modular Zipping for Transformable and Dynamic Systems Hagemann, Niklas There is a need for products, machines and environments that can change shape, transform and evolve according to their use. This thesis proposes the design of a simple, modular actuator based on reversible folding and interlocking (zipping) of flexible 3D printed strips. The proposed zipper design allows for continuous control states between a compact and fully deployed state. The modular actuators can be integrated into a variety of systems to enable compact, shape- and stiffness-changing structures, robots and other devices. Designs are presented for single- and double-zipper modules using the same basic zipper design. The modules can be used as modular components of compact robotic systems with the ability to expand and contract according to their environment, or used as adjustable structural components to create deployable, shape-and stiffness-changing objects. The zipper design points the way towards simplified mono-material components that embed transformation and reversibility into everyday devices, products and spaces, and enabling objects that are as easy to transform, reconfigure and reverse as they are to manufacture.

Embodied Representation of Time in Virtual Reality

2026-01-21T04:08:03Z

Embodied Representation of Time in Virtual Reality Kim, Suwan Recent advancements in 3D graphics and AI-assisted generative techniques have accelerated the creation of realistic scenes for immersive technologies, including virtual reality, yet most systems continue to encode time as a linear parameter, relying on timeline-based playback. Mesh-based representations are typically constrained by fixed topologies and rely on predefined animations, which limit their capacity to encode temporal change as a spatial or perceptual phenomenon. In reality, human experience of time is embodied and dynamic, perceived through interaction and memory. Existing digital systems fail to capture this dimension, reducing time to a passive parameter. This thesis proposes a framework for representing time as an embodied and spatial dimension within virtual reality by embedding it directly into the geometry and interaction logic of point cloud data. The system consists of three parts: (1) processing 2D images into layered volumetric point clouds to enable structural fluidity and temporally responsive spatial form; (2) enabling perceptual and spatial modulation in response to user distance and contact, with color influencing the character of change and opacity shaping its perceptual reveal at both global and local scales; and (3) enabling real-time visualization of modulated point cloud through a custom pipeline optimized for mobile virtual reality. By embedding temporal dynamics directly into geometry and interaction logic, this thesis contributes a novel representational approach to spatiotemporal modeling in immersive systems. By doing so, we create new opportunities for architectural visualization, interactive simulations, game design, and reimagining how we perceive and construct digital spaces.

Essays on Behavioral Economics and Sophisticated Procrastination

2026-02-20T03:14:30Z

Essays on Behavioral Economics and Sophisticated Procrastination Chen, Xi Procrastination is a widespread yet complex behavior that resists simple explanation. This dissertation integrates theoretical modeling with experimental evidence to examine procrastination through the lens of sophisticated decision-making. It reframes procrastination not merely as a deviation from rationality, but as a behavior shaped by strategic trade-offs, self-awareness, and individual heterogeneity. The first essay develops a theoretical model of Perfectionistic Procrastination, proposing that individuals with high internal standards may delay tasks not as a simple lapse in self-control, but as a strategic response to the anticipated costs of sustained effort. In this framework, deadlines act as external constraints that help perfectionists limit open-ended striving and bring tasks to completion. An accompanying experiment tests the model’s prediction and finds that perfectionists are more likely to prefer deadlines. These results suggest that, in some cases, procrastination may reflect a structured strategy rather than a purely irrational failure of self-control. The second essay explores the phenomenon of Sophisticated Procrastination, challenging traditional models that attribute procrastination to naïveté. Instead, it proposes that even individuals who are aware of their tendency to delay may struggle to act on that awareness. Two experimental studies using a menu-choice framework examine how people choose task timings. In Study 1, participants preferred earlier deadlines when flexibility was available but shifted toward later options when required to commit, revealing a gap between intention and action. Study 2 identified diverse patterns of deadline preferences: while many participants actively avoided the latest possible deadline, their hesitation to commit to any specific deadline suggests a deeper tension rooted in uncertainty or discomfort with commitment. These findings provide early empirical support for Sophisticated Procrastination, indicating that self-awareness alone may not be sufficient to overcome procrastination. The third essay introduces the idea of Prosocial Procrastination, describing the tendency to delay tasks that benefit others, such as charitable activities, more than those with self-interested outcomes. Using two distinct experimental designs, one based on conjoint analysis and the other on single-attribute choice, the studies show that individuals are more likely to prefer longer deadlines when working for a charity than when working for themselves. These findings offer suggestive evidence for Prosocial Procrastination and contribute to the growing literature on the intersection of social preferences and time preferences.

Patent Visibility and the Diffusion of Trapped Knowledge: Evidence from US Grants

2026-01-21T04:08:04Z

Patent Visibility and the Diffusion of Trapped Knowledge: Evidence from US Grants Yao, Randol H. Valuable knowledge developed in one part of the world may remain “trapped" locally due to frictions in how knowledge is recognized and shared globally. This paper examines how granting US patents to foreign-origin inventions—by elevating their visibility and credibility— untraps the knowledge and facilitates global diffusion. Using examiner leniency as an instrument, complemented by a difference-in-differences design, I find that US grants of home country patents significantly increase both the likelihood and intensity of forward citations, including marked increases from third countries. A novel measure of “trappedness” reveals that knowledge from historically more trapped countries and sectors sees larger diffusion benefits after the US grants. These findings highlight the central role of the US as a platform of global knowledge recognition and diffusion, particularly in turning overlooked ideas into globally relevant innovations.

Characterization of the East China Sea Continental Shelf Circulation Northeast of Taiwan Surrounding Mien-Hua Canyon

2026-01-21T04:07:59Z

Characterization of the East China Sea Continental Shelf Circulation Northeast of Taiwan Surrounding Mien-Hua Canyon Rafferty, Lieutenant Commander Keefe Submarine canyons have a proven and direct influence on continental shelf circulation and flow dynamics, especially in relation to western boundary currents. There are two key circulation features northeast of Taiwan on the East China Sea continental shelf: (1) the cold dome, a cyclonic feature that appears primarily in summer and is associated with upwelling, and (2) Kuroshio intrusions onto the continental shelf in the vicinity of Mien-Hua Canyon. This paper is a descriptive physical oceanography study with a focus on characterizing the circulation patterns northeast of Taiwan surrounding Mien-Hua Canyon, closely correlating these patterns with the migration of the Kuroshio and its variability and intrusions onto the southern East China Sea continental shelf, leading to the formation of the cold dome. The Institute of Oceanography at the National Taiwan University and WHOI executed a joint international field survey at Mien-Hua Canyon aiming to improve the understanding of canyon flow dynamics between the East China Sea continental shelf northeast of Taiwan and the Kuroshio as the North Pacific Gyre westward boundary current. This joint oceanographic expedition expands on previous joint US/Taiwan physical oceanographic and ocean acoustic studies in the China Seas dating back to ASIAEX in the South China Sea during 2000-2001 and QPE in the East China Sea during 2008-2009. The strengthening and weakening of Kuroshio transport and intensity northeast of Taiwan is closely correlated to the timescales of mesoscale westward propagating eddies arriving to the East Taiwan Channel. When a canyon has a Rossby number ~1 or Rossby radius equivalent to the width of the canyon in a region of left-bounded flow, induced cyclonic flow will experience an upwelling regime within the canyon system with dominant upwelling located at the downstream canyon rim vertically constrained by Rossby Height. Observational analysis of canyon bottom-moored ADCPs and vertical temperature arrays supports previous theory on submarine canyon dynamics on a continental shelf. Satellite sea surface temperature and absolute dynamic topography observations render the formation of a cold dome northeast of Taiwan coincident with this joint oceanographic survey.

From Wallets to Wages: Consumer Income, Job Design, and Pay Disparities

2026-01-21T04:08:01Z

From Wallets to Wages: Consumer Income, Job Design, and Pay Disparities Roh, Soohyun Pay differences between organizations are a key source of wage inequality. I propose a novel account of these differences by starting from the consumers that these businesses serve. Firms that serve high-income consumers specialize jobs into higher-paying and higher-skilled positions focused on quality, while those that serve lower-income consumers emphasize cost minimization by requiring workers to perform a wider range of general tasks. Matching consumer foot traffic data and establishment-level wage records, I find that establishments serving higher-income consumers pay their workers more. This effect holds comparing among establishments in the same neighborhoods and industries. Longitudinally, establishments increase wages when they shift toward higher-income customers. Analysis of online job postings further reveals that jobs at higher-income-serving firms involve a narrower set of tasks that command higher market value. These findings show how consumer markets shape firms’ internal job design and contribute to pay inequality across organizations.

The Fate of Federal Buildings: Real Estate Disposition and the Future of Washington, D.C.

2026-01-21T04:07:56Z

The Fate of Federal Buildings: Real Estate Disposition and the Future of Washington, D.C. Mulcahy, Robby L. The United States federal government is the largest property owner in the country, with more than 370 million square feet of real estate under its control. Much of this portfolio is outdated, underutilized, and located in the urban cores of American cities. Nowhere is this more evident—or more consequential—than in Washington, D.C., where the federal government controls approximately 27% of the office market. As federal agencies adopt hybrid work models, and as the operational needs of government evolve, the existing real estate footprint has become increasingly inefficient, expensive, and misaligned with civic and market realities. This thesis investigates the opportunity to rethink federal land ownership and management as a catalyst for urban regeneration, civic stewardship, and housing production. Using the James V. Forrestal Building as a focal case study, the research examines the historical, policy, and spatial dynamics that have led to the current moment of reckoning. Located on Independence Avenue SW, straddling 10th Street between the National Mall and the Wharf, Forrestal is emblematic of the postwar federal design ethos: monumental, inward-facing, and hostile to street life. Once a symbol of bureaucratic permanence, the building now stands as a physical and symbolic barrier to urban connectivity and civic vitality. The case of Forrestal is used to explore broader questions: How can the federal government dispose of surplus property more effectively? What policy tools exist—or are needed—to unlock value and enable redevelopment? And what role should cities play in shaping the outcomes of federal land disposition? The thesis employs a mixed-methods approach that includes policy analysis, stakeholder interviews, precedent case studies, and spatial analysis of Southwest D.C. The work identifies a range of obstacles to effective disposition, including Title V of the McKinney-Vento Homeless Assistance Act, opaque OMB budget scoring rules, jurisdictional fragmentation, and the absence of a coordinating authority across federal agencies. It also identifies key lessons from successful projects such as The Yards, Walter Reed, and the Volpe Center, where thoughtful structuring and strong federal-local partnerships enabled transformative redevelopment of surplus land. The thesis concludes with ten detailed recommendations for reform, including reauthorization of the Federal Assets Sale and Transfer Act (FASTA), modernization of Title V and OMB scoring, the creation of Federal Redevelopment Zones, and the prioritization of housing, civic infrastructure, and design quality in disposition strategy. It argues that the federal government must shift from a passive landlord to an active steward of public land—one that collaborates with cities, integrates public benefit, and reflects democratic values through the built environment. In this moment of shifting federal needs, declining office demand, and urban transformation, the question is not whether federal real estate reform is needed—it is whether we will seize the opportunity. The fate of buildings like Forrestal will shape not only the skyline of Washington, D.C., but also the federal government’s legacy in America’s cities for generations to come.

Analyst Incentives

2026-01-21T04:08:00Z

Analyst Incentives Green, Brice Analyst forecasts have been shown to reflect substantial behavioral biases and predict a number of macroeconomic phenomena. While we typically treat reported forecasts as statistical expectations, under uncertainty the reported point estimate will be sensitive to the payoff structure facing the forecaster. Using data on careers from LinkedIn, I describe the incentive structures faced by analysts, shedding light the extent to which pay and career success are tied to performance. Further, I extend a causal estimator to identify credible counterfactual forecasts and provide tentative causal evidence of the relationship between forecast errors and promotions.

Maverick Neuroscientist: What Does a Life in Science Look Like Outside the Ivory Tower? : Dr. Eugenio Vargas-Pena is a renowned psychiatrist in Paraguay, who conducts neuroscience research without university ties, funding, or peer review. ls his embodiment of the gentleman scientist an alternate path for those who want to break away from institutionalized science?

2026-01-21T04:07:57Z

Maverick Neuroscientist: What Does a Life in Science Look Like Outside the Ivory Tower? : Dr. Eugenio Vargas-Pena is a renowned psychiatrist in Paraguay, who conducts neuroscience research without university ties, funding, or peer review. ls his embodiment of the gentleman scientist an alternate path for those who want to break away from institutionalized science? Chomik-Morales, Jessica This longterm narrative investigates the life and work of Dr. Eugenio Vargas-Peña, a neuropsychiatrist in Asunción, Paraguay who built a fully functional lab in his countryside home. Vargas-Peña conducts brain research independently, guided by decades of self-study, clinical practice, and an unwavering belief in the value of curiosity-driven inquiry. The piece interweaves historical context, character study, and personal narrative, using the author's own background in neuroscience and science communication to frame an inquiry into legitimacy, recognition, and alternative pathways in science. It asks: What defines a scientist today? Who gets to decide which ideas are taken seriously? And what are the consequences-creative or catastrophic-of working outside institutional boundaries? Through the lens of one man's eccentric yet earnest intellectual journey, this thesis invites broader reflection on the pressures shaping contemporary research and the enduring romance of unorthodox scholarship.

Machine Learning Systems for Unsupervised Time Series Anomaly Detection

2026-01-21T03:23:36Z

Machine Learning Systems for Unsupervised Time Series Anomaly Detection Alnegheimish, Sarah Modern assets – from launched satellites to electric vehicles – output dense, multivariate time series data that must be monitored for deviations from “normal” behavior. This monitoring task is referred to as time series anomaly detection. The current state of the industry still depends on fixed or heuristic thresholds that often drown operators in false alarms, and can miss the subtle, context-dependent faults that matter most. This thesis addresses unsupervised time series anomaly detection as an end-to-end problem, asking how we can learn, evaluate, and deploy models that judiciously flag anomalies while remaining intuitive to the end user. This thesis provides contributions in the form of both algorithms and systems. First, it introduces three models that enlarge the design space of unsupervised time series anomaly detection: TadGAN, which leverages adversarial reconstruction; AER, which unifies predictive and reconstructive objectives in a single hybrid score; and MixedLSTM, which explicitly incorporates interdependencies to improve anomaly detection in multivariate time series. We propose two range-based evaluation metrics that quantify detection quality over temporal intervals. Second, it presents our system Orion, which abstracts anomaly detection pipelines as directed acyclic graphs of reusable primitives, providing user-friendly APIs and enabling interactive visual inspection. Building on this infrastructure, OrionBench performs periodic, fully reproducible benchmarks, producing leaderboards that align research innovations with the needs of end users. Third, the thesis explores a new paradigm – foundation models for unsupervised time series anomaly detection – by formulating SigLLM , which employs large language models and time series foundation models for zero-shot anomaly detection via prompting and forecasting. This paradigm indicates a promising path to developing scalable models for anomaly detection. Finally, beyond evaluating our systems on publicly available datasets, we provide extensive experiments on two industrial case studies that demonstrate improved detection accuracy and practical usability of our system.

Techniques for Reliability and Robustness in Integrated Electronic and Photonic Systems

2026-01-21T03:23:15Z

Techniques for Reliability and Robustness in Integrated Electronic and Photonic Systems Chakraborty, Uttara Reliability and robustness are key concerns in the development of novel electronic and photonic materials, devices, and systems. This thesis presents statistical and machine learning techniques for reliability analysis of heterogeneously-integrated systems, extraction of variations from photonic test structure measurements, making smart decisions about test configurations in the face of time and resource constraints, and robust design of photonic components. To estimate reliability model parameters from lifetime datasets where multiple underlying failure mechanisms are present, a differential evolution framework and a boundconstrained expectation maximization algorithm are developed; both these approaches significantly outperform the gradient-based L-BFGS-B algorithm. New schemes for strategic failure analysis on a subset of the failed units are presented, both for detecting the presence of a second failure mechanism and for improving two-mechanism reliability models. A regression-based protocol is also presented for optimally selecting reliability test conditions to verify physical failure mechanism models. A maximum-likelihood-estimation-based approach is demonstrated for the simultaneous extraction of waveguide index and thickness variations using integrated photonic direction couplers and Mach-Zehnder interferometers. Schemes are proposed for optimal selection of cut-back test structures and for propagation loss estimation with a Bayesian prior distribution for fiber-coupling error. Finally, a robust Bayesian optimization algorithm using a new tunable acquisition function is presented for photonic component design. The methods developed in this thesis are expected to be broadly applicable to a wide variety of electronic and photonic devices and systems.

Molecularly Thin Polyaramid Nanomechanical Resonators

2026-01-17T06:32:32Z

Molecularly Thin Polyaramid Nanomechanical Resonators Gress, Hagen; Ritt, Cody L; Shomakhov, Inal; Altmisdort, Kaan; Quien, Michelle; Wei, Zitang; Lawall, John R; Boddeti, Narasimha; Strano, Michael S; Bunch, J Scott; Ekinci, Kamil L Two-dimensional polyaramids exhibit strong hydrogen bonding to create molecularly thin nanosheets analogous to graphene. Here, we report the first nanomechanical resonators made out of a two-dimensional polyaramid, 2DPA-1, with thicknesses as small as 8 nm. To fabricate these molecular-scale resonators, we transferred nanofilms of 2DPA-1 onto chips with previously etched arrays of circular microwells. We then characterized the thermal resonances of these resonators under different conditions. When there is no residual gas inside the 2DPA-1-covered microwells, the eigenfrequencies are well-described by a tensioned plate theory, providing the Young's modulus and tension of the 2DPA-1 nanofilms. With gas present, the nanofilms bulge up and mechanical resonances are modified due to the adhesion, bulging and slack present in the system. The fabrication and mechanical characterization of these first 2DPA-1 nanomechanical resonators represent a convincing path toward molecular-scale polymeric NEMS with high mechanical strength, low density, and synthetic processability.

Interferometric Deflection Analysis of Suspended 2D Polyaramid Thin Films

2026-01-17T06:32:30Z

Interferometric Deflection Analysis of Suspended 2D Polyaramid Thin Films Quien, Michelle; Ritt, Cody L; Garimella, Sanjay S; Gress, Hagen; Ekinci, Kamil L; Bunch, Joseph Scott; Strano, Michael S The 2D nanofilm bulge test, which uses an Atomic Force Microscope (AFM) to measure the deflection of a suspended film under various conditions, has emerged as an important measurement platform for understanding mechanical, barrier, and permeability properties of 2D materials as thickness approaches the angstrom scale. The problem considered in this work is the limitation of such bulge analyses imposed by the AFM whereby dynamic measurements under high pressure, high temperature, and chemically corrosive conditions are limited. In this work, a technique is developed for measuring nanofilm deflection using only visible light interferometry. Both theoretical and semi‐empirical models are applied to translate multicolor interference patterns from broadband excitation into estimates of nano‐film deflection, allowing nanoscale precision in most cases. The technique and algorithm advanced in this work allows the use of widespread optical microscopy to widen the study of these important 2D nanofilm systems to more relevant conditions.

Quantum One-Time Programs, Revisited

2026-01-17T06:31:55Z

Quantum One-Time Programs, Revisited Gupte, Aparna; Liu, Jiahui; Raizes, Justin; Roberts, Bhaskar; Vaikuntanathan, Vinod One-time programs (Goldwasser, Kalai and Rothblum, CRYPTO 2008) are programs that can be run on any single input of a user’s choice, but not on a second input. Classically, they are unachievable without trusted hardware, but the destructive nature of quantum measurements seems to provide an alternate path to constructing them. Unfortunately, Broadbent, Gutoski and Stebila (CRYPTO 2013) showed that even with quantum techniques, a strong notion of one-time programs, similar to ideal obfuscation, cannot be achieved for any non-trivial quantum function. On the positive side, Ben-David and Sattath (Quantum, 2023) showed how to construct a quantum one-time program for a certain (probabilistic) digital signature scheme, under a weaker notion of one-time program security. There is a vast gap between achievable and provably impossible notions of one-time program security, and it is unclear what functionalities are one-time programmable and which are not, under the achievable notions of security. In this work, we present new, meaningful, yet achievable definitions of one-time program security for probabilistic classical functions. We show how to construct one time programs satisfying these definitions for all functions in the classical oracle model and for constrained pseudorandom functions in the plain model. Finally, we examine the limits of these notions: we show a class of functions which cannot be one-time programmed in the plain model, as well as a class of functions which appears to be highly random given a single query, but whose quantum one-time program leaks the entire function even in the oracle model. STOC ’25, Prague, Czechia

Learning the Closest Product State

2026-01-17T06:31:54Z

Learning the Closest Product State Bakshi, Ainesh; Bostanci, John; Kretschmer, William; Landau, Zeph; Li, Jerry; Liu, Allen; O'Donnell, Ryan; Tang, Ewin We study the problem of finding a product state with optimal fidelity to an unknown n-qubit quantum state ρ, given copies of ρ. This is a basic instance of a fundamental question in quantum learning: is it possible to efficiently learn a simple approximation to an arbitrary state? We give an algorithm which finds a product state with fidelity ε-close to optimal, using N = npoly(1/ε) copies of ρ and poly(N) classical overhead. We further show that estimating the optimal fidelity is NP-hard for error ε = 1/poly(n), showing that the error dependence cannot be significantly improved. For our algorithm, we build a carefully-defined cover over candidate product states, qubit by qubit, and then demonstrate that extending the cover can be reduced to approximate constrained polynomial optimization. For our proof of hardness, we give a formal reduction from polynomial optimization to finding the closest product state. Together, these results demonstrate a fundamental connection between these two seemingly unrelated questions. Building on our general approach, we also develop more efficient algorithms in three simpler settings: when the optimal fidelity exceeds 5/6; when we restrict ourselves to a discrete class of product states; and when we are allowed to output a matrix product state. STOC ’25, Prague, Czechia

Breaking the T^(2/3) Barrier for Sequential Calibration

2026-01-17T06:31:46Z

Breaking the T^(2/3) Barrier for Sequential Calibration Dagan, Yuval; Daskalakis, Constantinos; Fishelson, Maxwell; Golowich, Noah; Kleinberg, Robert; Okoroafor, Princewill A set of probabilistic forecasts is calibrated if each prediction of the forecaster closely approximates the empirical distribution of outcomes on the subset of timesteps where that prediction was made. We study the fundamental problem of online calibrated forecasting of binary sequences, which was initially studied by Foster and Vohra. They derived an algorithm with O(T2/3) calibration error after T time steps, and showed a lower bound of Ω(T1/2). These bounds remained stagnant for two decades, until Qiao and Valiant improved the lower bound to Ω(T0.528) by introducing a combinatorial game called sign preservation and showing that lower bounds for this game imply lower bounds for calibration. In this paper, we give the first improvement to the O(T2/3) upper bound on calibration error of Foster and Vohra. We do this by introducing a variant of Qiao and Valiant’s game that we call sign preservation with reuse (SPR). We prove that the relationship between SPR and calibrated forecasting is bidirectional: not only do lower bounds for SPR translate into lower bounds for calibration, but algorithms for SPR also translate into new algorithms for calibrated forecasting. We then give an improved upper bound for the SPR game, which implies, via our equivalence, a forecasting algorithm with calibration error O(T2/3 − ) for some > 0, improving Foster and Vohra’s upper bound for the first time. Using similar ideas, we then prove a slightly stronger lower bound than that of Qiao and Valiant, namely Ω(T0.54389). Our lower bound is obtained by an oblivious adversary, marking the first ω(T1/2) calibration lower bound for oblivious adversaries. STOC ’25, Prague, Czechia

PrEP attitudes, willingness, and preferences among men incarcerated in jail in Massachusetts

2026-01-17T06:32:28Z

PrEP attitudes, willingness, and preferences among men incarcerated in jail in Massachusetts Al Abosy, Jude; Kalavacherla, Sruthi; Koutoujian, Peter J.; Siddiqi, Kashif; Senst, Thomas; Caro, Jose; Grossman, Anna; Dong, Kimberly R. Background People who inject drugs (PWID) are both disproportionately incarcerated and affected by HIV infection. Systemic inequities perpetuate the cyclic nature of injection drug use (IDU) and incarceration, and both IDU and incarceration are linked to higher rates of HIV infection. Pre-exposure prophylaxis (PrEP) is highly effective in HIV prevention and is currently available as a daily oral pill. Longer-acting PrEP options, such as injectables and implants, are also in development to improve accessibility and adherence. Despite these advancements, PrEP uptake remains low among PWID and individuals recently released from jail, and there is limited literature exploring the preferences for PrEP uptake within this population. Methods We conducted qualitative interviews using a semi-structured interview guide with 20 male participants (19 incarcerated in a Massachusetts jail and 1 recently released) to assess perceived HIV risk, knowledge of PrEP, barriers to PrEP uptake, and preferences for PrEP modality and frequency. The data were analyzed using a directed content analysis approach. Results Most participants were aware of their HIV risk but were largely unaware of PrEP and had never been educated about PrEP by a healthcare provider. Participants cited a lack of access to healthcare, stigma around HIV infection, and feasibility as barriers to uptake. While participants expressed interest in longer-acting PrEP, most preferred the oral pill due to distrust of the safety and efficacy of injectables and implants, countering the assumption that modality changes alone can improve low PrEP uptake. Conclusions Our findings underscore the urgent need for targeted education and interventions to improve HIV prevention in vulnerable populations impacted by incarceration. While long-acting injectables have been touted to help address barriers to accessing healthcare among this population, skepticism about the efficacy of long-acting injectables among this population may prevent these efforts. It is important to further research the willingness to uptake PrEP and modality preferences among this population to meet their needs.

Search for a new scalar resonance decaying to a Higgs boson and another new scalar particle in the final state with two bottom quarks and two photons in proton-proton collisions at $$\sqrt{s}=13$$ TeV

2026-01-17T06:32:26Z

Search for a new scalar resonance decaying to a Higgs boson and another new scalar particle in the final state with two bottom quarks and two photons in proton-proton collisions at $$\sqrt{s}=13$$ TeV Hayrapetyan, A.; Makarenko, V.; Tumasyan, A.; Adam, W.; Andrejkovic, J. W.; Benato, L.; Bergauer, T.; Dragicevic, M.; Giordano, C.; Hussain, P. S.; Jeitler, M.; Krammer, N.; Li, A.; Liko, D.; Matthewman, M.; Mikulec, I. A search is presented for a new scalar resonance, X, decaying to a standard model Higgs boson and another new scalar particle, Y, in the final state where the Higgs boson decays to a $$\text{b}\overline{\text{b} }$$ pair, while the Y particle decays to a pair of photons. The search is performed in the mass range 240–1000 GeV for the resonance X, and in the mass range 70–800 GeV for the particle Y, using proton-proton collision data collected by the CMS experiment at $$\sqrt{s}=13$$ TeV, corresponding to an integrated luminosity of 132 fb−1. In general, the data are found to be compatible with the standard model expectation. Observed (expected) upper limits at 95% confidence level on the product of the production cross section and the relevant branching fraction are extracted for the X → YH process, and are found to be within the range of 0.05–2.69 (0.08–1.94) fb, depending on mX and mY. The most significant deviation from the background-only hypothesis is observed for X and Y masses of 300 and 77 GeV, respectively, with a local (global) significance of 3.33 (0.65) standard deviations.

Public Service Provision and the Virtuous Circle: Evidence from Malawi

2026-01-17T06:32:24Z

Public Service Provision and the Virtuous Circle: Evidence from Malawi Chen, Nuole; Grady, Christopher; Dulani, Boniface; Masumbu, Mwayi; Chiona, Busta; Bowers, Jake; Winters, Matthew S. Many governments struggle to obtain the resources they need to govern effectively. In the virtuous circle model of state development, tax revenue allows governments to provide public goods and services to citizens, and citizens comply with taxation when governments provide sufficient levels of goods and services. The model, however, also suggests a vicious version of the circle, where citizens do not pay taxes, governments lack revenue to provide public goods and services, and citizens therefore continue to not pay taxes. Under this suboptimal equilibrium, governments cannot deliver on their governing and service provision mandates. We study whether a shock to public service provision in a major city in Malawi can induce citizens to pay taxes, thereby shifting the relationship between the city and its citizens from a vicious circle to a virtuous circle. With a difference-in-differences-style analysis, we show that households exposed to new government-provided waste collection expressed more trust in and better perceptions of the local government. Most importantly, these households were more likely to make tax payments. We find that this increase in tax payments largely came from people paying more of what they owed rather than from new taxpayers entering the rolls.

NonlinearSolve.jl: High-Performance and Robust Solvers for Systems of Nonlinear Equations in Julia

2026-01-17T06:32:22Z

NonlinearSolve.jl: High-Performance and Robust Solvers for Systems of Nonlinear Equations in Julia Pal, Avik; Holtorf, Flemming; Larsson, Axel; Loman, Torkel; Rajput, Utkarsh; Sch?fer, Frank; Qu, Qingyu; Edelman, Alan; Rackauckas, Chris Efficiently solving nonlinear equations underpins numerous scientific and engineering disciplines, yet scaling these solutions for challenging system models remains a challenge. This paper presents NonlinearSolve.jl -- a suite of high-performance open-source nonlinear equation solvers implemented natively in the Julia programming language. NonlinearSolve.jl distinguishes itself by offering a unified API that accommodates a diverse range of solver specifications alongside features such as automatic algorithm selection based on runtime analysis, support for static array kernels for improved GPU computation on smaller problems, and the utilization of sparse automatic differentiation and Jacobian-free Krylov methods for large-scale problem-solving. Through rigorous comparison with established tools such as PETSc SNES, Sundials KINSOL, and MINPACK, NonlinearSolve.jl demonstrates robustness and efficiency, achieving significant advancements in solving nonlinear equations while being implemented in a high-level programming language. The capabilities of NonlinearSolve.jl unlock new potentials in modeling and simulation across various domains, making it a valuable addition to the computational toolkit of researchers and practitioners alike.

Property Testing with Online Adversaries

2026-01-17T06:32:20Z

Property Testing with Online Adversaries Ben Eliezer, Omri; Kelman, Esty; Meir, Uri; Raskhodnikova, Sofya The online manipulation-resilient testing model, proposed by Kalemaj, Raskhodnikova and Varma (Theory of Computing 2023), studies property testing in situations where access to the input degrades continuously and adversarially. Our main contributions are as follows: - An extension of the model, introducing \emph{batch queries} where multiple queries are made and answered between each round of manipulation, and \emph{fractional manipulation rate}, where the adversary makes less than one manipulation per round. - New optimal testers for linearity testing of Boolean functions in the original online and offline models. - A new lower-bound for testing low-degree of Boolean functions in the original model which can be overcome by an algorithm using batch queries. - Efficient testers for local properties of sequences when the manipulation rate is fractional. Specifically, for sortedness, we show a sharp transition from optimal query complexity to the impossibility of testability, depending on the manipulation rate.

Engineered yeast tolerance enables efficient production from toxified lignocellulosic feedstocks

2026-03-08T03:39:24Z

Engineered yeast tolerance enables efficient production from toxified lignocellulosic feedstocks Lam, Felix H; Turanlı-Yıldız, Burcu; Liu, Dany; Resch, Michael G; Fink, Gerald R; Stephanopoulos, Gregory Lignocellulosic biomass remains unharnessed for the production of renewable fuels and chemicals due to challenges in deconstruction and the toxicity its hydrolysates pose to fermentation microorganisms. Here, we show in Saccharomyces cerevisiae that engineered aldehyde reduction and elevated extracellular potassium and pH are sufficient to enable near-parity production between inhibitor-laden and inhibitor-free feedstocks. By specifically targeting the universal hydrolysate inhibitors, a single strain is enhanced to tolerate a broad diversity of highly toxified genuine feedstocks and consistently achieve industrial-scale titers (cellulosic ethanol of >100 grams per liter when toxified). Furthermore, a functionally orthogonal, lightweight design enables seamless transferability to existing metabolically engineered chassis strains: We endow full, multifeedstock tolerance on a xylose-consuming strain and one producing the biodegradable plastics precursor lactic acid. The demonstration of “drop-in” hydrolysate competence enables the potential of cost-effective, at-scale biomass utilization for cellulosic fuel and nonfuel products alike.

Removal of lycopene substrate inhibition enables high carotenoid productivity in Yarrowia lipolytica

2026-03-08T03:39:23Z

Removal of lycopene substrate inhibition enables high carotenoid productivity in Yarrowia lipolytica Ma, Yongshuo; Liu, Nian; Greisen, Per; Li, Jingbo; Qiao, Kangjian; Huang, Sanwen; Stephanopoulos, Gregory Substrate inhibition of enzymes can be a major obstacle to the production of valuable chemicals in engineered microorganisms. Here, we show substrate inhibition of lycopene cyclase as the main limitation in carotenoid biosynthesis in Yarrowia lipolytica. To overcome this bottleneck, we exploit two independent approaches. Structure-guided protein engineering yields a variant, Y27R, characterized by complete loss of substrate inhibition without reduction of enzymatic activity. Alternatively, establishing a geranylgeranyl pyrophosphate synthase-mediated flux flow restrictor also prevents the onset of substrate inhibition by diverting metabolic flux away from the inhibitory metabolite while maintaining sufficient flux towards product formation. Both approaches result in high levels of near-exclusive β-carotene production. Ultimately, we construct strains capable of producing 39.5 g/L β-carotene at a productivity of 0.165 g/L/h in bioreactor fermentations (a 1441-fold improvement over the initial strain). Our findings provide effective approaches for removing substrate inhibition in engineering pathways for efficient synthesis of natural products.

Isotope tracing in health and disease

2026-03-08T03:39:26Z

Isotope tracing in health and disease Dong, Wentao; Rawat, Eshaan S; Stephanopoulos, Gregory; Abu-Remaileh, Monther Biochemical characterization of metabolism provides molecular insights for understanding biology in health and disease. Over the past decades, metabolic perturbations have been implicated in cancer, neurodegeneration, and diabetes, among others. Isotope tracing is a technique that allows tracking of labeled atoms within metabolites through biochemical reactions. This technique has become an integral component of the contemporary metabolic research. Isotope tracing measures substrate contribution to downstream metabolites and indicates its utilization in cellular metabolic networks. In addition, isotopic labeling data are necessary for quantitative metabolic flux analysis. Here, we review recent work utilizing metabolic tracing to study health and disease, and highlight its application to interrogate subcellular, intercellular, and in vivo metabolism. We further discuss the current challenges and opportunities to expand the utility of isotope tracing to new research areas.

Engineering a universal and efficient platform for terpenoid synthesis in yeast

2026-03-08T03:39:25Z

Engineering a universal and efficient platform for terpenoid synthesis in yeast Ma, Yongshuo; Zu, Yuexuan; Huang, Sanwen; Stephanopoulos, Gregory Engineering microbes for the production of valuable natural products is often hindered by the regulation of native competing metabolic networks in host. This is particularly evident in the case of terpenoid synthesis in yeast, where the canonical terpenoid precursors are tightly coupled to the biosynthesis of sterols essential for yeast viability. One way to circumvent this limitation is by engineering product pathways less connected to the host native metabolism. Here, we introduce a two-step isopentenol utilization pathway (IUP) in Saccharomyces cerevisiae to augment the native mevalonate pathway by providing a shortcut to the synthesis of the common terpenoid precursors, isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP). As such, the IUP was capable of elevating the IPP/DMAPP pool by 147-fold compared with the native pathway. We further demonstrate that cofeeding isoprenol and prenol enhances geranyl diphosphate (GPP) content for monoterpene biosynthesis. More importantly, we established a synthetic three-step route for efficient synthesis of di-and tetraterpene precursor geranylgeranyl diphosphate (GGPP), circumventing the competition with farnesyl diphosphate (FPP) for sterol biosynthesis and elevating the GGPP level by 374-fold. We combine these IUP-supported precursor-forming platforms with downstream terpene synthases to harness their potential and improve the production of industrially relevant terpenoids by several fold. Our exploration provides a universal and effective platform for supporting terpenoid synthesis in yeast.

Oscillatory control of cortical space as a computational dimension

2026-01-16T03:07:45Z

Oscillatory control of cortical space as a computational dimension Chen, Zhen; Brincat, Scott L.; Lundqvist, Mikael; Loonis, Roman F.; Warden, Melissa R.; Miller, Earl K. Flexible cognition depends on the ability to represent and apply relevant information to the current task at hand. This allows the brain to interpret sensory input and guide behavior in a context-dependent manner. Recent work has proposed “spatial computing” as a mechanism for this flexibility, suggesting that task-related signals organize information processing through spatial patterns of oscillatory activity across the cortical surface. These patterns are proposed to act as “inhibitory stencils” that constrain where sensory-related information (the “content” of cognition) can be expressed in spiking activity. Here, we provide a comprehensive empirical test of spatial computing using multi-electrode recordings from the lateral prefrontal cortex in non-human primates performing a range of cognitive tasks (object working memory, sequence working memory, and categorization). We found that alpha/beta oscillations encoded task-related information, were organized into spatial patterns that changed with task conditions, and inversely correlated with the spatial expression of sensory-related spiking activity. Furthermore, we found that alpha/beta oscillations reflected misattributions of task conditions and correlated with subjects’ trial-by-trial decisions. These findings validate core predictions of spatial computing, suggesting that oscillatory dynamics not only gate information in time but also shape where in the cortex cognitive content is represented. This framework offers a unifying principle for understanding how the brain flexibly coordinates cognition through structured population dynamics.

Working memory readout varies with frontal theta rhythms

2026-01-16T03:07:46Z

Working memory readout varies with frontal theta rhythms Han, Hio-Been; Brincat, Scott L.; Buschman, Timothy J.; Miller, Earl K. Increasing evidence suggests that attention varies rhythmically, phase locked to ongoing cortical oscillations. Here, we report that the phase of theta oscillations (3–6 Hz) in the frontal eye field (FEF) is associated with the spatiotemporal variation of information readout from working memory (WM). Non-human primates were briefly shown a sample array of colored squares. A short time later, they viewed a test array and were rewarded for identifying which square changed color (the target). Behavioral performance varied systematically with theta phase at the time of test array onset, as well as with the target’s location. This is consistent with theta “scanning” across the FEF and thus visual space from top to bottom. Theta was coupled, on opposing phases, to both spiking and beta (12–20 Hz). These results could be explained by a wave of activity that moves across the FEF, modulating the readout of information from WM.

Nuclear Ship Safety Handbook

2026-01-21T15:07:41Z

Nuclear Ship Safety Handbook Valiaveedu, Anthony; Edmonds, Nat; Izurieta, Jose At present, there exists no clear, unified public document in the incorporation of design safety for nuclear civilian ships. Historically, there has been developed research into this area due to political development in the “Atoms for Peace” era. However, as of recent, the only development has been through standards institutions related to Floating Nuclear Power Plants (commonly known as FLOPPS) and by the Russian Federation with their nuclear icebreaker development. This paper uses this research data and standards and combines it with the operational experiences during civilian maritime nuclear operations to provide unique insights into potential issues and resolutions in the design efficacy of maritime nuclear operations. The goal, therefore, is to provide a strong basis for initial safety on key areas that require nuclear and maritime regulatory research and development in the coming years to prepare for nuclear propulsion in the maritime industry. The paper is isolated into multiple chapters in the areas that involve overlapping nuclear/maritime safety design decisions that will be encountered by engineers. Chapter 1 establishes the principles andm philosophy behind the safety discussion for nuclear maritime and discusses key topics that relate to the overall ship design. Chapter 2 provides design details on the reactor compartment and other considerations when designing the reactor compartment. Chapter 3 describes the various hazards the reactor plant should be resilient against and avenues in establishing resiliency. Chapter 4 discusses the propulsion system and key considerations when evaluating different propulsion designs. Chapter 5 provides emergency power considerations for design determinations. Chapter 6 provides an event tree analysis on the major initiating events when operating a nuclear ship. Chapter 7 outlines the port operating procedures including avenues for establishing porting requirements for nuclear ships. Contact information: Anthony Valiaveedu (arv7@mit.edu); Nat Edmonds (edmondsn@mit.edu)

Exploring the Emotional Effects of Enhanced Interoception via Heartbeat-Synchronized Haptic Feedback

2026-03-08T03:38:53Z

Exploring the Emotional Effects of Enhanced Interoception via Heartbeat-Synchronized Haptic Feedback Kim, Minsol; Whitmore, Nathan; Chua, Phoebe; Pei, Serena; Abdalla, Malak; Maes, Pattie This study examines how amplifying real-time heartbeat feedback affects emotion regulation. Accurate heartbeat perception—a key facet of cardiac interoception—has been linked to emotional awareness and mental well-being, yet the causal role of interoceptive feedback in emotion regulation remains underexplored. We empirically tested whether making heart rate signals more perceptible through wearable haptic feedback could facilitate implicit emotion regulation during emotionally evocative experiences. Using a custom Fitbit-based system, thirty participants received real-time, sham, or no heartbeat-synchronized vibrations while viewing fear- and amusement-inducing film clips. Interoceptive accuracy, emotional disturbance, and the linguistic complexity of emotion descriptions were measured. Exploratory analyses showed that real-time feedback reduced emotional disturbance during fear stimuli, especially among individuals attentive to bodily sensations, though effects did not remain significant after multiple comparisons correction. Feedback primarily modulated arousal rather than valence and did not significantly affect heartbeat counting or linguistic complexity. As one of the first causal, empirical investigations of interoceptive feedback and emotion regulation, this work identifies boundary conditions for its effectiveness and offers insights for designing personalized, interoception-aware wearable technologies.

EcoLearn: Optimizing the Carbon Footprint of Federated Learning

2026-03-08T03:39:04Z

EcoLearn: Optimizing the Carbon Footprint of Federated Learning Mehboob, Talha; Bashir, Noman; Iglesias, Jesus Oma?a; Zink, Michael; Irwin, David Federated Learning (FL) distributes machine learning (ML) training across edge devices to reduce data transfer overhead and protect data privacy. Since FL model training may span hundreds of devices and is thus resource- and energy-intensive, it has a significant carbon footprint. Importantly, since energy's carbon-intensity differs substantially (by up to 60×) across locations, training on the same device using the same amount of energy, but at different locations, can incur widely different carbon emissions. While prior work has focused on improving FL's resource- and energy-efficiency by optimizing time-to-accuracy, it implicitly assumes all energy has the same carbon intensity and thus does not optimize carbon efficiency, i.e., work done per unit of carbon emitted. To address the problem, we design EcoLearn, which minimizes FL's carbon footprint without significantly affecting model accuracy or training time. EcoLearn achieves a favorable tradeoff by integrating carbon awareness into multiple aspects of FL training, including i) selecting clients with high data utility and low carbon, ii) provisioning more clients during the initial training rounds, and iii) mitigating stragglers by dynamically adjusting client over-provisioning based on carbon. We implement EcoLearn and its carbon-aware FL training policies in the Flower framework and show that it reduces the carbon footprint of training (by up to 10.8×) while maintaining model accuracy and training time (within ~1%) compared to state-of-the-art approaches. SEC ’25, Arlington, VA, USA

Democratizing Multi-Granularity Spatio-Temporal Intelligence with Multi-Agent Systems

2026-03-08T03:39:00Z

Democratizing Multi-Granularity Spatio-Temporal Intelligence with Multi-Agent Systems Wu, Che-Cheng; Huang, Syuan-Bo; Song, Yu-Lun; Lin, Po-Han; Lin, Michael; Lin, Yu-Ta We propose a system that democratizes multi-granularity spatio-temporal analysis by integrating a Discrete Global Grid System (DGGS) data pipeline with a Multi-Agent System (MAS). Unlike existing single-agent spatial AI solutions that primarily target experts and lack support for heterogeneous data, persistent memory, and validation, our platform converts diverse datasets into standardized H3-indexed cells, enabling consistent analysis across scales. To enhance usability for non-experts, the system interactively guides users to refine queries, which are decomposed into sub-tasks managed by specialized agents for data retrieval, transformation, analysis, and visualization. Agents communicate through a decentralized framework with shared memory, supporting persistent reasoning and multi-turn dialogue. Reflection modules and human-in-the-loop validation further strengthen robustness. Demonstrated through real-world scenarios, such as analyzing the relationship between aging rate patterns and average income to inform social welfare policy in Taiwan, the system illustrates how natural language queries, combined with intuitive map- and chart-based visualizations, can support evidence-based decision-making. GeoGenAgent ’25, Minneapolis, MN, USA

One String to Pull Them All: Fast Assembly of Curved Structures from Flat Auxetic Linkages

2026-03-08T03:39:20Z

One String to Pull Them All: Fast Assembly of Curved Structures from Flat Auxetic Linkages Zaman, Akib; Aslarus, Jacqueline; Li, Jiaji; Mueller, Stefanie; Konakovic Lukovic, Mina We present a computational approach for designing freeform structures that can be rapidly assembled from initially flat configurations by a single string pull. The target structures are decomposed into rigid spatially varied quad tiles that are optimized to approximate the user-provided surface, forming a flat mechanical linkage. Our algorithm then uses a two-step method to find a physically realizable string path that controls only a subset of tiles to smoothly actuate the structure from flat to assembled configuration. We initially compute the minimal subset of tiles that are required to be controlled with the string considering the geometry of the structure and interaction among the tiles. We then find a valid string path through these tiles that minimizes friction, which will assemble the flat linkage into the target 3D structure upon tightening a single string. The resulting designs can be easily manufactured with computational fabrication techniques such as 3D printing, CNC milling, molding, etc. in flat configuration that, in addition to manufacturing, facilitates storage and transportation. We validate our approach by developing a series of physical prototypes and showcasing various application case studies, ranging from medical devices, space shelters, to architectural designs.

Discovering Folding Lines for Surface Compression

2026-03-08T03:39:22Z

Discovering Folding Lines for Surface Compression Aoki, Toshiki; Tachi, Tomohiro; Konakovic Lukovic, Mina The miniaturization of shell structures presents a versatile and complex challenge, bridging geometry with diverse practical applications. In this paper, we introduce a novel approach for computing origami crease patterns to compress arbitrary 3D shell objects. First, we employ the adapted Material Point Method (MPM) to simulate the compression of a target surface and obtain an initial folded configuration. Since MPM produces overly smooth curved surfaces, their crease patterns are unsuitable for practical origami fabrication. We then propose a novel Folding Line Extraction (FLE) method that optimizes these smoothed surfaces to extract folding lines that achieve the target compression with minimal deformation and stretching outside the crease lines. This method produces smooth curved folding lines. Fabrication and experimental validation of the extracted patterns demonstrate their effectiveness and applicability in real-world scenarios. SA Conference Papers ’25, Hong Kong, Hong Kong

3DPR: Single Image 3D Portrait Relighting with Generative Priors

2026-03-08T03:39:16Z

3DPR: Single Image 3D Portrait Relighting with Generative Priors Rao, Pramod; Meka, Abhimitra; Zhou, Xilong; Fox, Gereon; B R, Mallikarjun; Zhan, Fangneng; Weyrich, Tim; Bickel, Bernd; Pfister, Hanspeter; Matusik, Wojciech; Beeler, Thabo; Elgharib, Mohamed; Habermann, Marc; Theobalt, Christian Rendering novel, relit views of a human head, given a monocular portrait image as input, is an inherently underconstrained problem. The traditional graphics solution is to explicitly decompose the input image into geometry, material and lighting via differentiable rendering; but this is constrained by the multiple assumptions and approximations of the underlying models and parameterizations of these scene components. We propose 3DPR, an image-based relighting model that leverages generative priors learnt from multi-view One-Light-at-A-Time (OLAT) images captured in a light stage. We introduce a new diverse and large-scale multi-view 4K OLAT dataset of 139 subjects to learn a high-quality prior over the distribution of high-frequency face reflectance. We leverage the latent space of a pre-trained generative head model that provides a rich prior over face geometry learnt from in-the-wild image datasets. The input portrait is first embedded in the latent manifold of such a model through an encoder-based inversion process. Then a novel triplane-based reflectance network trained on our lightstage data is used to synthesize high-fidelity OLAT images to enable image-based relighting. Our reflectance network operates in the latent space of the generative head model, crucially enabling a relatively small number of lightstage images to train the reflectance model. Combining the generated OLATs according to a given HDRI environment maps yields physically accurate environmental relighting results. Through quantitative and qualitative evaluations, we demonstrate that 3DPR outperforms previous methods, particularly in preserving identity and in capturing lighting effects such as specularities, self-shadows, and subsurface scattering. SA Conference Papers ’25, December 15–18, 2025, Hong Kong, Hong Kong

Shoot-Bounce-3D: Single-Shot Occlusion-Aware 3D from Lidar by Decomposing Two-Bounce Light

2026-03-08T03:39:03Z

Shoot-Bounce-3D: Single-Shot Occlusion-Aware 3D from Lidar by Decomposing Two-Bounce Light Klinghoffer, Tzofi; Somasundaram, Siddharth; Xiang, Xiaoyu; Fan, Yuchen; Richardt, Christian; Dave, Akshat; Raskar, Ramesh; Ranjan, Rakesh 3D scene reconstruction from a single measurement is challenging, especially in the presence of occluded regions and specular materials, such as mirrors. We address these challenges by leveraging single-photon lidars. These lidars estimate depth from light that is emitted into the scene and reflected directly back to the sensor. However, they can also measure light that bounces multiple times in the scene before reaching the sensor. This multi-bounce light contains additional information that can be used to recover dense depth, occluded geometry, and material properties. Prior work with single-photon lidar, however, has only demonstrated these use cases when a laser sequentially illuminates one scene point at a time. We instead focus on the more practical – and challenging – scenario of illuminating multiple scene points simultaneously. The complexity of light transport due to the combined effects of multiplexed illumination, two-bounce light, shadows, and specular reflections is challenging to invert analytically. Instead, we propose a data-driven method to invert light transport in single-photon lidar. To enable this approach, we create the first large-scale simulated dataset of ~100k lidar transients for indoor scenes. We use this dataset to learn a prior on complex light transport, enabling measured two-bounce light to be decomposed into the constituent contributions from each laser spot. Finally, we experimentally demonstrate how this decomposed light can be used to infer 3D geometry in scenes with occlusions and mirrors from a single measurement. Our code and dataset are released on our project webpage. SA Conference Papers ’25, Hong Kong, Hong Kong

PhysiOpt: Physics-Driven Shape Optimization for 3D Generative Models

2026-03-08T03:39:00Z

PhysiOpt: Physics-Driven Shape Optimization for 3D Generative Models Zhan, Xiao; Jambon, Cl?ment; Thompson, Evan; Ng, Kenney; Konakovi? Lukovi?, Mina Generative models have recently demonstrated impressive capabilities in producing high-quality 3D shapes from a variety of user inputs (e.g., text or images). However, generated objects often lack physical integrity. We introduce PhysiOpt, a differentiable physics optimizer designed to improve the physical behavior of 3D generative outputs, enabling them to transition from virtual designs to physically plausible, real-world objects. While most generative models represent geometry as continuous implicit fields, physics-based approaches often rely on the finite element method (FEM), requiring ad hoc mesh extraction to perform shape optimization. In addition, these methods are typically slow, limiting their integration in fast, iterative generative design workflows. Instead, we bridge the representation gap and propose a fast and effective differentiable simulation pipeline that optimizes shapes directly in the latent space of generative models using an intuitive and easy-to-implement differentiable mapping. This approach enables fast optimization while preserving semantic structure, unlike traditional methods relying on local mesh-based adjustments. We demonstrate the versatility of our optimizer across a range of shape priors, from global and part-based latent models to a state-of-the-art large-scale 3D generator, and compare it to a traditional mesh-based shape optimizer. Our method preserves the native representation and capabilities of the underlying generative model while supporting user-specified materials, loads, and boundary conditions. The resulting designs exhibit improved physical behavior, remain faithful to the learned priors, and are suitable for fabrication. We demonstrate the effectiveness of our approach on both virtual and fabricated objects. SA Conference Papers ’25, Hong Kong, Hong Kong

Low-Rank Adaptation of Neural Fields

2026-03-08T03:39:13Z

Low-Rank Adaptation of Neural Fields Truong, Anh; Mahmoud, Ahmed; Konakovi? Lukovi?, Mina; Solomon, Justin Processing visual data often involves small adjustments or sequences of changes, e.g., image filtering, surface smoothing, and animation. While established graphics techniques like normal mapping and video compression exploit redundancy to encode such small changes efficiently, the problem of encoding small changes to neural fields—neural network parameterizations of visual or physical functions—has received less attention. We propose a parameter-efficient strategy for updating neural fields using low-rank adaptations (LoRA). LoRA, a method from the parameter-efficient fine-tuning LLM community, encodes small updates to pre-trained models with minimal computational overhead. We adapt LoRA for instance-specific neural fields, avoiding the need for large pre-trained models and yielding lightweight updates. We validate our approach with experiments in image filtering, geometry editing, video compression, and energy-based editing, demonstrating its effectiveness and versatility for representing neural field updates. Anh Truong, Ahmed H. Mahmoud, Mina Konaković Luković, and Justin Solomon. 2025. Low-Rank Adaptation of Neural Fields. In Proceedings of the SIGGRAPH Asia 2025 Conference Papers (SA Conference Papers '25). Association for Computing Machinery, New York, NY, USA, Article 86, 1–12.

Participatory Evolution of Artificial Life Systems via Semantic Feedback

2026-03-08T03:39:34Z

Participatory Evolution of Artificial Life Systems via Semantic Feedback Li, Shuowen; Wang, Kexin; Fang, Minglu; Huang, Danqi; Asadipour, Ali; Mi, Haipeng; Sun, Yitong We present a semantic-feedback framework that treats natural language as a regulatory signal for evolving artificial-life systems. Instead of using prompts to select finished images, text in our system shapes the dynamics of an interactive ecosystem, allowing audiences to cultivate behaviors over time. The framework couples a learned mapping from prompts to simulation parameters with evolutionary search and vision–language evaluation, so user intent modulates both visible outcomes and the underlying generative rules. It supports iterative prompt refinement, multi-agent interaction, and the synthesis of new collective rules from community input. In a user study, participants achieved higher semantic alignment and reported a greater sense of control than with manual tuning, while behaviors remained diverse across generations. As an art-led contribution, the work reframes authoring as participatory cultivation and advances open-ended evolution as a socially distributed, not solely algorithmic, process; as a tool contribution, it offers a practical platform for co-creative generative design. SA Art Papers ’25, Hong Kong, Hong Kong

Physical Manifestation of Generative AI Music Systems for Live Performance

2026-03-08T03:39:33Z

Physical Manifestation of Generative AI Music Systems for Live Performance Naseck, Perry; Blanchard, Lancelot; Lavakare, Madhav; Lecamwasam, Kimaya; Paradiso, Joseph This paper explores the physical manifestation of generative AI music systems for live performance, focusing on bridging the expressive gap between AI-generated music and audience perception. Through a year-long collaboration with a human performer, we constructed a kinetic sculpture that visualizes the outputs of an AI jam_bot during concerts. The sculpture, powered by ML-based and pattern-driven mapping methodologies, interprets real-time AI musical decisions as expressive movements. Audience feedback indicates increased engagement and curiosity, although interpretability remains a challenge. Our work highlights the potential of embodied visualization to establish communicative presence for AI performers and suggests avenues for future research. SA Art Papers ’25, Hong Kong, Hong Kong

Performant Unified GPU Kernels for Portable Singular Value Computation Across Hardware and Precision

2026-03-08T03:39:31Z

Performant Unified GPU Kernels for Portable Singular Value Computation Across Hardware and Precision Ringoot, Evelyne; Alomairy, Rabab; Churavy, Valentin; Edelman, Alan This paper presents a portable, GPU-accelerated implementation of a QR-based singular value computation algorithm in Julia. The singular value decomposition (SVD) is a fundamental numerical tool in scientific computing and machine learning, providing optimal low-rank matrix approximations. Its importance has increased even more in large-scale machine learning pipelines, including large language models (LLMs), where it enables low-rank adaptation (LoRA). The implemented algorithm is based on the classic two-stage QR reduction, consisting of successive matrix reduction to band form and bidiagonal form. Our implementation leverages Julia’s multiple dispatch and metaprogramming capabilities, integrating with the GPUArrays and KernelAbstractions frameworks to provide a unified type and hardware-agnostic function. It supports diverse GPU architectures and data types, and is, to our knowledge, the first GPU-accelerated singular value implementation to support Apple Metal GPUs and half precision. Performance results on multiple GPU backends and data types demonstrate that portability does not require sacrificing performance: the unified function outperforms most linear algebra libraries (MAGMA, SLATE, rocSOLVER, oneMKL) for matrix sizes larger than 1024 × 1024, and achieves 80%-90% of the performance of cuSOLVER for large matrices. ICPP ’25, San Diego, CA, USA

UQGNN: Uncertainty Quantification of Graph Neural Networks for Multivariate Spatiotemporal Prediction

2026-03-08T03:39:29Z

UQGNN: Uncertainty Quantification of Graph Neural Networks for Multivariate Spatiotemporal Prediction Yu, Dahai; Zhuang, Dingyi; Jiang, Lin; Xu, Rongchao; Ye, Xinyue; Bu, Yuheng; Wang, Shenhao; Wang, Guang Spatiotemporal prediction plays a critical role in numerous real-world applications such as urban planning, transportation optimization, disaster response, and pandemic control. In recent years, researchers have made significant progress by developing advanced deep learning models for spatiotemporal prediction. However, most existing models are deterministic, i.e., predicting only the expected mean values without quantifying uncertainty, leading to potentially unreliable and inaccurate outcomes. While recent studies have introduced probabilistic models to quantify uncertainty, they typically focus on a single phenomenon (e.g., taxi, bike, crime, or traffic crashes), thereby neglecting the inherent correlations among heterogeneous urban phenomena. To address the research gap, we propose a novel Graph Neural Network with Uncertainty Quantification, termed UQGNN for multivariate spatiotemporal prediction. UQGNN introduces two key innovations: (i) an Interaction-aware Spatiotemporal Embedding Module that integrates a multivariate diffusion graph convolutional network and an interaction-aware temporal convolutional network to effectively capture complex spatial and temporal interaction patterns, and (ii) a multivariate probabilistic prediction module designed to estimate both expected mean values and associated uncertainties. Extensive experiments on four real-world multivariate spatiotemporal datasets from Shenzhen, New York City, and Chicago demonstrate that UQGNN consistently outperforms state-of-the-art baselines in both prediction accuracy and uncertainty quantification. For example, on the Shenzhen dataset, UQGNN achieves a 5% improvement in both prediction accuracy and uncertainty quantification. SIGSPATIAL ’25, Minneapolis, MN, USA

SONAR Web: A Platform-Agnostic Framework for Real-Time Decentralized Learning Across Heterogeneous Edge Clients

2026-03-08T03:39:37Z

SONAR Web: A Platform-Agnostic Framework for Real-Time Decentralized Learning Across Heterogeneous Edge Clients Yuan, Joyce; Le, Brian; Le, Kathryn; Shi, Yichuan; Singh, Abhishek; Sharma, Rishi; Patricio, Angel; Raskar, Ramesh Most federated learning (FL) frameworks assume reliable networks and homogeneous devices, limiting their applicability in mobile and edge environments where connectivity is intermittent and devices are highly heterogeneous. We introduce SONAR Web, an open-source framework for fully decentralized, cross-platform collaborative learning between browsers, servers, tablets, and smartphones. SONAR Web decouples the learning protocol from the underlying client platform through a platform-agnostic configuration interface—enabling Python, JavaScript, and mobile clients to seamlessly interoperate in real time. By combining peer-to-peer RTC protocols with communication-efficient techniques from FL, SONAR Web supports privacy-preserving training without centralized orchestration. We demonstrate SONAR Web's robustness through deployments on real-world devices and networks, showing resilience under heterogeneous network conditions and resource variability. SONAR Web provides a unified, language-agnostic interface for decentralized learning, enabling seamless collaboration across heterogeneous devices and runtimes—advancing scalable, inclusive, and real-time model training at the mobile and edge frontier. FLEdge-AI ’25, November 4-8, 2025, Hong Kong, China

Ferrozuit: Ferromagnetic Electronic Textile System for Zero-Gravity Spatial Anchoring

2026-03-08T03:39:36Z

Ferrozuit: Ferromagnetic Electronic Textile System for Zero-Gravity Spatial Anchoring Honnet, Cedric; Freire, Rachel; Cherston, Juliana; Guenther, Maximilian; Paradiso, Joseph; Wicaksono, Irmandy Long-duration human space missions introduce persistent physical, physiological, and psychological challenges stemming from the absence of gravity. Beyond major concerns like bone deterioration, cardiovascular deconditioning, and muscle atrophy, astronauts frequently experience spatial disorientation, discomfort during routine tasks, and difficulty maintaining stable body positioning. These subtle yet pervasive issues impact daily functioning, underscoring the need for lightweight, unobtrusive solutions that support orientation, comfort, and stability in microgravity environments. Ferrozuit introduces a solution to address these challenges in microgravity. It is a prototype crafted from custom ferromagnetic thread, woven and tailored to interact with programmable (electro)permanent magnets embedded within the microgravity environment. This system aims to provide an anchoring force intended to improve stability during tasks, enhance comfort during rest, and create a sense of orientation. This paper details the design rationale, the fabrication of the ferromagnetic textile, the magnetic docking system, initial technical evaluations, and potential applications. Ferrozuit reimagines spatial anchoring as an embedded, textile-driven experience, blending textile craft with advanced materials for adaptive wearable anchoring in microgravity environments. UbiComp Companion ’25, Espoo, Finland

Intelligent Soft Wearables

2026-03-08T03:39:35Z

Intelligent Soft Wearables Yu, Tianhong; Honnet, Cedric; Cheng, Tingyu; Takahashi, Ryo; Zhou, Bo; Zhang, Cheng; Lukowicz, Paul; Kawahara, Yoshihiro; Hester, Josiah; Paradiso, Joseph; Luo, Yiyue; Wicaksono, Irmandy Human bodies are almost always in contact with soft materials like clothing, for warmth, protection, self-expression, etc. Recent advancements in intelligent soft wearables have augmented these on-body soft objects with computational functions and intelligence with little compromise on the softness and comforts of wearables, allowing prolonged wear. These innovations, which combine advanced soft sensor design, fabrication, and computational power, offer unprecedented opportunities to improve our health, productivity, and overall well-being with monitoring and assistive capabilities. However, the inherent physical properties of soft materials present unique challenges in achieving practical interactions. The complexity of intelligent soft wearables, multiplexing intricate designs, soft materials, flexible electronics, advanced signal processing algorithms, and machine learning models, necessitates collaborative efforts from experts across diverse domains. This workshop aims to bring together interested researchers and practitioners across relevant domains to discuss the challenges and opportunities of intelligent soft wearables.

Quantum dots: A journey from fundamental discovery to technological impacts

2026-01-13T04:55:00Z

Quantum dots: A journey from fundamental discovery to technological impacts Hassan, Abeera; Kaur, Jaspreet; Chen, Ou; Bawendi, Moungi G. This article traces the evolution of quantum dots (QDs) from their initial discovery to growing technological impacts. We highlight the key breakthroughs in the development of colloidal QDs that have enabled precise control over their unique optical and optoelectronic properties. We also discuss a range of QD-based applications and address commercialization efforts. Finally, we examine ongoing challenges and emerging opportunities that are set to shape the future of QD research and technological advancement.

Proximity-labeling proteomics reveals remodeled interactomes and altered localization of pathogenic SHP2 variants

2026-03-08T03:39:31Z

Proximity-labeling proteomics reveals remodeled interactomes and altered localization of pathogenic SHP2 variants van Vlimmeren, Anne E.; Tang, Lauren C.; Jiang, Ziyuan; Iyer, Abhishek; Voleti, Rashmi; Krismer, Konstantin; Gaublomme, Jellert T.; Jovanovic, Marko; Shah, Neel H. Missense mutations in PTPN11, which encodes the protein tyrosine phosphatase SHP2, are common in several developmental disorders and cancers. While many mutations disrupt auto-inhibition and hyperactivate SHP2, several do not enhance catalytic activity. Both activating and non-activating mutations could potentially drive pathogenic signaling by altering SHP2 interactions or localization. We employed proximity-labeling proteomics to map the interaction networks of wild-type SHP2, ten clinically relevant mutants, and SHP2 bound to an inhibitor that stabilizes its auto-inhibited state. Our analyses reveal mutation- and inhibitor-dependent alterations in the SHP2 interactome, with several mutations also changing localization. Some mutants show increased mitochondrial localization and impact mitochondrial function. This study provides a resource for exploring SHP2 signaling and offers new insights into the molecular basis of SHP2-driven diseases. Furthermore, this work highlights the capacity for proximity-labeling proteomics to detect missense-mutation-dependent changes in protein interactions and localization.

Perspective on patient and non-academic partner engagement for the responsible integration of large language models in health chatbots

2026-03-08T03:39:30Z

Perspective on patient and non-academic partner engagement for the responsible integration of large language models in health chatbots Jaiswal, Nikhil; Ma, Yuanchao; Lebouché, Bertrand; Poenaru, Dan; Pomey, Marie-Pascale; Achiche, Sofiane; Lessard, David; Engler, Kim; Montiel, Zully; Acevedo, Hector; Gameiro, Rodrigo R.; Celi, Leo A.; Osmanlliu, Esli Uses of large language models (LLMs) in health chatbots are expanding into high-stakes clinical contexts, heightening the need for tools that are evidence-based, accountable, accurate, and patient-centred. This conceptual, practice-informed Perspective reflects on engaging patients and non-academic partners for the responsible integration of LLMs, grounded in the co-construction of MARVIN (for people living with HIV) and in an emerging collaboration with MIT Critical Data. Organised by the Software Development Life Cycle, we describe: conception/needs assessment with patient partners to identify use cases, acceptable trade-offs, and privacy expectations; development that prioritises grounding via vetted sources, structured human feedback, and data-validation committees including patient partners; testing and evaluation using patient-reported outcome measures (PROMs) and patient-reported experience measures (PREMs) chosen in collaboration with patients to capture usability, acceptability, trust, and perceived safety, alongside task performance and harmful-output monitoring; and implementation via diverse governance boards, knowledge-mobilisation materials to set expectations, and risk-management pathways for potentially unsafe outputs. Based on our experience with MARVIN, we recommend early and continuous engagement of patients and non-academic partners, fair compensation, shared decision-making power, transparent decision logging, and inclusive, adaptable governance that can evolve with changing models and standards. These lessons highlight how patient partnership can directly shape chatbot design and oversight, helping teams align LLM-enabled tools with patient-centred goals while building accountable, safe, and equitable systems.

Study of charm mixing and CP violation with D0 → K±π∓π±π∓ decays

2026-03-08T03:39:28Z

Study of charm mixing and CP violation with D0 → K±π∓π±π∓ decays Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J. A study of charm mixing and CP violation in D0 → K±π∓π±π∓ decays is performed using data collected by the LHCb experiment in proton-proton collisions from 2015 to 2018, corresponding to an integrated luminosity of 6 fb−1. The ratio of promptly produced D0 → K+π−π+π− to D0 → K−π+π−π+ decay rates is measured as a function of D0 decay time, both inclusive over phase space and in bins of phase space. Taking external inputs for the D 0 − D ¯ 0 mixing parameters x and y allows constraints to be obtained on the hadronic parameters of the charm decay. When combined with previous measurements from charm-threshold experiments and at LHCb, improved knowledge is obtained for these parameters, which is valuable for studies of the angle γ of the Unitarity Triangle. An alternative analysis is also performed, in which external inputs are taken for the hadronic parameters, and the mixing parameters are determined, including ∆x and ∆y, which are nonzero in the presence of CP violation. It is found that x = 0 . 85 − 0.24 + 0.15 % , y = 0 . 21 − 0.27 + 0.29 % , ∆x = (−0.02 ± 0.04) % and Δ y = 0.0 2 − 0.03 + 0.04 % . These results are consistent with previous measurements and the hypothesis of CP conservation.

Energy-energy correlator at hadron colliders: celestial blocks and singularities

2026-03-08T03:39:01Z

Energy-energy correlator at hadron colliders: celestial blocks and singularities Chen, Hao; Ruan, Hongyi; Zhu, Hua X. Energy-energy correlator (EEC) is an event shape observable that characterizes the distribution of energy flux in collision events. We initiate the study of full-range EEC at hadron colliders, generalizing the extensively studied EEC in e+e− collision as well as the transverse EEC in hadron collisions. We derive celestial blocks from Lorentz symmetry to perform partial wave decomposition of the EEC at hadron colliders. These celestial blocks are essentially conformal blocks on the 2d celestial sphere, which have additional dependence on the collinear spin of “light-ray transition matrix” along the collision axis. In this work, we perform the leading-order (LO) analytic calculation of this observable in pure Yang-Mills theory and use it as an example to illustrate the block decomposition. Numerically, the block expansion demonstrates superior accuracy in the collinear limit compared to conventional power series expansion. Analytically, we observe in this example that the block coefficients exhibit analyticity in both collinear and transverse spin. In addition, we analyze several kinematic limits at LO — collinear, back-to-back, opposite coplanar and Regge limit. While the first three limits naturally generalize their e+e− collision counterparts or transverse EEC and are governed by soft-collinear dynamics, the Regge limit requires complete angular dependence and reveals BFKL physics. Phenomenologically, we propose a realistic experimental setup and briefly discuss how the convolution of parton distribution function modifies the perturbative EEC result. Our work suggests that the full-range EEC at hadron colliders is an elegant observable which probes a broader kinematic space and connects various regimes of different QCD dynamics through a single measurement.

Synthesis and Applications of Large-Area Monolayer Graphene

2026-01-13T03:36:27Z

Synthesis and Applications of Large-Area Monolayer Graphene Wang, Zhien (Abigail) Graphene, renowned for its exceptional electrical, mechanical, and chemical properties, is a promising candidate for next-generation electronics, photonics, and biosensing. However, realizing its full potential depends critically on the ability to synthesize high-quality monolayer graphene. In this thesis, we present a robust chemical vapor deposition (CVD) approach for synthesizing large-area, adlayer-free, single-orientation graphene on Cu(111) foil and Cu(111) film/sapphire. A comparative analysis between these two substrates reveals critical differences in wrinkle density, grain size, and strain — offering insights for optimizing graphene growth. We further identify and characterize defective merging behavior in single-orientation graphene domains. Contrary to conventional assumptions, these merging regions contain permeable defects, revealing previously unrecognized limitations in using single-orientation stitched graphene as an impermeable barrier. To scale up production while reducing human error, we also develop an autonomous CVD platform with automated sample handling, growth and post-growth oxidation. This system enables high-throughput and reproducible graphene synthesis with minimal supervision. Building on these synthesis advances, we explore multiple applications of large-area monolayer graphene. We discover that graphene can promote interfacial oxidation of metals like aluminum and titanium during deposition, whereas metals such as nickel remain stable — a finding that informs the engineering of metal-graphene interfaces for electronic devices. In parallel, we explored diverse applications of graphene, including its role as a transparent, flexible electrode in organic solar cells, along with several collaborative efforts demonstrating its use as a sensor for cardiac microtissues, and as a tunable microheater in mid-infrared devices. Altogether, this work advances both the fundamental understanding and technological scalability of monolayer graphene, positioning it as a versatile platform for future applications across electronics, optoelectronics, and biointerfaces.

Mapping signaling networks and rapidly evolving genes in the developing Arabidopsis seed at single-nucleus resolution

2026-01-13T03:36:58Z

Mapping signaling networks and rapidly evolving genes in the developing Arabidopsis seed at single-nucleus resolution Martin, Caroline A. Seeds are an exceptional evolutionary innovation that enables the conditional allocation of maternal resources to successfully fertilized ovules. During early development, seeds accumulate nutrients that are utilized either by the embryo or by humans who harvest seed crops for food, biofuel, and livestock feed. Moreover, the grains of maize, rice, and wheat provide approximately 60% of the calories consumed worldwide. Although seeds are a cornerstone for ecosystems and modern agriculture, fundamental aspects of their development are incompletely understood. In this thesis, I develop a transcriptional atlas of seed development using the model plant Arabidopsis thaliana to clarify the functional compartmentalization, diversity, and developmental dynamics of cell types in the seed. I focus my analyses on how seed cell types communicate with one another to ensure successful propagation, and how genetic conflicts in the seed may drive rapid evolution in specific cell types. After characterizing the extent of short, secreted peptide expression in specific seed cell types, I perform in silico screens to match potential peptide hormones with their receptors. In total, I show that the seed coat shows functional compartmentalization around the gateway for maternal resources into seeds, that seed genes differentially expressed in a maternal resource transfer structure are rapidly evolving, and that genes underlying brassinosteroid biosynthesis and response are expressed in adjacent tissues, among other findings. This thesis illuminates potentially new mechanisms for inter-tissue coordination and provides a transcriptional reference for future seed studies.

Optimization in Deep Learning: Structured, Realistic and Interpretable Learning for Decision-Making

2026-01-13T03:36:43Z

Optimization in Deep Learning: Structured, Realistic and Interpretable Learning for Decision-Making Tsiourvas, Asterios In recent years, deep learning has emerged as a powerful tool for data-driven decisionmaking. However, its adoption in high-stakes applications is often constrained by challenges related to interpretability, fairness, and generalization in structured or complex environments. This thesis develops new optimization methodologies to enhance the realism, structureawareness, and interpretability of deep learning models in decision-making tasks. We begin, in Chapter 2, by addressing the challenge of optimizing trained neural networks for data-driven decision-making. Although neural networks can encode rich representations of preferences or outcomes, directly optimizing their outputs can be computationally intractable and often may produce unrealistic prescriptions. We introduce scalable algorithms that leverage the piecewise-linear structure of ReLU networks, reducing the original hard-to-solve mixed-integer program into tractable linear programs. To ensure realism, we introduce constraints that restrict decisions to lie on the data manifold. We then extend this framework to any differentiable neural network or MIP-expressible model and show that it scales for networks with millions of parameters. In Chapter 3, we focus on decision-making under observational data. First, we study personalized treatment recommendations under discrete treatments. We introduce the Prescriptive ReLU (P-ReLU) network, a piecewise-linear model that partitions the input space into polyhedral regions, assigning treatments uniformly within each, and that can be translated into an equivalent interpretable decision tree. We demonstrate that P-ReLU achieves strong prescriptive accuracy and accommodates structural/prescriptive constraints with ease. Next, we consider the problem of large language model (LLM) routing, where a query must be dynamically routed to the best model under competing metrics like accuracy and cost. We develop a causal, end-to-end approach that learns routing policies directly from logged observational data, minimizing directly decision-making regret. Finally, we tackle the problem of generating realistic, manifold-aligned counterfactual explanations. To address this problem, we present a MIP formulation where we explicitly enforce manifold alignment by reformulating the highly nonlinear Local Outlier Factor (LOF) metric as a set of mixed-integer constraints. To address the computational challenge, we leverage the geometry of the network and propose an efficient decomposition scheme that reduces the initial hard-to-solve problem into a series of significantly smaller, easier-to-solve problems. We further extend this framework to any differentiable neural network or MIP-expressible machine learning model. In Chapter 4, we focus on structured machine learning. We first address the problem of hierarchical time series forecasting, where predictions must be both accurate and consistent with the aggregation structure of the hierarchy. While prior methods rely on fixed projection matrices, we propose learning the optimal oblique projection directly from data. The proposed end-to-end approach jointly trains the forecasting model and projection layer, significantly improving accuracy and coherence. Next, we study the problem of creating a highly expressive, interpretable, and fair machine learning model. We propose Neural-Informed Decision Trees (NIDTs), a model that combines the predictive power of neural networks with the inherent interpretability of decision trees. NIDTs use axis-aligned splits on dataset features to form transparent decision paths, and at each leaf, apply a linear predictor based on both the original features and neural embeddings from a task-specific network to capture non-linearities. To generate NIDTs, we develop a decomposition training scheme that supports direct integration of fairness constraints via a constrained convex optimization problem solved at each leaf. Finally, in Chapter 5, we address fairness and efficiency in emergency department (ED) operations, where prolonged length of stay (LOS) has been linked to adverse outcomes such as increased mortality and higher risk of hospital-acquired infections. We focus on the patient prioritization and placement aspects of ED operations to improve throughput and reduce wait times. We propose a novel MIP predictive-prescriptive framework that decomposes predicted LOS into actionable components, enabling a more granular and operationally meaningful model of ED dynamics. Fairness considerations are explicitly incorporated into the formulation. To address uncertainty, we introduce a sampling-based solution approach. Our method increases ED throughput by 50–100% and reduces average wait time by 50–75%, depending on current utilization levels, while achieving near-optimal performance compared to a clairvoyant oracle. This work was conducted in collaboration with a major U.S. academic medical center. To facilitate practical implementation, we also design an interpretable metamodel that approximates the predictive-prescriptive algorithm with high fidelity. Together, these contributions provide a unified perspective on deep learning for reliable decision-making, grounded in optimization and encompassing interpretability, structure-awareness, and causal reasoning, well-suited for high-stakes operational environments.

Reverberation Mapping of Supermassive Black Holes using Machine Learning

2026-01-13T03:36:38Z

Reverberation Mapping of Supermassive Black Holes using Machine Learning Lewin, Collin Accreting supermassive black holes at the centers of galaxies, known as active galactic nuclei (AGN), offer a unique window into the physics of accretion and feedback that shape galactic evolution. Yet, the small spatial scales of these regions remain inaccessible to direct imaging. Reverberation mapping circumvents this limitation by using time delays between correlated emission at different wavelengths to infer physical size scales. While X-ray reverberation probes the innermost accretion flow, continuum reverberation in the UV, optical, and infrared (UVOIR) traces reprocessing by the accretion disk and broad-line region (BLR). In this thesis, I develop and apply frequency-domain timing techniques based on Gaussian Process (GP) regression to study AGN reverberation across X-ray and UVOIR regimes. By modeling the empirical variability of AGN light curves with GPs, I interpolate onto an evenly sampled time grid, enabling robust estimation of Fourier-resolved time lags despite irregular sampling or large time gaps. I apply this method to NuSTAR observations of the Narrow-line Seyfert 1 galaxy Ark 564, introducing a multi-task GP model that jointly learns kernel hyperparameters across light curves. This enables the first simultaneous modeling of lag and flux spectra from both NuSTAR and XMM-Newton using a relativistic reverberation model to constrain black hole mass and disk properties. Recent reverberation campaigns with the Neil Gehrels Swift Observatory and ground-based telescopes have revealed significant discrepancies between observed inter-band lags and standard accretion disk theory. These include unexpectedly large lag amplitudes (the “accretion disk size problem”) and weak correlations between X-ray and UV/optical light curves. To investigate further, I analyze recent Swift campaigns of Mrk 335 and Mrk 817 using GP-based frequency-resolved lag analysis. In both sources, standard disk lags appear only on short timescales (high frequencies), while longer-than-expected lags dominate at low frequencies. These lag excesses are consistent with reprocessing at larger radii, similar to the BLR. Mrk 817 offers a rare opportunity to connect the inner and outer accretion flow: I detect the first simultaneous measurement of X-ray and UVOIR lags, effectively mapping the full disk. These lags vary significantly over the campaign, with longer delays during periods of stronger X-ray obscuration. This suggests that a disk wind may modulate the observed lags by introducing additional reprocessing and/or blocking ionizing flux from reaching more-distant material. To test this obscuration effect across a population, I conduct the first statistical study of UV/optical lag excess versus physical parameters across the Swift campaigns. The results show that the lag excess is driven entirely by obscured AGN, while the lags of unobscured sources are, on average, consistent with thin-disk theory. Regression analysis reveals that X-ray column density explains over 80% of the variance in lag excess. As for the X-ray/UV connection, obscured AGN also tend to show weaker correlations and more variable lags, suggesting that line-of-sight absorption not only contributes additional reprocessed emission that extends the UV/optical lags, but may also decouple or delay the X-ray and UV variability. To make GP-based time series analysis accessible to the community, I developed the STELA Toolkit, a fully documented Python package for computing frequency-domain data products using GPs. I also benchmark GP performance against other interpolation methods, including state-of-the-art transformers, paving the way for scalable, ML-enabled timing analysis in the era of time-domain surveys like Vera Rubin.

Learning Nonlinear Dynamics: Methods and Applications

2026-01-13T03:36:32Z

Learning Nonlinear Dynamics: Methods and Applications Rossi, Baptiste T. Accurate modeling of dynamical systems through differential equations is essential for scientific prediction and prescriptive control. Traditional model development, which relies on expert knowledge, parameter fitting and validation, is often iterative, time-consuming, and complicated by real-world data complexities such as noise and missing observations. This thesis addresses these challenges by developing robust, scalable, and interpretable methods for learning nonlinear ordinary differential equations (ODEs) and partial differential equations (PDEs) directly from data, with a particular emphasis on applications in fluid dynamics. In Chapter 2, we introduce a novel methodology for learning arbitrary nonlinear ODEs using collocation methods combined with interpolation. This approach demonstrates enhanced robustness to noise and significant computational speed-ups compared to classical system identification techniques, including the popular SINDy framework. It also provides a constructive method for reconstructing unobserved system components, making it applicable to partially observed systems, and offers theoretical guarantees on accuracy traditionally absent in strong-form identification. In Chapter 3, we combine the approach from Chapter 2 with sparse regression to derive sparse ODEs from data, demonstrating enhanced robustness to observational noise. Our method shows improved performance in recovering the true structures and coefficients on canonical benchmark tests under significant noise, while the performance of traditional surrogate methods deteriorates even with minimal noise. In Chapter 4, we extend this methodology to Partial Differential Equations (PDEs) using the method of lines, addressing issues related to data scale and interpolation ill-posedness. With a focus on Computational Fluid Dynamics (CFD), we show that our method goes beyond recovering complex nonlinear PDEs, such as the Navier-Stokes equations, from simulation data. The method can also be used as an a-posteriori indicator of simulation quality, providing insights into the effective PDEs represented by a given simulation, and pinpointing error-generating areas to inform adaptive mesh techniques. Lastly, in Chapter 5, we introduce a novel data-driven framework for modeling turbulent phenomena, a long-standing challenge in aerospace and climate science. Our approach addresses the Reynolds-Averaged Navier-Stokes (RANS) closure problem, which involves modeling the unobserved eddy viscosity field. We tackle two interconnected inverse problems: reconstructing the eddy viscosity from flow data and discovering its governing partial differential equations (PDEs), thereby proposing a new pathway to uncover new or refined RANS closure models directly from high-fidelity simulations. This chapter establishes a tractable baseline using a composite loss function, which we evaluate on canonical turbulent flows. Our results demonstrate that while the approach can recover governing equations when the ground truth eddy viscosity is known, significant challenges remain due to noise and numerical errors. We conclude that a more advanced reconstruction methodology is essential for robustly discovering these models, underscoring the potential of this data-driven approach and identifying critical areas for future research.

The Electronic Compressibility of Rhombohedral Graphene Multilayers

2026-01-13T03:37:38Z

The Electronic Compressibility of Rhombohedral Graphene Multilayers Aronson, Samuel H. In condensed matter systems, energy bands with narrow dispersion frequently host correlated electronic phases that arise from strong Coulomb interactions. When these bands also have concentrated Berry curvature, the correlated phases may be topologically non-trivial. The low-energy bands of rhombohedral graphene multilayers possess both of these ingredients, making this a promising class of materials in which to search for correlated topological electronic ground states. This thesis describes our electronic compressibility measurements on rhombohedral graphene multilayers, with a particular focus on the pentalayer system (R5G). We utilize a planar capacitance technique that probes the thermodynamic density of states and enables us to extract energy gaps of incompressible phases. We observe a variety of correlated electronic phenomena including half and quarter metals, layer antiferromagnetism, correlation-driven Chern insulators, and thermodynamic signatures of potential Wigner crystallization. We also study the electronic compressibility of R5G aligned to a hexagonal boron nitride (hBN) substrate to form a moiré superlattice. Motivated by the recent discovery of the fractional quantum anomalous Hall effect in this system when the electrons are pushed away from the moiré interface by an external electric displacement field, we study the opposite moiré-proximal limit, in which the superlattice potential is considerably stronger. We observe integer and fractional Chern insulator states that persist down to low magnetic fields in addition to numerous trivial and topological charge density waves. We map out a phase diagram that is highly sensitive to both displacement and magnetic fields, establishing the R5G-hBN superlattice as a highly-tunable system for studying the interplay between intrinsic band topology and strong lattice effects.

Hadronic Structure with Classical and Quantum Computing

2026-01-13T03:36:35Z

Hadronic Structure with Classical and Quantum Computing Avkhadiev, Artur Calculations in lattice quantum chromodynamics (QCD) — presently the only known systematically improvable approach to describe the strong nuclear force in the nonperturbative regime from first principles — are playing an increasingly important role in revealing how hadrons emerge from the interactions of the underlying degrees of freedom: quarks and gluons. With computational and theoretical advances, more fruitful connections have emerged between lattice QCD and phenomenology, and the field is now well into a stage ripe for deriving tighter constraints on hadronic structure through joint analyses of numerical lattice QCD results with experimental data. This thesis summarizes lattice QCD calculations of the Collins-Soper (CS) kernel: a nonperturbative function whose inclusion in joint analyses has the potential to advance the study of multidimensional hadronic structure. The CS kernel is an anomalous dimension of transverse-momentum-dependent (TMD) distributions describing a three-dimensional structure of ultrarelativistic hadrons as a function of quark-gluon momenta collinear with and transverse to the hadron's motion. Constraints on the CS kernel at nonperturbative transverse-momentum scales are instrumental to relate TMDs across scales and processes. The kernel differs for quark and gluon TMDs, but is otherwise universal. This thesis presents the first lattice QCD determination of the quark CS kernel with systematic control over operator mixing, quark mass, and lattice discretization, and a proof-of-principle lattice calculation of the gluon CS kernel providing the first nonperturbative constraints on this quantity. Additionally, this thesis summarizes exploratory studies on how Hamiltonian calculations — realized with quantum-computer simulations and tensor networks — may be combined with conventional Monte Carlo calculations based on Lagrangian formulations in Euclidean space. These studies examine how constructions of interpolating operators, used in conventional calculations to map between the vacuum and a ground state of interest, may be optimized in Hamiltonian calculations to increase overlap with the target state. Results, limited to the Schwinger model, support further investigations of this approach in theories more closely resembling QCD as quantum-computing and tensor-network technologies continue to mature.

Accelerating RTL Simulation Through Fine-grained Task Dataflow and Selective Execution

2026-01-13T04:08:30Z

Accelerating RTL Simulation Through Fine-grained Task Dataflow and Selective Execution Elsabbagh, Fares Fast simulation of digital circuits is crucial to build modern chips. Current processors and SoCs integrate hundreds of complex components, including cores, accelerators, and memory hierarchies. Simulating these systems is necessary to verify correctness and explore the design space. Simulation can happen at different levels of abstraction. In this work we focus on Register-Transfer-Level (RTL) simulation. While RTL simulators are frequently used in development due to their quick compilation times, their runtime performance is slow. This is because as the designs are scaled up, multicore communication and scheduling overheads limit performance and scalability. We present ASH, a parallel architecture tailored to RTL simulation workloads. ASH consists of a tightly codesigned hardware architecture and compiler for RTL simulation. ASH exploits two key opportunities. First, it performs dataflow execution of small tasks to leverage the fine-grained parallelism in simulation workloads. Second, it performs selective event-driven execution to run only the fraction of the design exercised each cycle, skipping ineffectual tasks. ASH hardware provides a novel combination of dataflow and speculative execution, and ASH’s compiler features several novel techniques to automatically leverage this hardware. We evaluate ASH in simulation using large Verilog designs that represent different types of architectures. With 256 simple cores, ASH is gmean 1,485× faster than 1-core Verilator, and it is 32× faster than Verilator on a server CPU with 32 complex cores while using 3× less area.

Task Scheduling Techniques to Accelerate RTL Simulation

2026-01-13T04:08:25Z

Task Scheduling Techniques to Accelerate RTL Simulation Sheikhha, Shabnam Fast simulation of digital circuits is crucial to build modern chips. Slow simulation lengthens chip design time and makes bugs more frequent. While simulation can happen at different levels of abstraction, Register-Transfer-Level (RTL) simulation is the usual bottleneck in chip design, as it is needed for ongoing debugging and evaluation. Current simulators scale poorly across CPU cores, because they are unable to exploit the fine-grained parallelism inherent in simulation workloads. We present ASH, a parallel architecture tailored to simulation workloads. ASH consists of a tightly codesigned hardware architecture and compiler for RTL simulation. ASH exploits two key opportunities. First, it performs dataflow execution of small tasks to leverage the fine-grained parallelism in simulation workloads. Dataflow execution exposes abundant parallelism, as each task can run as soon as its inputs are available. Second, it performs selective event-driven execution to run only the fraction of the design exercised each cycle, skipping ineffectual tasks. Selective execution introduces dynamic data dependences since skipped tasks do not communicate data. ASH employs speculative execution to handle these dependencies. ASH’s hardware provides a novel combination of dataflow and speculative execution, and ASH’s compiler features several novel techniques to automatically leverage this hardware. The key compiler techniques include a novel partitioning for minimizing data communication while maintaining load balance, and a strategic coarsening mechanism to reduce the overheads of fine-grained tasks. We evaluate ASH in simulation using large Verilog designs. With 256 simple cores, ASH is gmean 1,485× faster than 1-core Verilator, and it is 32× faster than Verilator on a server CPU with 32 complex cores while using 3× less area.

Scheduling Strategies for Bus Operator Retention: A Mixed-Methods Evaluation of Bus Operator Preferences and 4-Day Workweek Feasibility

2026-01-13T04:08:29Z

Scheduling Strategies for Bus Operator Retention: A Mixed-Methods Evaluation of Bus Operator Preferences and 4-Day Workweek Feasibility Baum, Amelia Rose Public transit agencies face significant and growing challenges related to workforce shortages, absenteeism, and employee retention, which threaten service reliability. Reports found that 90% of U.S. transit agencies are experiencing a workforce shortage, with 84% claiming that the shortage affects their ability to provide scheduled service. Industry-wide, operator absence is a significant contributor to missed work at transit agencies nationwide and has, in many cases, delayed the full reinstatement of service at transit agencies following the COVID-19 pandemic. The quality of bus operators' work is significantly impacted by inflexible crew scheduling constraints. However, most studies focus on pay, benefits, and infrastructure, neglecting the importance of scheduling. This thesis aims to fill this gap by examining the potential for crew scheduling improvements to enhance the quality of life for bus operators through a three-part case study at the Chicago Transit Authority. Part 1 analyzes the historical work preferences of CTA bus operators, providing actionable insights for scheduling improvements. Part 2 presents a high-fidelity proof of concept in HASTUS, using block schedules (10-hour-a-day runs that are intended to be run by an operator 4 days a week) and rostering to reduce negative work traits, increase consecutive and weekend days off for most operators, while maintaining schedules for the top 20% of senior operators. Part 3 evaluates the new 10-hour, 4-day-per-week packaged schedules via an LLM-based paired alternatives survey of operators at one CTA garage, measuring the desirability of the proof of concept and collecting qualitative feedback. Overall, the new schedules substantially improve the quality of work for operators by guaranteeing at least one weekend day off, at least two consecutive days off, and increasing day-to-day schedule consistency and overnight rest time, while maintaining constant vehicle requirements and total pay hours. The survey results show that 72% of operators at the 74th Street garage support the new schedule paradigm, demonstrating strong support for their potential adoption and encouraging future exploration of a block schedule hybrid rostering paradigm at the CTA and other transit agencies.

A Charcuterie Platter of QCD Matter

2026-01-13T03:37:36Z

A Charcuterie Platter of QCD Matter Sun, Zhiquan One of the greatest current challenges in theoretical high energy physics is to understand the dynamics of Quantum Chromodynamics (QCD). In this thesis, I address a variety of questions in QCD using Effective Field Theory (EFT). The first question deals directly with the observed phenomenology of QCD: How can we use EFT to disentangle the complicated three-dimensional dynamics of how quarks and gluons, the fundamental degrees of freedom of QCD, combine to form the observed bound states in nature called hadrons? I initiate a new formalism using Heavy Quark Effective Theory to study this dynamical process known as hadronization. I shed new light on the transverse momentum-dependent fragmentation process of heavy (charm and bottom) quarks by making use of the fact that heavy quarks with masses much larger than the strong interaction scale decouple from the rest of the hadronization cascade. I also present exciting heavy quark phenomenology at existing colliders and the upcoming Electron-Ion Collider. The second question investigates the field theory structure of QCD: What can we learn about the nonperturbative structure of the quantum field theory through the abstruse emergent phenomenon in QCD called “confinement”, which traps quarks and gluons inside hadrons? I study a class of cleverly constructed observables known as energy correlators by using fieldtheory based methods to determine the leading nonperturbative contribution, and examine the universality of the nonperturbative matrix element that gives rise to this contribution. I also show that including the nonperturbative contribution has a significant impact on the extraction of the strong coupling constant, a fundamental parameter of the Standard Model, using tools such as factorization and resummation from EFT. Last but not least, the final question explores the underlying symmetry properties of QCD and its potential completions: How robust is the axion solution to the strong CP (ChargeParity) problem, and what are some of its implications beyond the realm of QCD? I examine the axion quality problem in post-inflationary QCD axion models with different symmetry properties and identify a new tension with standard cosmology. I further show that the axion string-domain wall dynamics is more complicated than commonly expected, undermining the reliability of a unique mass prediction for axion dark matter in post-inflationary models. I showcase the importance of considering both high-energy extensions and the EFT at low energy, and uncover new complexity of the axion solution to the strong CP problem.

Learning from pre-pandemic data to design and test future-proof therapeutics

2026-01-13T03:37:25Z

Learning from pre-pandemic data to design and test future-proof therapeutics Gurev, Sarah Effective pandemic preparedness relies on predicting immune-evasive viral mutations to enable early detection of variants of concern and design vaccines and therapeutics that are resilient to future viral evolution. However, current strategies for viral evolution prediction are not available early in a pandemic and have limited predictive power – experimental approaches require host polyclonal antibodies and existing computational methods draw heavily from current strain prevalence. In addition, vaccines and therapeutics have been designed with an eye towards past or circulating variants, not towards future evolution. To address these challenges, we developed EVEscape, a generalizable framework that integrates fitness predictions from a deep generative model of evolutionary sequences with biophysical and structural information. EVEscape quantifies the immune escape potential of viral strains at scale and is applicable before surveillance sequencing, experimental scans, or 3D structures of antibody complexes are available. We demonstrate that EVEscape, trained on sequences available prior to 2020, performs as accurately as high-throughput experimental scans at anticipating pandemic variation for SARS-CoV-2 and is generalizable to other viruses including Influenza A virus, HIV, and understudied viruses with pandemic potential such as Lassa and Nipah. We investigate both alignment-based and protein language models to explore the best model of mutation effects across pandemic-threat viral families. We demonstrate the utility of EVEscape in three critical applications: (1) Surveillance efforts flagging high escape SARS-CoV-2 variants from their first appearance (2) Design of panels of viral antigens that mimic future viral variants for early, proactive evaluation of the future protection of vaccines and therapeutic; and (3) Design of a pan-sarbecovirus nanoparticle-based vaccine capable of eliciting broad, long-lasting protection against sarbecoviruses, including future variants. This three-pronged approach represents a paradigm shift in pandemic preparedness, offering a novel strategy to preemptively address viral families with pandemic potential and significantly bolster global prevention efforts.

Topics in quantum information theory and quantum many-body physics

2026-01-13T03:37:33Z

Topics in quantum information theory and quantum many-body physics Balasubramanian, Shankar In this thesis we present two results relating to the intersection between quantum information theory and quantum many-body physics. The first pertains to quantum algorithms, where few computational problems are believed to exhibit exponential separation between quantum and classical performance. For those that are, natural generalizations remain elusive. One speedup that has especially resisted generalization is the use of quantum walks to traverse the welded tree graph, due to Childs, Cleve, Deotto, Farhi, Gutmann, and Spielman. We show how to generalize this to a large class of hierarchical graphs in which the vertices are grouped into “supervertices” that are arranged according to a d-dimensional lattice. Supervertices can have different sizes, and edges between supervertices correspond to random connections between their constituent vertices. The traversal time of quantum walks on these graphs are related to (a) the existence of small subspaces within which the quantum walk evolves and (b) the localization properties of the quantum walk within these subspaces. We find examples of hierarchical graphs that yield provable speedups over classical algorithms ranging from superpolynomial to exponential, depending on the underlying dimension and the random graph model. We also discuss how to relax criterion (a) to the existence of a small and approximate subspace by using techniques from graph sparsification. The second result pertains to fault-tolerant quantum memories. Storing a qubit in a noisy environment is crucial for developing full-scale quantum computers. While constructions of fault-tolerant quantum memories exist, they often assume that quantum operations are not local and assisting classical computation operates instantaneously and noislessly. In particular, constructing a topological quantum memory below four dimensions with local quantum and classical operations that is fault-tolerant under both quantum and classical noise is an open problem. We construct a local quantum memory for the 2D toric code using ideas from the classical cellular automata of Tsirelson and Gács. Our memory preserves a logical state for exponential time in the presence of both classical and quantum noise below a constant noise rate. While our 2D quantum memory is built from operations that depend on space and time, we construct a fault-tolerant quantum memory in 3D using stacks of 2D toric codes that can be built with time-independent operations.

Quantum Matter in the Era of Generalized Symmetries

2026-01-13T03:37:15Z

Quantum Matter in the Era of Generalized Symmetries Chatterjee, Arkya The discovery of generalized symmetries has led to powerful new insights into quantum matter. They have been used to classify new families of quantum phases, place constraints on phases realizable in a given physical system, and conceptually unify seemingly disparate phenomena. In many ways, they prove just as powerful as traditional symmetries at organizing and constraining the theories that describe quantum matter. In this thesis, we attempt a unification of such constraints by developing a holographic correspondence between (generalized) symmetries and topological orders, called the Sym/TO correspondence. For any (finite internal) symmetry of a quantum system in d (spatial) dimensions, we associate with it a unique topological order in d + 1 dimensions, called its Symmetry Topological Order (SymTO). We devise an operator algebraic recipe to compute the SymTO data for any lattice spin model, demonstrating it in a number of examples. We then use the SymTO to classify possible quantum phases allowed by the symmetry—we call this a generalized Landau paradigm. Besides classifying phases, we also identify constraints on the phase transitions between them using a SymTO-resolved modular bootstrap. We test this framework in a quantum spin chain with non-invertible symmetries. We discover a new Kramers-Wannier-like duality and a rich phase diagram including a noninvertible symmetry-enriched incommensurate phase. The translation symmetry of the spin chain has a nontrivial interplay with the lattice Kramers-Wannier duality, which matches the anomaly of the corresponding non-invertible symmetry in the low-energy effective field theory. Finally, we explore such unusual anomaly-matching mechanisms in more detail in the context of the chiral anomaly of a single massless Dirac fermion, demonstrating a novel lattice realization of chiral symmetries and their anomaly.

Systems Materials Design of Ordered Nanocomposite Assemblies

2026-01-13T03:37:31Z

Systems Materials Design of Ordered Nanocomposite Assemblies Thrasher, Carl James The ability to precisely organize matter across multiple length scales is a central challenge in modern materials science. In this dissertation, I develop a systems materials design approach to engineer hierarchically structured nanocomposite assemblies, integrating molecular recognition, supramolecular chemistry, colloidal assembly, and bulk processing into unified material platforms. At the molecular and nanoscale, I investigate how multivalent supramolecular interactions can be rationally programmed by controlling the architecture of polymer binders grafted to nanoparticle surfaces. Through systematic variations in polymer topology, recognition group density, and scaffold geometry, I demonstrate how polymer design dictates the thermodynamic strength and multivalency of nanoparticle superlattice assembly, enabling precise control of thermal stability, crystallographic symmetry, and collective bonding behaviors in massively multivalent systems. Building on these design principles, I develop a colloidal metallurgy framework to process selfassembled nanoparticle superlattices into dense macroscopic polycrystalline solids while preserving nanoscale order. By systematically studying the interplay of temperature, pressure, and time during colloidal sintering, I elucidate mechanisms of densification, defect evolution, and grain growth unique to colloidal systems, establishing processing–structure relationships that parallel but fundamentally diverge from atomic sintering. Finally, I extend these concepts to create stretchable nanocomposite supercrystals, embedding supramolecularly assembled superlattices into elastomeric matrices via co-engineered polymer chemistries that enable hierarchical strain transduction. These materials combine the nanoscale precision of supercrystals with mechanical robustness, reconfigurability, and stimuli-responsive optical properties, illustrating a scalable pathway to multifunctional metamaterials. Collectively, this work demonstrates how a systemslevel integration of molecular design, colloidal assembly, and bulk processing enables new paradigms for the synthesis of hierarchically ordered, functional nanocomposites.

The novel roles of BCL6 and BATF3 in regulating human CD8⁺ T cell dysfunction

2026-01-13T03:36:48Z

The novel roles of BCL6 and BATF3 in regulating human CD8⁺ T cell dysfunction Traunbauer, Anna Katharina Reduced effector function and elevated inhibitory receptor expression are hallmarks of exhausted CD8⁺ T cells, yet the underlying molecular and epigenetic drivers remain incompletely defined. Here, we developed an in vitro repeated stimulation model to recapitulate features of human CD8⁺ T cell dysfunction and delineate transcriptional and epigenetic landscapes. Our analyses revealed that BCL6 and BATF3 are robustly upregulated in dysfunctional CD8⁺ T cells, with ATAC-seq demonstrating enhanced chromatin accessibility at their gene loci. Transcription factor footprinting shows increased BATF3 motif occupancy in chronically stimulated cells and integrative multi-omic analysis combining footprints, open chromatin regions, RNA-seq and ChIP-seq data revealed that putative BATF3 target genes may include master regulators of exhaustion. Moreover, overexpression of BCL6 or BATF3 markedly upregulates TIM-3 expression and suppressed cytokine release, establishing their capacity to induce T cell dysfunction. We further validated these findings ex vivo in antigen-specific CD8⁺ T cells from patients with advanced melanoma, as well as HCV and HIV infections, where cells were enriched for BCL6^high and BATF3^high subsets co-expressing canonical exhaustion markers such as PD-1, TIM-3 and CD39. Notably, Single-cell RNA sequencing of HIV-specific CD8⁺ T cells identified a distinct BCL6^high PD1⁻ progenitor population that gives rise to two distinct subsets via divergent differentiation trajectories: one branch generates effector-like BCL6^high PD1⁺ cells, whereas the other produces BCL6^high PD1⁺ cells that retain an exhaustion gene signature alongside partial memory-like feature. Collectively, these findings identify BCL6 and BATF3 as key mediators of human CD8⁺ T cell dysfunction and illuminate novel transcriptional and epigenetic pathways that may be leveraged for therapeutic intervention in cancer and chronic viral infections.

Aspects of Nonperturbative Heavy Quark Physics

2026-01-13T03:37:27Z

Aspects of Nonperturbative Heavy Quark Physics Lin, Joshua The properties of charm and bottom quarks are an interesting corner of Quantum Chromo-Dynamics (QCD) due to the fact that their masses are much heavier than the typical QCD interaction energy ΛQCD. Due to this scale separation, it is possible to describe these heavy quarks by Effective Field Theories (EFTs) that simplify their equations of motion, make explicit additional symmetries that only appear for heavier quark masses, and simplify the theoretical calculations required for predictions. By discretising these EFTs in a lattice regularisation, nonperturbative calculations of observables of interest become possible. This thesis presents progress towards systematically controlled calculations of two such observables: the Spectator Effect contributions to the inclusive decay rates of b-hadrons, and the real-time dynamics of fermions propagating in a thermal medium. Standard EFT calculations in Lattice-QCD proceed by expressing observables as sums over perturbatively computed Wilson coefficients and nonperturbative matrix elements that can be calculated by path integral monte-carlo methods. Though it is possible to carry out this procedure within a regulator-independent renormalization scheme, in practice almost all such decompositions are computed in the modified minimal subtraction scheme MS which is only defined for the dimensional regulator (DR), due to its simplicity. Computing such observables therefore requires a matching between lattice regularised operators and operators renormalized in MS. In Chapter 2, both the dimensional regulator (DR) and the lattice regulator are reviewed, with a particular emphasis on techniques needed for calculations carried out in later sections. An interesting subtelty in DR is the need to introduce d-dimensional counterparts to the Dirac γ-matrices, which a-priori are only well defined in integer number of dimensions. This analytic continuation is of practical importance as it introduces additional Evanescent Operators (Sec. 2.1.4) that have physical consequences. In Sec. 2.1.5, traces of d-dimensional γ-matrices were related to Tutte polynomial evaluations [4], presenting a new graph-theoretic interpretation of the dimensionally regulated γ-matrices. One strategy of renormalizing lattice-regulated operators into MS involves first renormalizing into a regulator independent scheme, before perturbatively matching between the regulator independent scheme and MS. In Chapter 3, regulator independent position-space (X-space) schemes for renormalizing operators defined in the leading order Heavy Quark Effective Theory (HQET) are proposed [3]. Compared to other regulator independent renormalization schemes such as RI-xMOM, X-space schemes have the benefit that they are gauge invariant. The next to leading order matching calculations between X-space and MS are presented for heavy-light and heavy-light-light multiplicatively renormalizable operators, as well as ∆Q = 0 and ∆Q = 2 four quark operators relevant for heavy hadron decays and mixing, where Q refers to the static quark number. Due to their heavy masses, hadrons containing heavy quarks decay via the weak nuclear force. Experimental measurements of these lifetimes provide precision determinations of the fundamental parameters of the Standard Model. The Heavy Quark Expansion expresses the inclusive lifetimes of heavy hadrons in terms of matrix elements of HQET operators of increasing dimension. The Spectator Effects are contributions due to the ∆Q = 0 four-quark operators, where the light quark degrees of freedom within a heavy hadron participate in the decay. In Chapter 4, a Lattice-QCD determination of the static decay constant f HQET B and the isospin-nonsinglet portion of the Spectator Effect matrix elements for heavy-light mesons is presented. Fits of bare matrix elements were performed for three different lattice spacings, and renormalized with the schemes proposed in Chapter 3 before a continuum limit is taken. Due to the heavy masses mQ of the heavy quarks, it is possible to find temperatures T approximately satisfying a hierarchy ΛQCD ≪ T ≪ mQ. At these temperatures, QCD undergoes a deconfinement transition into the Quark-Gluon-Plasma (QGP) phase where the light degrees of freedom are no longer confined, and instead screen the long-range colour forces. The heavy quarks however are not thermalised, and act as probes of the QGP. Further understanding of the QGP requires first principles simulations of the heavy quark dynamics at finite temperature, however such calculations are difficult due to the enormous size of the Hilbert space. Variational approximations of the Hilbert space encode wavefunctions within a few parameters, and provide a practical method to simulate many particle systems. As a testcase, the variational approach was applied for the first time to simulate fermions at finite temperature in a simple QFT: the 1+1d U(1) gauge theory known as the massive Schwinger model. Both the real-time dynamics of string like states, and the properties of the thermal state were studied, and such variational methods are shown to be promising approaches to the more difficult case of a heavy quark effective theory in QCD.

Probing the Nonperturbative Physics of QCD with Normalizing Flows and a moderate number of Pions

2026-01-13T03:36:19Z

Probing the Nonperturbative Physics of QCD with Normalizing Flows and a moderate number of Pions Abbott, Ryan William Quantum Chromodynamics (QCD) is a cornerstone of the standard model of particle physics, and the best known theory of strong nuclear interactions. The only known systematically improvable ab-initio method for accessing the nonperturbative physics of QCD is Lattice QCD is, and this thesis presents two advances in our understanding QCD using lattice-based methods. The first is a calculation using many-pion systems to map out the entire zero temperature, nonzero isospin density region of the QCD phase diagram. The calculation uses novel methods for working with many-pion systems that enables working with thousands of pions, and furthermore provides rigorous constraints on the baryon-dense region of the QCD phase diagram. The second is an application of methods from machine learning (namely normalizing flows) in order to accelerate sampling. This approach has the promise of eliminating issues such as critical slowing down, as well as introducing novel tools and methods that enable methods of calculation that would be possible otherwise.

Limits of QCD

2026-01-13T03:37:21Z

Limits of QCD Gao, Anjie This thesis explores the fundamental kinematic limits of Quantum Chromodynamics (QCD), including the soft, collinear, and Regge limits, using soft-collinear effective theory (SCET). We begin by studying transverse momentum dependent (TMD) physics in semi-inclusive deep inelastic scattering (SIDIS), which probes the small transverse momentum regime arising from the soft and collinear limits of QCD. We derive all-order factorization theorems for azimuthal asymmetries in SIDIS at next-to-leading power (NLP). We also propose a new angular observable, q_∗, for probing TMD dynamics at the future Electron-Ion Collider (EIC), which enables an order-of-magnitude improvement in experimental resolution while retaining sensitivity to TMD distributions. Next, we apply the TMD formalism to a class of observables known as energy correlators. We study the transverse energy-energy correlator (TEEC) in the back-to-back limit, a dijet observable at hadron colliders, and the three-point energy correlator (EEEC) in the coplanar limit, a trijet observable at lepton colliders. For both observables, we derive allorder factorization theorems and resum large logarithms to next-to-next-to-next-to-leading logarithmic (N3LL) accuracy. Finally, we analyze the Regge limit of 2 → 2 QCD amplitudes. By factorizing these amplitudes into collinear jet and soft functions and studying their rapidity evolution, we define Regge-like anomalous dimensions in a gauge-invariant manner. At the level of the exchange of two Glauber gluons in the t-channel, we recover the BFKL equation from a purely collinear perspective. Extending to three-Glauber exchange, we derive the first closed-form renormalization group equations for Regge cut contributions in several nontrivial t-channel color representations, providing a systematic method for organizing non-planar QCD amplitudes at high energy.

Determining the Molecular Underpinnings of Iron Homeostasis in Human Cells

2026-01-13T03:35:45Z

Determining the Molecular Underpinnings of Iron Homeostasis in Human Cells Lee, April Precise regulation of nutrient availability is crucial for cellular function and survival. Iron, in particular, is tightly regulated as it serves as an essential cofactor for numerous enzymes but can catalyze the formation of toxic radicals at elevated levels. To maintain the necessary cytoplasmic iron concentration, cells store excess iron in large proteinaceous cages called ferritin and, when available iron levels fall, they degrade these cages, liberating the stored iron for use. This thesis focuses on the molecular mechanisms underlying cellular iron sensing, as well as the molecular interactions supporting regulated ferritin degradation and subsequent iron release. Specifically, this work interrogates the protein interactions involved in ferritinophagy, a form of selective autophagy that leads to the lysosomal degradation of ferritin. Extending prior work that identified key components supporting ferritinophagy, including the selective autophagy receptor protein NCOA4 and its cognate autophagosomal receptor GATE16, experiments described here uncover the molecular contacts between these proteins. I found that NCOA4 bears two short linear motifs that each bind to GATE16 with weak affinity. However, these binding motifs are highly avid and, in concert, support high-affinity binding of NCOA4 to oligomerized GATE16. I further describe that ferritin degradation in cultured human cells relies on the contacts I identified biochemically. Moreover, I found that iron decreases NCOA4’s affinity for GATE16, providing a plausible mechanism for irondependent regulation of ferritinophagy. Taken together, this work suggests a general mechanism by which selective autophagy receptors can distinguish between inactive monomeric GATE16 and the active oligomerized forms that primarily drive autophagy. In related studies, I have biochemically probed the NCOA4•ferritin interface, with these experiments suggesting a novel function of NCOA4 in modulating ferritin cage structure – either through cage dismantling or through the formation of higher order structures. Taken together, these studies further define the molecular mechanisms by which NCOA4 aids cells in maintaining iron homeostasis, and they provide the requisite reagents for future work aimed at building a unified model for how mammalian cells regulate this vital but toxic metal.

Sampling Methods for Fast and Versatile GNN Training

2026-01-13T04:08:28Z

Sampling Methods for Fast and Versatile GNN Training Alkhatib, Obada Graph neural networks (GNNs) have become a commonly used class of machine learning models that achieve state-of-the-art performance in various applications. A prevalent and effective approach for applying GNNs on large datasets involves mini-batch training with sampled neighborhoods. Numerous sampling algorithms have emerged, some tailored for specific GNN applications. In this thesis, I explore ways to improve the efficiency and expressivity of existing and emerging sampling schemes. First, I explore system solutions to facilitate the development of fast implementations of different sampling methods. I introduce FlexSample, a system for efficiently incorporating custom sampling algorithms into GNN training. FlexSample leverages the types of performance optimizations found in SALIENT, a state-of-the-art system for fast training of GNNs with node-wise sampling. In experiments with 4 GNN models which use layer-wise and subgraph sampling, FlexSample achieves up to 1.3× speed-up for end-to-end training over PyTorch Geometric with the same sampling code. Furthermore, FlexSample extends SALIENT with highly-optimized C++ implementations of FastGCN and LADIES layer-wise sampling, which achieve 2×–5× speed-up over their respective Python implementations. Second, I introduce a novel framework for learning neighbor sampling distributions as part of GNN training. Key components of this framework, which I name PertinenceSample, are: (i) a differentiable approximation of node-wise sampling for GNNs; and (ii) a parametrization of node sampling distributions as node- or edge-wise weights of attention-like GNN layers. I present an initial exploration of the potential of PertinenceSample for improving node classification accuracy in the presence of noisy edges. Specifically, in two synthetic experiments where roughly half of a node’s neighbors may have similar features but different labels, I demonstrate that extending a GraphSAGE model with a 2-layer perceptron for learning the PertinenceSample weights can improve classification accuracy from 50%–75% to (nearly) 100%.

Designing Electrocatalysts for the Production and Oxidation of Liquid Fuels

2026-01-13T03:35:35Z

Designing Electrocatalysts for the Production and Oxidation of Liquid Fuels Zheng, Daniel J. With the ever-rising CO₂ levels in the atmosphere, it is paramount to cease reliance on fossil fuels to meet global energy demands. While the cost of electricity from renewable sources, such as solar and wind, continues to decrease and has even fallen below that of fossil fuels since 2014, these renewable energy sources suffer from intermittency, potentially causing shortages at peak demands. Thus, methods to store or economically use excess renewable energy are needed for full decarbonization. One promising avenue is to store the excess generated electrical energy in chemical bonds, creating molecules and materials with industrial or energy storage utility. In this proposed scheme, the renewable electricity would be used to electrochemically convert earth-abundant molecules into value-added chemical or fuels. These generated products could then be utilized as feedstocks in industrial applications or as a fuel source to generate electricity when needed by transforming back into their earth-abundant forms. Central to transforming earth-abundant molecules into value-added chemicals or fuels is the oxygen evolution reaction (OER), which is found in nearly every process. The plentiful nature of OER’s main reactant, water, and moderate thermodynamic potential of 1.23 V vs. the reversible hydrogen electrode, make OER an ideal reaction to pair with other transformations. However, the slow kinetics of OER significant hinder the efficiency of these processes. As such, discovering new OER catalysts with high activity and stability would have wide-spread impacts. On the other hand, one of the most promising renewable fuel sources is methanol, which boasts about 3 times the energy density of hydrogen and can be used as an alternative to hydrogen in proton exchange membrane fuel cells. However, the sluggish kinetics of the methanol oxidation reaction (MOR), even with current state-of-the-art noble metal catalysts causes direct methanol fuel cells to reach an efficiency of <40%, limiting their practical usage. While significant research has been invested in discovering new MOR electrocatalysts, PtRu has reigned for 5 decades, highlighting the need for a true breakthrough. In this thesis, electrocatalysts for OER and MOR are examined in depth. For OER, metal-hydroxide organic frameworks (MHOFs), a promising new class of hybrid organic-inorganic materials with potential to mimic the superior functionality of enzymes, are studied. Operando vibrational and absorption spectroscopy methods are used to characterize the degradation mechanisms and lattice oxygen exchange capacity as a function of the linkers. Using such knowledge, defects are engineered into the MHOF that increase both the activity and stability compared to the pristine material. Furthermore, the traditionally reported MOR mechanism is studied using isotope-labeled reactants and operando mass spectrometry. These experiments revealed that, in contradiction to typically accepted mechanisms, the C-O bond in methanol can be cleaved during MOR, with the resulting CO₂ molecule containing two water-derived oxygen atoms, opening a new paradigm for MOR catalyst design. Driven by the need to discover new materials at scale, a fluorescence-based OER catalyst screening method is developed that can screen an entire composition space simultaneously. In addition, an AI-driven, automated platform for screening a high-dimensional multimetallic space for MOR is presented.

Bridging the Gap: From Artificial Intelligence and Optimization Theory to Action

2026-01-13T03:37:10Z

Bridging the Gap: From Artificial Intelligence and Optimization Theory to Action Petridis, Periklis S. Despite significant theoretical advances in Operations Research (OR) and Artificial Intelligence (AI), a persistent gap remains between these developments and their practical implementation in real-world settings. Despite significant progress in these fields, many OR and ML approaches struggle to scale to realistic problem sizes, lack robustness to uncertainty, or fail to address implementation constraints faced by practitioners in industry. Through four distinct works conducted in collaboration with industry partners, this research demonstrates how methodological advancements can bridge this theory-practice divide while maintaining rigorous theoretical foundations and guarantees. In the first part, we focus on optimization methodologies that scale traditional OR approaches to handle real-world problem sizes and uncertainty. In Chapter 2, we develop a stochastic Benders decomposition scheme for large-scale network design problems, a class of problems ubiquitous in logistics, transportation, and energy sectors. By incorporating sampling techniques within the decomposition framework, we achieve deterministic optimality guarantees while reducing computational costs, enabling solutions for networks with 700 nodes—an order of magnitude larger than previously tractable instances—while achieving optimality gaps of 5-7% compared to 16-27% for traditional deterministic approaches. In Chapter 3, we present a holistic framework for industrial decarbonization, developed with a major phosphate producer planning to quadruple energy consumption while transitioning to renewable sources. Our robust optimization approach combines strategic capacity expansion planning over a 25-year horizon with adaptive operational models, providing 95% reliability guarantees while balancing solar and wind integration with battery storage to meet a projected 12 TWh annual demand. In the second part, we shift our focus to developing AI systems that address the unique challenges of medical data abstraction and clinical decision support. In Chapter 4, we address the challenge of automating clinical data abstraction from electronic health records, collaborating with the Society of Thoracic Surgeons to populate their Adult Cardiac Surgery Database. Our AI pipeline combines 31 models per target variable with a two-tiered quality control system, achieving over 99% accuracy while automatically extracting 43-50% of registry variables, demonstrating how AI can dramatically reduce manual abstraction burden while maintaining clinical standards. In Chapter 5, we extend this healthcare AI focus by developing xHAIM (Explainable Holistic AI in Medicine), which addresses the limitations of current clinical AI systems in handling extensive patient records, providing interpretability, and incorporating medical knowledge. Through semantic similarity techniques and generative AI, xHAIM improves predictive performance while generating clinically grounded explanations that enhance trust and adoption by healthcare practitioners.

Multi-species genome-wide CRISPR screens identify conserved suppressors of cold-induced cell death

2026-01-13T03:36:20Z

Multi-species genome-wide CRISPR screens identify conserved suppressors of cold-induced cell death Lam, Breanna During hibernation of Syrian hamsters, the core body temperature shows a remarkable decrease, going from 37°C to 4°C. Although this ability to survive at low temperatures could in principle be due to systemic factors that occur during hibernation, we and others have seen that cells from hibernating rodents cultured in vitro maintain this ability. Although others have studied characteristics of cells from hibernating and non-hibernating organisms, the genes and pathways that are involved in cold-induced cell death have not been systematically explored. In this thesis, we conduct two genome-wide CRISPR-Cas9 screens in both a cold-sensitive (K562) and cold-resistant (BHK-21) cell line, and uncover GPX4 and related selenocysteine incorporation genes as important for protection against cold-induced cell death. Using genetic knockdowns, along with overexpression of GPX4, we confirm our findings and demonstrate that levels of GPX4 may be limiting in K562 cells, contributing to their cold sensitivity. Additionally, pharmacological validation using inhibitors of GPX4 reveal that the catalytic activity of GPX4 is dependent on the selenocysteine in the active site. Our findings are extended across multiple cell lines and cell types across six species. Taken together, our results suggest that GPX4 may be a powerful and conserved suppressor of cold-induced cell death. Building on our initial findings, we go on to show that cold exposure leads to increases in membrane permeability. This membrane permeability is transient, as rewarming of the cells reduces permeability to baseline levels. We also test the role of lipid peroxidation in contributing to membrane permeability and find that although it contributes in some cell lines, it is not the sole contributor as ferroptosis inhibitors do not fully mitigate membrane permeability. We go on to test different membrane channels and do not see decreases in membrane permeability, potentially indicating pathway-independent effects of temperature on membrane permeability. Altogether, this work provides a foundation for understanding how cold exposure influences mammalian cells.

Why Landfills Endure: Quantifying economic barriers to material and energy recovery from municipal solid waste in the United States

2026-01-13T03:37:18Z

Why Landfills Endure: Quantifying economic barriers to material and energy recovery from municipal solid waste in the United States Baidoo, Jacqueline E. Municipal solid waste (MSW) is a heterogeneous mixture of materials discarded by residential and nonresidential generators at end-of-life processing facilities for treatment and disposal. Conventional treatment methods reduce waste volumes through recycling via material recovery facilities, energy recovery via municipal solid waste incinerators, and biochemical conversion via composting. Even so, nearly 50% of total MSW generated in the United States was sent to landfills for final disposal in 2018 and almost half of all landfills currently in operation are expected to reach capacity by 2050. Waste planners seek to use developing resource recovery technologies like dry anaerobic digestion, gasification, and pyrolysis to narrow the gaps in end-of-life processing. Such technologies are posited to improve materials circularity and advance zero-waste landfill diversion goals by transforming residuals into electricity, fuels, and precursors to chemicals and fertilizers. However, despite demonstrated improvements to technical inefficiencies in waste valorization, numerous projects built on these technologies have failed to break through to commercial success. We investigate the contribution of regional and economic factors to the success of resource recovery projects through the lens of why landfills remain the predominant method of waste disposal. We build cost models of conventional and select developing treatment methods and use discounted cash flow analysis to estimate financial feasibility by local MSW compositions as reported in regional waste characterization studies. Findings indicate that the most critical factor to sustainable operation is consistent supply of waste materials at the quality and scale that maximize production efficiency, which is not achievable without rigorous data monitoring of MSW composition. Conversely, dependence on waste volume rather than composition makes land disposal a uniquely flexible pathway capable of subsidizing the costs of resource recovery. Progress towards landfill diversion is economically linked to the opportunity cost of avoiding landfill utilization. Unless municipalities are able to introduce subsidies elsewhere in the waste management ecosystem through gate fees and credits, projects will fail where marginal net costs of diversion exceed the revenues lost from avoided landfilling. Targeted processing of organic wastes can facilitate an average diversion of 24% for the compositions surveyed and was found to be viable for composting and dry anaerobic digestion projects at low to negligible financial losses compared to landfilling.

Enlightening Artificial Intelligence with Science

2026-01-13T03:37:08Z

Enlightening Artificial Intelligence with Science Liu, Ziming Today’s artifciail intelligence (AI) systems, while remarkably capable, are largely black boxes. The black-box nature raises concerns for those who build AI – “How can we construct an understand AI in scientifically grounded ways?”, and those who use AI – “How can we trust systems we do not understand?”. This thesis takes a humble step towards addressing the black-box problem. Building white boxes with science (Science for AI): The prevailing paradigm in AI today – “scaling is all you need" – focuses on scaling up existing models. However, this approach often yields systems that are neither interpretable nor efficient. I argue that scientific principles offer fresh perspectives for designing more transparent and effective AI systems. This is demonstrated through Kolmogorov-Arnold Networks (KANs) inspired by mathematics, Poisson Flow Generative Models (PFGM) rooted in physical intuition, and brain-inspired modular training (BIMT) drawing insights from neuroscience, etc. Opening black boxes (Science of AI): Modern AI models exhibit a range of puzzling behaviors – such as grokking, neural scaling laws and emergent representation learning – whose underlying mechanisms remain poorly understood. I employed simplified “spherical cow” models to investigate these phenomena from the perspective of phase transitions. I will show that grokking is a special phase in the hyperparameter space, which can be controlled and eliminated. The learned algorithms after grokking also display distinct phases, called clock or pizza algorithms. AI for Science: With greater interpretability, AI systems can begin to function as “AI Scientists” capable of (re)discovering deep scientific structures from data. These include conservation laws, hidden symmetries, integrable systems, Langrangian and Hamiltonian formulations, modular structures, and high-precion solutions. I believe my research work contributes to the emerging interdiscipinary field that unites AI and Science. Building opon the foundation laid in this thesis, I envision a future in which science guides AI out of its current era of alchemy and into a true era of scientific understanding.

Quantitative modeling of 5' splice site subclass regulation and evolution

2026-01-13T03:36:17Z

Quantitative modeling of 5' splice site subclass regulation and evolution Kenny, Connor Jens Pre-mRNA splicing is an essential molecular process required for eukaryotic gene expression. In this thesis, I present a previously unknown mechanism of splicing regulation where a family of splicing factors, the LUC7 family, compete to differentially impact 5→ splice site (5→ SS) selection in a sequence-dependent manner. I quantitatively characterize two major subclasses of 5→ SS in eukaryotes and outline distinctive features of 5→ SS in exons affected by the three human LUC7 paralogs: LUC7L2 and LUC7L enhance splicing of “right-handed” 5→ SS that exhibit stronger consensus matching on the intron side of the nearly-invariant / GU, while LUC7L3 boosts splicing of “left-handed” 5→ SS with stronger consensus matching upstream of the /GU. Using a range of experimental systems, from human cells to mutant plants, I show that LUC7 paralogs have opposing effects on these two 5→ SS subclasses and that this regulatory mechanism likely originated in the last common ancestor of animals and plants over 1.5 billion years ago. I further evaluate a competing model of 5→ SS subclass regulation involving METTL16- mediated U6 snRNA modification and reconcile both models by devising computational tools that identify sequence features predictive splicing dysregulation in transcriptome-wide datasets. Finally, I examine the evolutionary dynamics of left- and right-handed 5→ SS and propose a model of intron evolution in which codon and intron phase constraints in protein-coding genes shape both minor-to-major intron conversion and transitions between left- and right- 5→ SS subclasses.

Compiler-Hardware Co-Design for Pervasive Parallelization

2026-01-13T03:37:12Z

Compiler-Hardware Co-Design for Pervasive Parallelization Ying, Victor A. Modern computer systems have hundreds of processor cores, so highly parallel programs are critical to achieve high performance. But parallel programming remains difficult on current systems, so many programs are still sequential. This dissertation presents new compilers and hardware architectures that can parallelize complex programs while retaining the simplicity of sequential code. Our new systems allow real-world programs to use hundreds of cores without burdening programmers with concurrency, deadlock, or data races. This dissertation follows a novel approach that eliminates the burden of explicit parallel programming to make parallel execution pervasive. This approach relies on four guiding principles. First, exploiting implicit parallelism preserves the simplicity of sequential execution. Second, dividing computation into tiny tasks, as short as tens of instructions each, unlocks plentiful fine-grained parallelism in challenging programs. Hardware-compiler co-design techniques can create many tasks in parallel and reduce per-task overheads to make tiny tasks scale to many cores. Third, new hardware and software mechanisms can compose parallelism across entire programs, removing serializing barriers to overlap executions of nested parallel subroutines. Finally, exploiting static and dynamic information for data locality reduces data movement costs while maintaining load balance on large multicore systems. This dissertation presents three systems that embody these four principles. First, T4 introduces automatic program transformations that exploit a novel hardware architecture to parallelize sequential programs. As a result, T4 scales hard-to-parallelize real-world programs to tens of cores, resulting in order-of-magnitude speedups. Second, S5 builds on T4 with novel transformations to remove needless serialization in a broad class of challenging data structures. Thus, S5 scales complex real-world programs to hundreds of cores, delivers additional order-of-magnitude speedups over T4, and outperforms manually parallelized code tuned by experts. Finally, ASH is an accelerator that demonstrates the same approach can be applied with simpler mechanisms tailored for digital circuit simulation. A small ASH implementation is 32x faster than a large multicore CPU running a state-of-the-art parallel simulator.

Optimizing Graph Neural Network Training on Large Graphs in A Distributed Setting

2026-01-13T04:08:27Z

Optimizing Graph Neural Network Training on Large Graphs in A Distributed Setting Murzynowski, Philip Graph neural networks (GNNs) are an important class of methods for leveraging the information present in graph structures to perform various learning tasks. Distributed GNNs can improve the performance of GNN execution by dividing computation among multiple machines and scale to large graphs by partitioning graph features and the graph structure. Although distributed GNNs are able to achieve self-relative speedup, they are often slower than well-optimized code running on a single machine. For example, evaluation of the prevalent Distributed DGL system on graphs in the Open Graph Benchmark shows Distributed DGL can achieve speedup of over 2× when moving from one to four nodes, but execution of Distributed DGL on 4 nodes is 2× slower than a well-optimized GNN system, such as the SALIENT system, on a single machine. In my thesis, I argue that it is possible for a distributed GNN system to be both fast and scalable. Specifically, I show that it is possible to match the performance of well-optimized, non-distributed codes for GNN training and also achieve good scalability when running in the distributed setting. I present a system called Distributed SALIENT and motivate its design through profiling and identifying bottlenecks that arise in the distributed setting. Key components of Distributed SALIENT include the use of well-optimized code for local computations, pipelining of inter-machine communication, and a careful trade-off between data partitioning and partial replication. I evaluate Distributed SALIENT on the Open Graph Benchmark (OGB) and show that Distributed SALIENT achieves good speedup compared to SALIENT’s well-optimized single-node code while only using replication factors of roughly 5%. In fact, in experiments with training a 3-layer GraphSAGE model on the large OGB papers100M data set, Distributed SALIENT on 8 nodes is 8.6x faster than SALIENT on 1 node.

Atomistic Insights into Alloy Solidification using Machine-Learning Potentials

2026-01-13T03:35:39Z

Atomistic Insights into Alloy Solidification using Machine-Learning Potentials Cao, Yifan Alloy solidification is a critical process in materials design and manufacturing, as it governs the formation of microstructures that determines the mechanical, thermal, and chemical properties of materials. However, direct in situ observation remains extremely challenging due the need for high spatial and temporal resolution under elevated temperatures. On the theory side, solidification is a complex phenomenon often studied using phase-field simulations, which rely on empirically fitted parameters and simplified assumptions about interfacial kinetics, limiting their predictive capability. Capturing this process at the atomistic level can yield more fundamental insights, but is hindered by the need for interatomic models that are both accurate and computationally efficient across relevant timescales and length scales. To overcome these challenges, this thesis develops and applies machine-learning interatomic potentials (MLPs) that capture the chemical complexity of metallic alloys, providing a physically accurate and computationally efficient backbone for large-scale atomistic simulations of complex alloy solidification. We first address a foundational challenge in deploying MLPs: the systematic construction of robust and transferable training datasets. Using CrCoNi as a model system, we evaluate various strategies for training MLPs to capture chemical short-range order (SRO), a critical feature in high-entropy alloys, and its effects on materials quantities of relevance for mechanical properties, such as stacking-fault energy and phase stability. It is demonstrated that energy accuracy on test sets often does not correlate with accuracy in capturing material properties, which is fundamental in enabling large-scale atomistic simulations of metallic alloys with high physical fidelity. Based on this analysis we systematically derive design principles for the rational construction of MLPs that capture SRO in the crystal and liquid phases of alloys. The resulting MLPs are validated against experimental measurements on key thermophysical properties, including melting points, heat capacities, thermal expansion coefficients, and enthalpy of SRO formation, confirming their suitability for predictive simulations. With these validated potentials, we then investigate the evolution of SRO during rapid solidification processes. Our simulations reveal that alloy processing can lead to nonequilibrium steady states of SRO that differ qualitatively from any equilibrium configuration. We attribute this behavior to an inherent ordering bias introduced by nonequilibrium dynamics during solidification. These findings suggest that conventional manufacturing processes offer new opportunities to tailor alloy properties by accessing a broader spectrum of nonequilibrium SRO states, expanding the alloy design space beyond the equilibrium spectrum. Finally, we conduct predictive solidification simulations of chemically complex alloys across experimentally relevant growth rates (0.15–2 m/s) , alloy compositions, interface orientations, and undercooling levels. These simulations capture the dynamic build up of solute partitioning at the solid-liquid interface and reveal kinetics-dependent segregation patterns that deviate markedly from equilibrium predictions. The developed framework enables direct evaluation of key kinetic properties under realistic growth conditions, including interface mobility, liquid diffusivity, and solute trapping. Altogether, this thesis develops machine-learning potentials capable of capturing the chemical complexity of metallic alloys with near DFT-level accuracy, and establishes a framework for extracting key kinetic properties through predictive simulations of alloy solidification. When combined with emerging advances in continuum-scale modeling, these results lay the groundwork for truly multiscale investigations of alloy solidification, enabling DFT-level predictive capabilities at scales directly comparable to experimental alloy design and additive manufacturing processes.

Sequential Resource Allocation and Applications in Revenue Management

2026-01-13T03:36:41Z

Sequential Resource Allocation and Applications in Revenue Management Zhou, Zijie Sequential resource allocation is a fundamental problem in operations research, encompassing a wide range of applications where decisions must be made dynamically under uncertainty. This thesis develops new theoretical foundations, explores practical applications, and establishes evaluation methodologies for sequential resource allocation, with a focus on revenue management, robustness and fairness, and experiment design. On the theoretical side, this thesis advances the study of classical network revenue management, a long-standing challenge in dynamic resource allocation. We introduce the first LP-free algorithm, improving the regret bound from O(T ^1/2) to O(T ^3/8)—a significant step toward closing the gap between existing algorithms and the theoretical lower bound of O(1). Additionally, we enhance robustness in sequential resource allocation by developing algorithms that incorporate machine-learned advice, striking a balance between overly conservative worst-case models and overly optimistic stochastic assumptions. Furthermore, we integrate individual fairness into sequential decision-making, ensuring equitable resource allocation without compromising competitive performance. On the application side, we demonstrate the impact of sequential resource allocation in the hospitality management domain. Collaborated with Oracle Lab, we design an online upgrading mechanism that enables hotels to dynamically determine when and at what price to offer room upgrades. Additionally, we propose near-optimal, fast approximation algorithms for this mechanism, achieving a regret bound of O(logT), which is close to the natural lower bound of O(1). We also incorporate our upgrading algorithm to a hotel dataset, and improves more than 20% revenue in 2022. Finally, we introduce new methodologies for evaluating sequential decision-making policies, with a focus on online experiment design. Traditional A/B testing methods struggle with dynamically arriving data, leading to biased or inefficient experimental results. Our pigeonhole experimental design effectively reduces bias and outperforms several well-known experimental design policies, including matched pair design and completely randomized design, making it a more reliable approach for evaluating sequential decision-making strategies. By unifying theoretical insights, real-world applications, and online evaluation frameworks, this thesis contributes to the broader field of sequential resource allocation, providing fundamental advancements with practical implications across revenue management and experimental design.

Aspects of Moiré Quantum Matter

2026-01-13T03:36:10Z

Aspects of Moiré Quantum Matter Paul, Nisarga The advent of moiré quantum matter has newly unified disparate themes in modern condensed matter physics, chief among them band theory, correlations, and topology. This thesis investigates how the interplay between these foundational elements leads to novel electronic phenomena uniquely enabled by moiré superlattices. We focus on modulated Landau levels, which is one of the simplest settings with all three of band dispersion, correlations and topology, yet is rich enough to capture much of the interesting phenomena of moiré quantum matter. We characterize emergent quantum phases that are newly unlocked by the moiré regime. Specifically, we discuss directional localization, formation of Hall crystals with tunable Chern numbers, and novel fractional Chern insulator collective mode physics in the context of modulated Landau levels. We also show that a class of models comprising itinerant electrons strongly coupled to skyrmion-like magnetic textures, closely connected with moiré transition metal dichalcogenides in which the fractional quantum anomalous Hall effect was observed, can host flat Chern bands, emergent Landau levels, and zero-field non-Abelian topological order. This thesis provides a framework for the study of the essential features of moiré quantum matter and demonstrates how moiré systems provide unprecedented opportunities to explore, design, and manipulate strongly correlated topological quantum matter.

An Optimized Bayesian Analysis Framework for the KATRIN Experiment

2026-01-13T03:35:29Z

An Optimized Bayesian Analysis Framework for the KATRIN Experiment Xu, Weiran Neutrinos, which were originally predicted to be massless within the Standard Model of particle physics, have been confirmed to possess non-zero masses through the discovery of neutrino flavor oscillations. These oscillations precisely measure mass-squared splittings between neutrino mass eigenstates, establishing lower limits for the effective electron-neutrino mass at 0.009 eV for normal mass ordering and 0.050 eV for inverted mass ordering. However, the absolute neutrino mass scale remains a fundamental open question in both particle physics and cosmology. Precise spectroscopy of beta-decay spectrum provides a model-independent probe of the absolute neutrino mass via decay kinematics. The KArlsruhe TRItium Neutrino (KATRIN) experiment, utilizing a Magnetic Adiabatic Collimation and Electrostatic (MAC-E) filter spectrometer, sets the world's tightest upper limit of m_v < 0.45 eV (90% C.L.) based on its first five measurement campaigns. KATRIN is scheduled to complete its 1,000-day data-taking period by the end of 2025, targeting a final sensitivity of m_v < 0.3 eV}. Future improvements on neutrino mass measurements will depend on advances in differential detection techniques and the development of atomic tritium sources. This thesis presents an optimized modeling of the KATRIN beta spectrum and a comprehensive analysis of the first five measurement campaigns. An improved framework for computing the theoretical beta spectrum and the KATRIN response function is developed to address the complexities arising from the asymmetric field configurations in the main spectrometer. Benefiting from a computational speedup of four orders of magnitude and improved numerical stability, frequentist best-fit values for individual campaigns are reported, together with an upper limit on neutrino mass using the Lokhov-Tkachov confidence belt construction method. Parallel Bayesian analyses are conducted on the same dataset, yielding an independent and complementary statistical interpretation of the experimental results. Posterior distributions for the squared neutrino mass are sampled for each campaign under a flat prior on m²ᵥ using the parallel Stretch-Move algorithm, and are subsequently combined with a novel approach developed in this work to enhance computational efficiency. Convergence of each Markov chain is assessed through autocorrelation time analysis, and the robustness of the results is validated through cross-team comparison and consistency checks with profile likelihood. The Bayesian results reported here enable straightforward integration with constraints from oscillation measurements and cosmological observations, and the methodologies developed in this work are directly applicable to the final KATRIN dataset, providing a foundation for future neutrino mass analyses and searches for physics beyond the Standard Model.

Redox-Mediated Processes Toward Modular Electrochemical Systems

2026-01-13T03:36:30Z

Redox-Mediated Processes Toward Modular Electrochemical Systems Mallia, Christopher T. Electrochemical technologies offer an attractive path toward a sustainable future where conventional methods of storing energy or producing critical materials are increasingly coupled to renewable electricity generation. To enable such a future, it is imperative that we have strong foundational understanding of electrochemical reactions that are useful to our needs. Redox flow batteries (RFBs) have emerged as a promising architecture for large scale storage of electricity to bridge the gap when renewable generation is unavailable. These devices operate by storing charge in the form of redox-active species that are dissolved into an electrolyte, and subsequently passed through an electrochemical cell to either store or release electrical energy. An extension of the concept of RFBs toward more general applications is to use the dissolved redox-active species to drive a reaction with another material, either to increase the energy storage density through an electrochemically active charge-dense material, or to drive a useful chemical reaction. This extension is termed a redox-mediated (RM) process, and inherits many of the complexities and intricacies of conventional electrochemical technologies, specifically that of RFB-type devices. The subject of this thesis is the development of knowledge and techniques for studying RM processes toward practical embodiments. While technical implementations of this concept are still nascent, many promising early results have been found in devices that use redox-mediated reactions to store electricity. Despite this, progress is frequently hindered by a lack of foundational knowledge from which to ideate better systems, and techniques to experimentally determine underlying physics. First, I establish the development of the RM concept over the past years as primarily through proof-of-concept electrochemical reactors which mimic RFBs. Second, we establish that the underlying nature of some RM reactions can be quantified and understood through corrosion principles, which guide our intuition for selecting chemistries and operating conditions. Third, I demonstrate that the behavior of many desirable RM chemistries is intrinsically coupled to passivation phenomena, and that this must be accounted for in reaction design. Fourth and finally, I provide experimental and practical guidance for researchers in this field, coupled with the design of some apparatus and techniques useful for characterizing RM reactions in specific and electrochemical processes in general. This body of work is broadly intended to advance understanding of electrochemically active interfaces and enable technology concepts which promote a sustainable future.

Predictive Modeling of Chemical Reactivity for Sustainability

2026-01-13T03:35:43Z

Predictive Modeling of Chemical Reactivity for Sustainability Singhal, Avni Priya Predicting and controlling chemical reactivity is key to sustainable material and process design. However, modeling reactivity at scale remains challenging due to the computational demands of quantum chemical methods and the complexity of reaction mechanisms. This thesis explores how high-throughput computational approaches, rooted in quantum chemistry and enabled by automation, can be used to interrogate reactivity across large chemical spaces. We focus on two domains where reactivity governs process efficiency and sustainability: solvent-based carbon capture and polymer, specifically thermoset, manufacturing. We first investigate pi-conjugated heterocyclic nucleophiles as alternative carbon capture solvents to address the high regeneration energy and degradation rates of conventional amine-based systems. We combine synthetic template-based library enumeration, density functional theory (DFT), and machine learning models to evaluate binding energies, capture capacity, regeneration thermodynamics, and oxidative stability. Structure–property analysis reveals design strategies to enhance capture strength while balancing tradeoffs with desorption temperature and degradation resistance. We next focus on designing monomers for frontal ring-opening metathesis polymerization (FROMP), a polymerization mode that enables rapid, energy-efficient manufacturing of polymers. This self-propagating process harnesses exothermic reactions to sustain a polymerization front without continuous external heating, but it requires monomers with a finely tuned balance of thermodynamic and kinetic parameters. We develop a multi-level screening pipeline that integrates DFT-calculated properties with a reaction-diffusion model to predict front behavior directly from the atomistic structure of the monomer. We experimentally validate a preliminary pipeline, identifying a new class of FROMP-capable furan-benzyne monomers, and uncover additional candidates from unexplored chemical spaces that overcome limitations of known systems. Together, these studies demonstrate how high-throughput, mechanism-informed modeling can guide the discovery of molecules and materials that meet complex reactivity and performance criteria.

When Your Home Becomes a Panda Park: The Opportunity and Upheaval of China's Giant Panda National Park

2026-01-13T04:08:22Z

When Your Home Becomes a Panda Park: The Opportunity and Upheaval of China's Giant Panda National Park Zhao, Celina In December 2016, China launched the Giant Panda National Park (GPNP). A massive ecological initiative aimed at safeguarding its beloved national symbol and international icon of conservation, the park marked an unequivocal win for giant pandas. But for the 100,000 people already living in and around the borders, the outcome was not as clear. The GPNP seeks to establish a harmonious balance between biodiversity protection and human development. But the vast amount of land covered by the park means not all places are equally primed to achieve that goal. A handful of communities have been designated as exclusive entrance communities, with lavish funding to become the face of the national park. In others, a persistent question simmers: Are pandas more important than people? Central to this story is how individuals are adapting to and reimagining their futures. Rather than a binary of winners and losers, the GPNP has sparked a wide range of human responses -showing that the path to a sustainable future between people and pandas is far from black and white.

The Sequence Landscape of Bacterial Genes is Shaped by Long-Range mRNA Folding

2026-01-13T03:36:07Z

The Sequence Landscape of Bacterial Genes is Shaped by Long-Range mRNA Folding Gill, Manraj Singh An evolutionary selection for optimal expression of genes in regulatory networks has led to discernable sequence patterns in bacterial genomes observed in nature. Such patterns result from gene regulatory strategies that leverage sequence-dependent interactions with key cellular machineries and regulatory molecules. While numerous regulatory strategies that shape bacterial gene sequence have been characterized, predicting functional consequences from sequence alone remains challenging due to the sheer vastness of the possible sequence space. Moreover, the primary gene sequence encodes information on secondary and tertiary topologies that the molecules of the central dogma can fold into. Specifically, though local messenger RNA (mRNA) structures are known to regulate bacterial gene expression, the role of long-range mRNA folding remains unclear despite the predicted prevalence of such interactions across mRNAs. In bacteria, a major regulator of mRNA decay and translation rates is accessibility of the ribosome binding site (RBS) to the ribosome. Sequences in the mRNA’s 5´ untranslated region (UTR) complementary to the RBS can decrease gene expression by base pairing and occluding ribosomes from binding. To determine whether such antagonistic sequences are also the primary determinants of sequence choice along the rest of the mRNA transcript, we measured the effect of all possible 8-nucleotide substitutions (65,536 variants) on mRNA levels when placed in multiple positions along a bacterial transcript. We find that, while the vast majority of substitutions in the middle of genes negligibly affect RNA level, 8mers with complementarity to parts of the RBS exhibit the strongest effects by increasing RNA degradation rates up to 4-fold. RBS-complementary sequences also decrease translation initiation rates when placed in a coding sequence, and are able to occlude ribosome binding even when they are located hundreds of nucleotides away from the start codon. The inhibitory effect of such secondary structures on gene expression likely explains a strong selection against sequences complementary to conserved parts of RBSs throughout coding sequences of genes from diverse bacterial genomes, which we uncover through computational analysis. Together, this thesis reveals the widespread impact of RNA intramolecular interactions in vivo on both mRNA stability and translation and uncovers a key constraint on gene sequences.

Engineering Materials for Non-Compressible Torso Hemorrhage and Internal Bleeding

2026-01-13T03:35:37Z

Engineering Materials for Non-Compressible Torso Hemorrhage and Internal Bleeding Hong, Celestine Jia Huey Non-compressible torso hemorrhage (NCTH) and internal bleeding results in a significant number of preventable casualties worldwide among civilians and in the field. In particular, internal bleeding can only be diagnosed through changes in vital signs and then through imaging modalities that may only be available in a hospital setting. Over the past few decades, researchers in the field have sought to address these needs by developing hemostats that can rapidly expand, bind, or seal an exposed wound, or interact with wound-specific components when delivered intravenously to enhance preexisting hemostatic processes. The first part of this thesis investigates the effect of hemostatic nanoparticle size on their interactions with platelets. Small nanoparticles were observed to result in an increased percentage of specifically-bound single platelets under flow and intermediate nanoparticles were observed to result in the greatest degree of platelet recruitment to a platelet-collagen surface. Large nanoparticles were observed to result in the most nanoparticle mass bound to a surface, the shortest circulation time and retention, and the highest pulmonary accumulation. Ultimately, intermediate nanoparticles were shown to result in the most significant increase in survival relative to the saline control in a lethal inferior vena cava (IVC) injury model (84.6% vs 26.7%), as well as the greatest accumulation at the injured IVC relative to uninjured vessel controls. Subsequently, the intermediate nanoparticles from the prior study were functionalized with bio-orthogonal click-crosslinkable azide groups to achieve targeted crosslinking behavior. Commercial multiarm PEG functionalized with the corresponding clickable moiety, dibenzylcyclooctyne (DBCO), and DBCO-PEG-b-PLGA nanoparticles were delivered as the second part of this two-component system. This system was demonstrated to increase platelet recruitment, and decrease fibrin loss during plasminolysis in vitro. When challenged in a mouse liver resection model, the two-component system resulted in significantly increased survival relative to the nanoparticle-only system and higher accumulation in the remnant liver. Finally, a charge-inverting polymer was synthesized through controlled radical polymerization. The material was demonstrated to undergo rapid charge inversion when exposed to physiological pH, resulting in the near-complete lift-off within a minute of a layer-by-layer drug film into the dermis when coated on microneedles. This versatile release platform could be coated on wound dressings to facilitate the release of therapeutics to aid in healing, or other applications involving charged films. In sum, this thesis has investigated several new materials and assays for the treatment of traumatic hemorrhage, opening potential avenues for the development of more effective hemostats.

Advances in Nonconvex and Robust Optimization

2026-01-13T03:35:24Z

Advances in Nonconvex and Robust Optimization Koukouvinos, Theodoros Nonconvex optimization presents significant challenges, as identifying the global optimum is often difficult. This thesis introduces novel algorithms to find the exact solution of a broad class of nonconvex optimization problems. The thesis is structured into four parts. In Chapter 2, we propose a novel method for solving nonconvex optimization problems, in which the nonconvex components are sums of linear times convex (SLC) functions. We introduce a new technique, called the Reformulation-Perspectification Technique (RPT), to obtain a convex approximation of the original nonconvex optimization problem. We then incorporate RPT within branch and bound to obtain the global optimal solution of the nonconvex optimization problem. By using the RPT, we obtain a convex relaxation by forming the perspective of each convex function and linearizing all product terms with newly introduced variables. To further tighten the approximation, we pairwise multiply constraints. Therefore, in Chapter 3, we analyze all possibilities of multiplying conic constraints, a very wide class of constraints. Further, we delineate methods for deriving new, valid linear and second-order cone inequalities for pairwise constraint multiplications involving the power cone and exponential cone, thereby enhancing the strength of the approximation. In Chapter 4, we address nonconvex optimization problems that involve polynomials. We derive valid SLC decompositions for polynomials, in which the linear functions are inequalities of the feasible region and the convex functions are quadratics. We prove the existence of such SLC decompositions for arbitrary degree polynomials. Further, out of the many possible SLC decompositions, we obtain the one that results in the tightest lower bound. Finally, in the numerical experiments we show that our method often outperforms state-of-the-art approaches for polynomial optimization. In Chapter 5, we propose a robust optimization framework that immunizes some of the central linear algebra problems in the presence of data uncertainty. Namely, we formulate linear systems, matrix inversion, eigenvalues-eigenvectors and matrix factorization under uncertainty, as robust optimization problems using appropriate descriptions of uncertainty. We show that for both linear systems and matrix inversion, the robust approach leads to more accurate solutions than the nominal, in the case of nearly singular matrices.

Fabricating and Tailoring Halide Perovskites for Photovoltaic Applications

2026-01-13T03:36:12Z

Fabricating and Tailoring Halide Perovskites for Photovoltaic Applications Kadosh Zhitomirsky, Tamar Green energy is a contemporary global concern, and research of materials for solar energy harvesting is the heart of potential solutions for the energy crisis. Halide perovskites are leading candidates to replace silicon in next generation solar cells. This thesis focuses on halide perovskite materials, aiming to understand their structure, electronic and ionic properties and photo-activity; and to re-direct their fabrication techniques to address global market needs and requirements. In this work we developed alternative, vapor-based fabrication techniques, based on manufacturing-compatible, safe, rapid and scalable processes, that have the potential to improve material stability and efficiency. Vapor Transport Deposition (VTD) is investigated as a promising fabrication method for thin film halide perovskites and beyond. We explored the deposition parameter space and elucidated relationships and trends regarding composition, structure and deposition rate. We examined the morphology, crystal phase formation, optical and electrical properties, and finally the performance of the deposited films when incorporated into solar cells. We begin by exemplifying the viability of vapor transport co-deposition in fabricating active perovskite films, utilizing methylammonium lead iodide (MAPbI3) as a simplified model system. We then design an improved version of the vapor transport deposition system and transition to the more technologically attractive perovskite composition formamidinium lead iodide (FAPbI3). Learning from previous attempts to fabricate this material, we developed a novel technique that we call Hybrid two-step vapor-solution deposition in which we use VTD to deposit the inorganic 4 precursor, not readily dissolved in industry acceptable solvents, and then react it with a solution of the organic precursors dissolved in a benign solvent. This technique allowed us to fabricate functioning FAPbI3 based solar cell devices, in a safe, fast-paced, scalable and manufacturing compatible fashion. The deposition rate is significantly influenced by chamber pressure and source temperature, and by controlling all deposition parameters, we systematically reached rates of up to 1200 nm/min, that is orders of magnitude faster than current comparable techniques. We found the technique to be reproducible, yielding 13% efficient devices, with champion efficiencies of up to 15.3%. Based on the proposed novel fabrication process, we believe it offers an avenue for further improvement in solar cell stability and efficiency. CsPbBr3, a fully inorganic halide perovskite, also shows great promise as a photo and gamma ray detector and like the other halide perovskites is known to support halide ion conductivity that contributes to device instability and reduced sensitivity to irradiation. We choose this as a model system to apply concepts from defect chemistry and demonstrate the ability to measure and manipulate the ionic conductivity in the material by stoichiometry control and doping.

Membrane protein conformational dynamics and ligand-binding interactions in bacterial glycoconjugate biosynthesis

2026-01-13T03:35:41Z

Membrane protein conformational dynamics and ligand-binding interactions in bacterial glycoconjugate biosynthesis Higinbotham, Hugh Membrane associated proteins are an essential component of the complex biochemistry that is carried out at the membrane interface and perform essential functions for cellular life. Biophysical characterization of protein structure-function relationships faces a unique set of challenges due to the constraints of phospholipid bilayer chemistry and geometry. Advances in x-ray crystallography and cryo electron microscopy have made progress in this regard, but dynamic structural features remain difficult to study. Small membrane proteins, such as those responsible for bacterial glycosylation, remain challenging to structurally characterize at all. Bacterial glycan synthesis pathways are essential for cell function yet highly variable between strains, making them promising systems for targeted antibiotic development. Many pathways have initiating SmPGTs that show incredible specificity for minute changes in glycan chemistry despite being small enough to streamline many computational methods, which makes them ideal model systems for developing multidisciplinary strategies to study membrane protein dynamics. This thesis presents a strategy that employs structural bioinformatics in Chapter 2, molecular dynamics simulation (MD) in Chapter 3, and single-molecule FRET microscopy (smFRET) in Chapter 4 to observe the ligand-dependent conformational dynamics of integral membrane proteins in situ. It focuses on representative members of the small monotopic phosphoglycosyl transferase (SmPGT) superfamily, which catalyze transfer of a phosphosugar from a soluble nucleotide-sugar donor to a membrane-embedded polyprenol phosphate acceptor in the initiating step of glycoconjugate biosynthesis in prokaryotes. The pipeline is employed to confirm the role of SmPGT conformational dynamics in substrate binding and informs the design of non-hydrolyzable substratemimetic inhibitors. Chapter 5 further sets the stage for the use of structural bioinformatics and molecular simulation to characterize subsequent glycosyl transferase (GT) enzymes down pathway and presents initial results characterizing inter-protein cooperative interactions. The integrated approach to incorporate computational and experimental characterization methods has significantly contributed to the understanding of SmPGT structure-function relationships and opened up new directions of inquiry into specific PGTligand interactions, the development of new inhibitory compounds, and the role of interprotein interactions in bacterial glycan synthesis.

Functional Genomic and Image-Based Screening Approaches for Probing Host-Pathogen Interactions

2026-01-13T03:35:32Z

Functional Genomic and Image-Based Screening Approaches for Probing Host-Pathogen Interactions Carlson, Rebecca J. Host-pathogen interactions represent a complex interplay between hosts and pathogens that can evolve over millions of years. Interactions between bacteria or viruses and human cells, and the resulting evolved antipathogenic signaling pathways, are processes responsible for pathologies ranging from infectious diseases to autoimmune conditions and cancer. In addition, engineered designs inspired by pathogen interactions with hosts are increasingly being used to both treat and diagnose many pathologies that need not originate from infection with a pathogen. Therefore, it is critical to build and deploy scalable tools to better understand host-pathogen dynamics in order to both better treat conditions where pathogens or antipathogenic signaling contribute directly to disease pathology as well as to engineer new treatments to address a broader range of disease states. In this thesis, I describe approaches to leverage functional genomics and image-based screening to perturb and profile host-pathogen interactions, including responses to two RNA viruses, Sendai virus and Ebola virus. These provide case studies highlighting the utility of high-content image-based screening for revealing new genes regulating predefined phenotypes of interest as well as for generating single-cell imaging profiles that can be used to infer new genetic functions and phenotypic states directly from screening data without a priori specification. I also highlight an example of a genetic screen that revealed a robust negative result, leading to hypothesis and validation of a novel function of the STING protein as a proton channel.

Measurement of the B0 → ρ(770)0γ branching fraction

2026-03-08T03:39:33Z

Measurement of the B0 → ρ(770)0γ branching fraction Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J. The ratio between the branching fractions of the B0 → ρ(770)0γ and B0 → K*(892)0γ decays is measured with proton-proton collision data collected by the LHCb experiment at centre-of-mass energies of 7, 8, and 13 TeV, corresponding to an integrated luminosity of 9 fb−1. The measured value is B B 0 → ρ 770 0 γ B B 0 → K ∗ 892 0 γ = 0.0189 ± 0.0007 ± 0.0005 , where the first uncertainty is statistical and the second systematic. The branching fraction for B0 → ρ(770)0γ decays is hence obtained as B B 0 → ρ 770 0 γ = 7.9 ± 0.3 ± 0.2 ± 0.2 × 10 − 7 , where the last uncertainty is due to the branching fraction of the normalisation mode. This result assumes that both the ρ(770)0 and K*(892)0 decays saturate the dihadron mass spectra considered in the analysis. It is consistent with the current world-average value and by far the most precise measurement to date.

A Beamdump facility at Jefferson Lab

2026-03-08T03:39:32Z

A Beamdump facility at Jefferson Lab Achenbach, Patrick; Afanasev, Andrei; Ambrozewicz, Pawel; Ashkenazi, Adi; Banerjee, Dipanwita; Battaglieri, Marco; Benesch, Jay; Bondí, Mariangela; Brindza, Paul; Camsonne, Alexandre; Christy, Eric M.; Cline, Ethan W.; Cuevas, Chris; Dilling, Jens; Doria, Luca; Fegan, Stuart; Filippini, Marco; Fulci, Antonino; Giovannella, Simona; Grazzi, Stefano The potential of the intense secondary muon, neutrino, and (hypothetical) light dark matter beams at the Thomas Jefferson National Accelerator Facility (Jefferson Lab) is explored. These are produced in the high-power dumps with high-current electron beams. Light dark matter searches with the approved Beam Dump eXperiment (BDX) are driving the realization of a new underground vault behind Hall A that could be extended to a Beamdump Facility with little additional installations. High-energy muons created via the Bethe–Heitler process uniquely do not proceed through the more common pion production and decay channels. Several possible muon physics applications are highlighted. Neutrino detector technologies and experiments suitable for a beamdump facility are outlined.

Linking Chemical Phase and Mechanical Properties to Evaluate the Use of Millimeter-Wave Induced Vitrified Basalt in Enhanced Geothermal Systems

2026-01-10T03:08:21Z

Linking Chemical Phase and Mechanical Properties to Evaluate the Use of Millimeter-Wave Induced Vitrified Basalt in Enhanced Geothermal Systems Meltzer, Eve R.; Stefaniuk, Damian; Einstein, Herbert H. Extraction of geothermal energy from Earth’s heat could significantly contribute to long-term energy needs, yet the current geothermal drilling process faces significant technical limitations. A promising advancement in enhanced geothermal systems is the use of a millimeter-wave (MMW) gyrotron, which enables faster and more efficient drilling. The MMW drilling process offers two key advantages over traditional methods: (1) rock is melted rather than mechanically drilled, leading to faster well hole advancement, and (2) the molten rock solidifies into a vitrified wall, eliminating the need for additional casing materials. This integrated drilling and casing method has the potential to save costs, time, and materials. This paper examines the strength, structural integrity, and microscale mechanical and chemical properties of the vitrified material formed during the mm-wave process, focusing on basalt as the test material. By employing a suite of experimental and analytical characterization techniques, this study aims to provide a comprehensive comparison of the structural, mechanical, and chemical changes in the rock before and after melting, offering insights into the effectiveness and implications of mm-wave drilling for enhanced geothermal systems. Highlights There is a clear change of phase between the basalt, the transition zone, and the melt, due to mm-wave exposure. The region exposed to mm-waves is completely vitrified, while there is partial melting of minerals within the zone right outside of the mm-wave beam. The transition zone created from mm-waves poses high risk to wellbore stability due to its variable mechanical strength and chemical composition. A better understanding of this new material can be achieved by overlaying a series of chemical and mechanical characterization data.

Reimagining Commercial Health Insurance in India: A System-Dynamics Approach to Complex Stakeholder Incentives and Policy Outcomes

2026-01-10T03:08:27Z

Reimagining Commercial Health Insurance in India: A System-Dynamics Approach to Complex Stakeholder Incentives and Policy Outcomes Mor, Nachiket; Gupta, Aakriti; Roy, Rahul Most low- and middle-income governments are unwilling and unable to adequately fund their health systems using tax resources. Despite this route’s popularity in public discourse, it is neither a feasible nor a desirable route for financing Universal Health Coverage (UHC), given competing public finance priorities and limited citizen demand, among other challenges. It thus becomes essential to study the underlying mechanisms behind commercial health insurance and offer citizens the best possible product, which ensures that they not only receive a high degree of protection from health and financial risk on a sustained basis but also find reasonable access and support to improve their health outcomes. In this paper, we build a system-dynamics model that simulates the aggregate behavior of the Indian health-insurance industry, with interacting feedbacks between decisions by stakeholders such as the insurer, healthy and chronically ill populations, and the regulator to outcomes like insurance penetration among segments, overall coverage, health status over the long run, a mechanism of market-discovered premium, and financial viability of the private insurer. We then investigate policy choices and scenarios to explore contrast between design choices and ideal or targeted states of this market, such as a market with 100% enrollment, risk selection by insurers, group insurance models, and managed care, and study the impact on our outcomes of interest, i.e., insurance penetration and pricing, the financial sustainability of the insurers, and the population’s health outcomes. The simulations show that even while insurers and the different population segments optimize for their respective near-term objectives, the best outcomes for all come from the managed-care policy option, which has greater insurance penetration, lower premiums, higher profitability for insurers, and better long-term health outcomes. All other choices and scenarios yield suboptimal, imbalanced systemic outcomes. We thus recommend managed care as a desirable policy alternative for low-income countries intending to improve UHC by leveraging commercial health insurance.

A Symbiotic Digital Environment Framework for Industry 4.0 and 5.0: Enhancing Lifecycle Circularity

2026-01-10T03:08:29Z

A Symbiotic Digital Environment Framework for Industry 4.0 and 5.0: Enhancing Lifecycle Circularity Ponce, Pedro; Maldonado-Romo, Javier; Anthony, Brian W.; Bradley, Russel; Montesinos, Luis This paper introduces a Symbiotic Digital Environment Framework (SDEF) that integrates Human Digital Twins (HDTs) and Machine Digital Twins (MDTs) to advance lifecycle circularity across all stages of the CADMID model (i.e., Concept, Assessment, Design, Manufacture, In-Service, and Disposal). Unlike existing frameworks that address either digital twins or sustainability in isolation, SDEF establishes a bidirectional adaptive system where human, machine, and environmental digital entities continuously interact to co-optimize performance, resource efficiency, and well-being. The framework’s novelty lies in unifying human-centric adaptability (via HDTs) with circular economy principles to enable real-time symbiosis between industrial processes and their operators. Predictive analytics, immersive simulation, and continuous feedback loops dynamically adjust production parameters based on operator states and environmental conditions, extending asset lifespan while minimizing waste. Two simulation-based scenarios in VR using synthetic data demonstrate the framework’s capacity to integrate circularity metrics (material throughput, energy efficiency, remanufacturability index) with human-machine interaction variables in virtual manufacturing environments. SDEF bridges Industry 4.0’s automation capabilities and Industry 5.0’s human-centric vision, offering a scalable pathway toward sustainable and resilient industrial ecosystems by closing the loop between physical and digital realms.

Development of a Test Bed to Investigate Wetting Behaviours of High-Temperature Heavy Liquid Metals for Advanced Nuclear Applications

2026-01-10T03:08:39Z

Development of a Test Bed to Investigate Wetting Behaviours of High-Temperature Heavy Liquid Metals for Advanced Nuclear Applications Saraswat, Abhishek; Bhattacharyay, Rajendraprasad; Chaudhuri, Paritosh; Gedupudi, Sateesh Specifically engineered heavy liquid metals are proposed as candidate coolants and tritium breeders for advanced nuclear applications. Understanding the wetting behaviours of these liquids on relevant substrate configurations is crucial to tackle the challenges associated with corrosion protection and flow diagnostics development. However, detailed investigations are scarce in the literature. In this experimental study, an apparatus is designed to measure contact angles of different liquid metals over a mirror-polished horizontal SS-304 substrate. This paper presents design aspects of the developed test facility, as well as initial results obtained using direct imaging and the Low-Bond Axisymmetric Drop Shape Analysis algorithm-based image processing technique. Methodological validation is achieved through surrogate liquids/liquid metals (H2O, Hg, Ga, GaInSn), prior to taking measurements from molten lead (Pb) droplets at 425 °C. Estimated contact angles obtained using the two techniques lie within ±10% deviation. Towards the end, the paper lays out plans for future upgrades for studies of wetting behaviours of molten Pb/Pb alloys on substrates with relevant surface properties, including bare P-91 and reduced-activation ferritic–martensitic steels, along with Al2O3/Er2O3-coated versions of these materials, to generate a database for Gen-IV fission reactors and fusion power plants.

Future Circular Collider Feasibility Study Report

2026-01-10T03:08:37Z

Future Circular Collider Feasibility Study Report Benedikt, M.; Zimmermann, F.; Auchmann, B.; Bartmann, W.; Burnet, J. P.; Carli, C.; Chancé, A.; Craievich, P.; Giovannozzi, M.; Grojean, C.; Gutleber, J.; Hanke, K.; Henriques, André; Janot, P.; Lourenço, C.; Mangano, M.; Otto, T.; Poole, J.; Rajagopalan, S.; Raubenheimer, T. Volume 1 of the FCC Feasibility Report presents an overview of the physics case, experimental programme, and detector concepts for the Future Circular Collider (FCC). This volume outlines how FCC would address some of the most profound open questions in particle physics, from precision studies of the Higgs and EW bosons and of the top quark, to the exploration of physics beyond the Standard Model. The report reviews the experimental opportunities offered by the staged implementation of FCC, beginning with an electron-positron collider (FCC-ee), operating at several centre-of-mass energies, followed by a hadron collider (FCC-hh). Benchmark examples are given of the expected physics performance, in terms of precision and sensitivity to new phenomena, of each collider stage. Detector requirements and conceptual designs for FCC-ee experiments are discussed, as are the specific demands that the physics programme imposes on the accelerator in the domains of the calibration of the collision energy, and the interface region between the accelerator and the detector. The report also highlights advances in detector, software and computing technologies, as well as the theoretical tools/reconstruction techniques that will enable the precision measurements and discovery potential of the FCC experimental programme. The content and structure of this report are guided by the scope and priorities defined in the mandate of the FCC Feasibility Study. It is therefore not intended to serve as an exhaustive review of the full physics potential of FCC. Several topics, already covered in earlier reports such as the FCC CDR, are not reiterated here or are addressed only briefly, in alignment with the study’s focus. This volume reflects the outcome of a global collaborative effort involving hundreds of scientists and institutions, aided by a dedicated community-building coordination, and provides a targeted assessment of the scientific opportunities and experimental foundations of the FCC programme.

Unregulated Vertical Urban Growth Alters Microclimate: Coupling Building-Scale Digital Surface Models with High-Resolution Microclimate Simulations

2026-01-10T03:08:51Z

Unregulated Vertical Urban Growth Alters Microclimate: Coupling Building-Scale Digital Surface Models with High-Resolution Microclimate Simulations Falcão, Jonatas Goulart Marinho; Furtado, Luiz Felipe de Almeida; Barbosa, Gisele Silva; Teixeira Coelho, Luiz Carlos Rio de Janeiro’s favelas house over 20% of the city’s population in just 5% of its territory, with Rio das Pedras emerging as a critical case study: ranking as Brazil’s fifth most populous favela and its most vertically intensified. This study quantifies how uncontrolled vertical growth in informal settlements disrupts microclimate dynamics, directly impacting thermal comfort. Using high-resolution geospatial analytics, we integrated digital surface models (DSMs) derived from LiDAR and photogrammetric data (2013, 2019, and 2024) with microclimatic simulations to assess urban morphology changes and their thermal effects. A spatiotemporal cadastral analysis tracked vertical expansion (new floors) and demolition patterns, while ENVI-met simulations mapped air temperature anomalies across decadal scenarios. Results reveal two key findings: (1) rapid, unregulated construction has significantly altered local airflow and surface energy balance, exacerbating the urban heat island (UHI) effect; (2) microclimatic simulations consistently recorded elevated temperatures, with the most pronounced impacts in densely built zones. These findings underscore the need for public policies to mitigate such negative effects observed in informal settlement areas.

The Symbolic Bridge: A Monograph on Niela Miller Life’s Work

2026-01-10T03:05:14Z

The Symbolic Bridge: A Monograph on Niela Miller Life’s Work Labrune, Jean-Baptiste This monograph examines the interdisciplinary contributions of Niela Miller, specifically her development of Symbolic Modeling (SymMod) and its role in bridging humanistic psychology with technological innovation. Situated within the MIT Media Lab’s framework of unconventional synthesis, the study explores how Miller’s focus on tacit, pre-verbal, and intuitive knowledge complements data-driven paradigms. The research archives her transition from bodily-based psychological practices to pioneering work in virtual learning environments and "metaliteracy." By analyzing Miller’s methodology for unlocking human potential through symbolic expression, this document provides a formal architecture for integrating and extending human consciousness into the design of future technologies. In an era dominated by code and explicit data, the work of Niela Miller serves as a vital reminder that human innovation is rooted in the intuitive and the symbolic. This document offers an immersive look into Miller’s lifelong exploration of the "inner landscape," tracing her journey from foundational humanistic psychology to her visionary use of virtual spaces as laboratories for authentic interaction. Through the lens of the MIT Media Lab, we explore her Symbolic Modeling methodology—a replicable system designed to translate deep, pre-verbal insights into tangible creation. Whether she is utilizing bodily performance to map the psyche or defining new frontiers of digital literacy, Miller’s work challenges the boundary between the human experience and technological advancement. This is more than an archive; it is a celebration of the belief that our most profound breakthroughs come from what we can symbolically express but not always logically articulate.

Striking a Pose: DIY Computer Vision Sensor Kit to Measure Public Life Using Pose Estimation Enhanced Action Recognition Model

2026-01-09T06:25:43Z

Striking a Pose: DIY Computer Vision Sensor Kit to Measure Public Life Using Pose Estimation Enhanced Action Recognition Model Williams, Sarah; Kang, Minwook Observing and measuring public life is essential for designing inclusive, vibrant, and climate-resilient public spaces. While urban planners have traditionally relied on manual observation, recent advances in open-source Computer Vision (CV) now enable automated analysis. However, most CV sensors in urban studies focus on transportation analysis, offering limited insight into nuanced human behaviors such as sitting or socializing. This limitation stems in part from the challenges CV algorithms face in detecting subtle activities within public spaces. This study introduces the Public Life Sensor Kit (PLSK), an open-source, do-it-yourself system that integrates a GoPro camera with an NVIDIA Jetson edge device, and evaluates whether pose estimation-enhanced CV models can improve the detection of fine-grained public life behaviors, such as sitting and social interaction. The PLSK was deployed during a public space intervention project in Sydney, Australia. The resulting data were measured against data collected from the Vivacity sensor, a commercial transportation-focused CV system, and traditional human observation. The results show that the PLSK outperforms the commercial sensor in detecting and classifying key public life activities, including pedestrian traffic, sitting, and socializing. These findings highlight the potential of the PLSK to support ethically collected and behavior-rich public space analysis and advocate for its adoption in next-generation urban sensing technologies.

Pretreatment of Mice with 830 nm Light Enhances Endurance During Acute Exercise

2026-01-09T06:25:47Z

Pretreatment of Mice with 830 nm Light Enhances Endurance During Acute Exercise Cheema, Nashwa; Ghag, Namrata; Pham, Linh; Wise, Emma; Fuchs, Christiane; Anderson, Rox; Tam, Joshua Light therapy has been shown to produce several beneficial physiological effects in a wide range of tissues. The musculoskeletal system can be irradiated with deeply penetrating wavelengths in near infrared (NIR) regions. Photobiomodulation therapy (PBMT) reduces pain and inflammation and enhances physical performance. However, the mechanism(s) of cellular responses to PBMT in muscle is not clearly understood. Therefore, the goal of this study is to improve our understanding of the mechanism(s) of action of PBMT effects in exercised and sedentary muscle. In sedentary mice, PBMT using a wavelength of 830 nm increased the gene expression for muscle tissue development, including cFos, which is critical for activating interstitial and satellite cells that repair muscle. Immunostaining for cFOS expression confirmed an increase in the number of activated cells in PBMT-treated muscle. We observed that PBMT-treated mice showed increased performance on the treadmill, reduced muscle fiber damage, and altered mitochondrial structure. RNA sequencing from fatigued TA tissue suggested that PBMT treatment increased the gene expression of tissue regeneration and remodeling, suggesting tissue adaptation and muscle repair after exercise with PBMT. In conclusion, our study suggests that the 830 nm wavelength may have altered the muscle by activating regenerative genes that protect the tissue from exercise-induced cellular stress.

Integrating Physiologic Assessment into Virtual Reality-Based Pediatric Pain Intervention: A Feasibility Study

2026-01-09T06:25:49Z

Integrating Physiologic Assessment into Virtual Reality-Based Pediatric Pain Intervention: A Feasibility Study Marwah, Harsheen; Moldovanu, Stefania R.; Reks, Talis; Anthony, Brian; Logan, Deirdre E. This feasibility study explored the integration of physiological monitoring into a virtual reality (VR) intervention for pediatric pain management. The goal of this study is to identify a feasible strategy for collecting physiologic data in the context of a VR intervention currently being developed for youth with chronic pain. We assess the potential of Cognitive Load (CL)—derived from heart rate and pupillometry/eye-tracking data—as a marker of arousal and user engagement in a VR simulation to promote school functioning in youth with chronic pain. The HP Reverb G2 Omnicept headset and Polar H10 heart-rate sensor were utilized. The Child Presence Questionnaire (CPQ) assessed participants’ self-reported immersion and engagement. Data collection focused on feasibility and utility of physiologic data in assessing arousal and correlations with self-reported experience. Nine participants engaged in the simulation, with eight yielding complete data. The simulation and headset were well tolerated. CPQ Transportation subscale showed trend-level correlation with mean CL. Due to small sample and feasibility focus, individual-level results were examined. Combining multiple physiologic markers into a construct like CL is intriguing, but data interpretability was limited. Pupillometry and related metrics show promise as feasible markers of engagement and arousal for VR-based intervention but require appropriate expertise to fully interpret. The study found that integration of physiologic monitoring is feasible, but further work is needed to standardize metrics and identify the most useful and user-friendly markers.

Two Decades of CARICOMP Mangrove Monitoring (1992–2013) Reveal Variability in Tree Structure and Productivity of Rhizophora mangle Across the Wider Caribbean

2026-01-09T06:25:50Z

Two Decades of CARICOMP Mangrove Monitoring (1992–2013) Reveal Variability in Tree Structure and Productivity of Rhizophora mangle Across the Wider Caribbean Kjerfve, Björn; Oxenford, Hazel A.; Collin, Rachel; Pestana, Inácio Abreu; Samper-Villarreal, Jimena; Medina-Gómez, Israel; Cortés, Jorge; Smith, Struan R.; Koltes, Karen; Feller, Ilka C.; Bastidas, Carolina; Juman, Rahanna; Geraldes, Francisco X.; Filippo, Alessandro; Varela, Ramon; McCoy, Croy; Garzón-Ferreira, Jaime; Polanía, Jaime; Capelo, Juan C.; Ogden, John The Caribbean Coastal Marine Productivity (CARICOMP) program was conceptualized in 1985 to monitor coral reefs, seagrass beds, and mangrove forests at multiple sites across the wider Caribbean. Mangrove monitoring was focused on the dominant Caribbean species, red mangrove (Rhizophora mangle). Forest structure and productivity were monitored at 21 sites (18 countries) across different geomorphological settings, from tropical to subtropical mainland and island systems. Here, we provide the key findings from the CARICOMP mangrove data collected, mostly from 1992 to 2013, to assess spatial and temporal variability across the region. Red mangrove above-ground biomass averaged 190 t ha−1 (far higher than previously reported) but ranged widely across sites from 33 to 590 t ha−1, equating to an average above-ground ‘blue carbon’ of 84 t ha−1 (range 15–260 t ha−1). Tree density averaged 3237 trees ha−1, tree basal area averaged 19.7 m2 ha−1, tree height averaged 6.1 ± 2.8 m, and seedling density varied from 1.2 to 74 seedlings m−2 across the sites. Among the environmental factors that influence mangroves, local temperature and rainfall explained 48% of the variability in measured tree structure parameters. Annual litterfall, as a proxy for productivity, measured on average 1.24 ± 0.70 kg m−2 yr−1, with 60% of the total litterfall composed of leaves. Litterfall varied seasonally by 42%. No relationship was apparent between litterfall and seasonal ocean–atmosphere climate indices (ONI and AMM). With exception of the three most southwesterly CARICOMP sites, hurricanes and tropical storms impacted the mangrove sites repeatedly, resulting in considerable damage. A direct strike by a category-4 hurricane in 1998 in Dominican Republic killed 67% of the red mangrove trees, lowered above-ground biomass by 91%, basal area by 89%, litterfall by 63%, and resulted in the subsequent growth of many tall and thin saplings, totally changing the structure of the forest ecosystem in the first few years after the hurricane. In comparing mangrove systems, major differences may be explained by time elapsed since the last destructive event (hurricane) affecting each site. This highlights the fact that despite an increasing focus on preserving these valuable ecosystems, they are still highly vulnerable to natural hazards and likely to face a poor outcome under ongoing climate change.

Thallium(I) Uptake and Accumulation by Wheat and Rice Plants

2026-01-09T06:25:52Z

Thallium(I) Uptake and Accumulation by Wheat and Rice Plants Yang, Puu-Tai; Chang, Hsin-Fang; Huang, Liang-Sin; Chuang, Tsung-Ju; Wang, Shan-Li Thallium (Tl) is a highly toxic trace metal of increasing concern in agricultural soils. This study investigated the uptake, accumulation, and tissue-level distribution of Tl(I) in rice (Oryza sativa L.) and wheat (Triticum aestivum L.) grown in three agricultural soils differing in soil pH and texture. In the seedling pot experiment (0–100 mg kg−1 soil Tl), plant Tl concentrations increased dose-dependently, and were at least an order of magnitude lower in the alkaline soil than in the acidic soils. Bioaccumulation factors of roots and shoots generally exceeded unity and declined with increasing Tl dose in acidic soils, consistent with uptake saturation and physiological stress at high exposure. To elucidate how soil Tl speciation and pH regulate Tl availability, X-ray absorption spectroscopy (XAS) was used; it showed that Tl(I)—sorbed on illite was the predominant species in all soils (89–95%), with a minor fraction (5–11%) associated with non-specific adsorption. In maturity pots (5 mg kg−1 soil Tl), both crops grown in the moderately acidic, coarse-textured soil translocated a small fraction of absorbed Tl to grains, with wheat and rice containing 0.24 and 0.10 mg kg−1 Tl, respectively. Comparatively, plants in the more acidic soil failed to reach maturity, and grain Tl was not detected in the alkaline soil. LA-ICP-MS mapping revealed Tl enrichment in the bran and embryo of rice and in the crease, bran, and embryo of wheat, indicating that unpolished grains may pose higher dietary exposure risks than polished products. Overall, these findings demonstrate the key roles of soil pH and mineral composition in governing soil Tl availability and plant Tl uptake, whereas plant transport processes regulate grain Tl loading. In the absence of food-safety standards for Tl, the results of this study underscore the need to better understand and mitigate Tl transfer from contaminated soils into human food chains via cereal crops.

A study of the field control operation of railway motors : a thesis

2026-01-07T03:24:36Z

A study of the field control operation of railway motors : a thesis Davis, Stanley W. (Stanley Whitcomb) Thesis: B.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1925; Includes bibliographical references (leaf 91).

Reviewing I.S. : how to handle legacy systems?

2026-01-07T03:23:47Z

Reviewing I.S. : how to handle legacy systems? Orlando, Ricardo, 1966- Thesis: S.M.M.O.T., Massachusetts Institute of Technology, Sloan School of Management, Management of Technology Program, 1999; Includes bibliographical references (leaves 100-106).

The effects of changing economic conditions on energy costs in stainless-steel clad pressurized water reactors

2026-01-07T03:23:33Z

The effects of changing economic conditions on energy costs in stainless-steel clad pressurized water reactors Trapp, Donald L. Thesis: M.S., Massachusetts Institute of Technology, Department of Nuclear Engineering, 1962; Appendix contains numerous pamphlets.; Includes bibliographical references (leaves 135-136).

The politics of metropolitan transportation.

2026-01-07T03:04:18Z

The politics of metropolitan transportation. Colcord, Frank Carlton. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics and Social Science, 1964

The design of a control system for the terminal phase of a satellite rendezvous

2026-01-07T03:23:50Z

The design of a control system for the terminal phase of a satellite rendezvous Hollister, Walter M., 1930- Thesis: M.S., Massachusetts Institute of Technology, Department of Aeronautics and Astronautics, 1959; Includes bibliographical references (leaf 47).

Compliance in a gyroscope gimbal

2026-01-07T03:24:32Z

Compliance in a gyroscope gimbal Graham, James William. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1958

On hypergraphs and hypergeometries.

2026-01-07T03:04:03Z

On hypergraphs and hypergeometries. Helgason, Thorkell. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Mathematics, 1971; Vita.; Bibliography: leaves 158-159.

Noise analysis of circuit models representing maser operation.

2026-01-07T03:23:54Z

Noise analysis of circuit models representing maser operation. Hempstead, Robert Douglas. Thesis: M.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1965; Bibliography: leaves 106-108.

Cocoa in the Ghanaian economy.

2026-01-07T03:03:55Z

Cocoa in the Ghanaian economy. Bateman, Merril Joseph. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Economics, 1965

Even denominator quantum numbers and termination of the fractional series in the fractional quantum hall effect

2026-01-07T03:04:22Z

Even denominator quantum numbers and termination of the fractional series in the fractional quantum hall effect Willett, Robert L. (Robert Lee) Thesis: Ph. D., Massachusetts Institute of Technology, Department of Physics, 1989; Includes bibliographical references (leaves 6-7).

Improving railroad terminal control systems : a case study of Southern Railway's Brosnan yard

2026-01-07T03:23:43Z

Improving railroad terminal control systems : a case study of Southern Railway's Brosnan yard Ferguson, William Lloyd. Thesis: M.S., Massachusetts Institute of Technology, Department of Civil Engineering, 1979; Bibliography: leaves 194-195.

Investigations in the theory of quantum corrections to classical solutions of the Yang-Mills equations

2026-01-07T03:04:06Z

Investigations in the theory of quantum corrections to classical solutions of the Yang-Mills equations Callias, Constantine John. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Physics, 1979; Includes bibliographical references.

Wave equations, particles and chronometric geometry.

2026-01-07T03:03:58Z

Wave equations, particles and chronometric geometry. Orsted, Bent. Thesis: Ph. D., Massachusetts Institute of Technology, Department of Mathematics, 1976; Includes bibliographical references.

Planteez(tm) : business plan and preliminary research

2026-01-07T03:24:18Z

Planteez(tm) : business plan and preliminary research Sanchez, Manuel A. (Manuel Andres), 1979- Thesis: S.B., Massachusetts Institute of Technology, Department of Mechanical Engineering, 2001; Includes bibliographical references (p. 15).

The Refutability Gap: Challenges in Validating Reasoning by Large Language Models

2026-01-07T03:01:39Z

The Refutability Gap: Challenges in Validating Reasoning by Large Language Models Mossel, Elchanan Recent reports claim that Large Language Models (LLMs) have achieved the ability to derive new science and exhibit human-level general intelligence. We argue that such claims are not rigorous scientific claims, as they do not satisfy Popper’s refutability principle (often termed falsifiability), which requires that scientific statements be capable of being disproven. We identify several methodological pitfalls in current AI research on reasoning, including the inability to verify the novelty of findings due to opaque and non-searchable training data, the lack of reproducibility caused by continuous model updates, and the omission of human-interaction transcripts, which obscures the true source of scientific discovery. Additionally, the absence of counterfactuals and data on failed attempts creates a selection bias that may exaggerate LLM capabilities. To address these challenges, we propose guidelines for scientific transparency and reproducibility for research on reasoning by LLMs. Establishing such guidelines is crucial for both scientific integrity and the ongoing societal debates regarding fair data usage.

Proximity Loses: Real-Time Resolution of Ambiguous Wh-Questions in Japanese

2026-03-08T03:39:35Z

Proximity Loses: Real-Time Resolution of Ambiguous Wh-Questions in Japanese Nakamura, Chie; Flynn, Suzanne; Miyamoto, Yoichi; Yusa, Noriaki This study investigated how Japanese speakers interpret structurally ambiguous wh-questions, testing whether filler–gap resolution is guided by syntactic resolution based on hierarchical structure or linear locality based on surface word order. We combined behavioral key-press responses with fine-grained eye-tracking data and applied cluster-based permutation analysis to capture the moment-by-moment time course of syntactic interpretation as sentences were processed in real time. Key-press responses revealed a preference for resolving the dependency at the main clause (MC) gap position. Eye-tracking data showed early predictive fixations to the MC picture, followed by shifts to the embedded clause (EC) picture as the embedded event was described. These shifts occurred prior to the appearance of syntactic cues that signal the presence of an EC structure, such as the complementizer -to, and were therefore most likely guided by referential alignment with the linguistic input rather than by syntactic reanalysis. A subsequent return of the gaze to the MC picture occurred when the clause-final question particle -ka became available, confirming the interrogative use of the wh-phrase. Both key-press and eye-tracking data showed that participants did not commit to the first grammatically available EC interpretation but instead waited until clause-final particle information confirmed the interrogative use of the wh-phrase, ultimately favoring the MC interpretation. This pattern supports the view that filler–gap resolution is guided by structural locality rather than linear locality. By using high-resolution temporal data and statistically robust analytic techniques, this study demonstrates that Japanese comprehenders engage in predictive yet structurally cautious parsing. These findings challenge earlier claims that filler–gap resolution in Japanese is primarily driven by linear locality and instead showed a preference for resolving dependencies at the structurally higher MC position, consistent with parsing biases previously observed in English, despite typological differences in word order between the two languages. This preference also reflects sensitivity to language-specific morpho-syntactic cues in Japanese, such as clause-final particles.

Engineering peroxisomal biosynthetic pathways for maximization of triterpene production in Yarrowia lipolytica

2026-03-08T03:39:27Z

Engineering peroxisomal biosynthetic pathways for maximization of triterpene production in Yarrowia lipolytica Ma, Yongshuo; Shang, Yi; Stephanopoulos, Gregory Constructing efficient cell factories for product synthesis is frequently hampered by competing pathways and/or insufficient precursor supply. This is particularly evident in the case of triterpenoid biosynthesis in Yarrowia lipolytica, where squalene biosynthesis is tightly coupled to cytosolic biosynthesis of sterols essential for cell viability. Here, we addressed this problem by reconstructing the complete squalene biosynthetic pathway, starting from acetyl-CoA, in the peroxisome, thus harnessing peroxisomal acetyl-CoA pool and sequestering squalene synthesis in this organelle from competing cytosolic reactions. This strategy led to increasing the squalene levels by 1,300-fold relatively to native cytosolic synthesis. Subsequent enhancement of the peroxisomal acetyl-CoA supply by two independent approaches, 1) converting cellular lipid pool to peroxisomal acetyl-CoA and 2) establishing an orthogonal acetyl-CoA shortcut from CO2-derived acetate in the peroxisome, further significantly improved local squalene accumulation. Using these approaches, we constructed squalene-producing strains capable of yielding 32.8 g/L from glucose, and 31.6 g/L from acetate by employing a cofeeding strategy, in bioreactor fermentations. Our findings provide a feasible strategy for protecting intermediate metabolites that can be claimed by multiple reactions by engineering peroxisomes in Y. lipolytica as microfactories for the production of such intermediates and in particular acetyl-CoA-derived metabolites.

Drug screening in human physiologic medium identifies uric acid as an inhibitor of rigosertib efficacy

2026-03-08T03:39:27Z

Drug screening in human physiologic medium identifies uric acid as an inhibitor of rigosertib efficacy Rawat, Vipin; DeLear, Patrick; Prashanth, Prarthana; Ozgurses, Mete Emir; Tebeje, Anteneh; Burns, Philippa A; Conger, Kelly O; Solís, Christopher; Hasnain, Yasir; Novikova, Anna; Endress, Jennifer E; González-Sánchez, Paloma; Dong, Wentao; Stephanopoulos, Greg; DeNicola, Gina M; Harris, Isaac S; Sept, David; Mason, Frank M; Coloff, Jonathan L The nonphysiological nutrient levels found in traditional culture media have been shown to affect numerous aspects of cancer cell physiology, including how cells respond to certain therapeutic agents. Here, we comprehensively evaluated how physiological nutrient levels affect therapeutic response by performing drug screening in human plasma-like medium. We observed dramatic nutrient-dependent changes in sensitivity to a variety of FDA-approved and clinically trialed compounds, including rigosertib, an experimental cancer therapeutic that recently failed in phase III clinical trials. Mechanistically, we found that the ability of rigosertib to destabilize microtubules is strongly inhibited by the purine metabolism end product uric acid, which is uniquely abundant in humans relative to traditional in vitro and in vivo cancer models. These results demonstrate the broad and dramatic effects nutrient levels can have on drug response and how incorporation of human-specific physiological nutrient medium might help identify compounds whose efficacy could be influenced in humans.

Metabolic Engineering of E. coli for Enhanced Diols Production from Acetate

2026-03-08T03:38:53Z

Metabolic Engineering of E. coli for Enhanced Diols Production from Acetate Ricci, Luca; Cen, Xuecong; Zu, Yuexuan; Antonicelli, Giacomo; Chen, Zhen; Fino, Debora; Pirri, Fabrizio C; Stephanopoulos, Gregory; Woolston, Benjamin M; Re, Angela Effective employment of renewable carbon sources is highly demanded to develop sustainable biobased manufacturing. Here, we developed Escherichia coli strains to produce 2,3-butanediol and acetoin (collectively referred to as diols) using acetate as the sole carbon source by stepwise metabolic engineering. When tested in fed-batch experiments, the strain overexpressing the entire acetate utilization pathway was found to consume acetate at a 15% faster rate (0.78 ± 0.05 g/g/h) and to produce a 35% higher diol titer (1.16 ± 0.01 g/L) than the baseline diols-producing strain. Moreover, singularly overexpressing the genes encoding alternative acetate uptake pathways as well as alternative isoforms of genes in the malate-to-pyruvate pathway unveiled that leveraging ackA-pta and maeA is more effective in enhancing acetate consumption and diols production, compared to acs and maeB. Finally, the increased substrate consumption rate and diol production obtained in flask-based experiments were confirmed in bench-scale bioreactors operated in fed-batch mode. Consequently, the highest titer of 1.56 g/L achieved in this configuration increased by over 30% compared to the only other similar effort carried out so far.

Constant Degree Networks for Almost-Everywhere Reliable Transmission

2026-03-08T03:22:36Z

Constant Degree Networks for Almost-Everywhere Reliable Transmission Bafna, Mitali; Minzer, Dor In the almost-everywhere reliable message transmission problem, introduced by [Dwork, Pippenger, Peleg, Upfal’86], the goal is to design a sparse communication network G that supports efficient, fault-tolerant protocols for interactions between all node pairs. By fault-tolerant, we mean that that even if an adversary corrupts a small fraction of vertices in G, then all but a small fraction of vertices can still communicate perfectly via the constructed protocols. Being successful to do so allows one to simulate, on a sparse graph, any fault-tolerant distributed computing task and secure multi-party computation protocols built for a complete network, with only minimal overhead in efficiency. Previous works on this problem achieved either constant-degree networks tolerating o(1) faults, constant-degree networks tolerating a constant fraction of faults via inefficient protocols (exponential work complexity), or poly-logarithmic degree networks tolerating a constant fraction of faults. We show a construction of constant-degree networks with efficient protocols (i.e., with polylogarithmic work complexity) that can tolerate a constant fraction of adversarial faults, thus solving the main open problem of Dwork et al. Our main contribution is a composition technique for communication networks, based on graph products. Our technique combines two networks tolerant to adversarial edge-faults to construct a network with a smaller degree while maintaining efficiency and fault-tolerance. We apply this composition result multiple times, using the polylogarithmic-degree edge-fault tolerant networks constructed in a recent work of [Bafna, Minzer, Vyas’24] (that are based on high-dimensional expanders) with itself, and then with the constant-degree networks (albeit with inefficient protocols) of [Upfal’92]. STOC ’25, Prague, Czechia

Quasi-Linear Size PCPs with Small Soundness from HDX

2026-03-08T03:22:29Z

Quasi-Linear Size PCPs with Small Soundness from HDX Bafna, Mitali; Minzer, Dor; Vyas, Nikhil; Yun, Zhiwei We construct 2-query, quasi-linear size probabilistically checkable proofs (PCPs) with arbitrarily small constant soundness, improving upon Dinur’s 2-query quasi-linear size PCPs with soundness 1 − Ω(1). As an immediate corollary, we get that under the exponential time hypothesis, for all𝜀 > 0 no approximation algorithm for 3-SAT can obtain an approximation ratio of 7/8+𝜀 in time 2 𝑛/log𝐶 𝑛 , where 𝐶 is a constant depending on 𝜀. Our result builds on a recent line of independent works by Bafna, Lifshitz and Minzer, and Dikstein, Dinur and Lubotzky, that showed the existence of linear size direct product testers with small soundness. The main new ingredient in our proof is a technique that embeds a given 2-CSP into a 2-CSP on a prescribed graph, provided that the latter is a graph underlying a sufficiently good high-dimensional expander (HDX). We achieve this by establishing a novel connection between PCPs and fault-tolerant distributed computing, more precisely, to the almost-everywhere reliable transmission problem introduced by Dwork, Peleg, Pippenger and Upfal (1986). We instantiate this connection by showing that graphs underlying HDXs admit routing protocols that are tolerant to adversarial edge corruptions, also improving upon the state of the art constructions of sparse edge-fault-tolerant networks in the process. Our PCP construction requires variants of the aforementioned direct product testers with poly-logarithmic degree. The existence and constructability of these variants is shown in the full version. STOC ’25, Prague, Czechia

Approximately Counting and Sampling Hamiltonian Motifs in Sublinear Time

2026-03-08T03:22:49Z

Approximately Counting and Sampling Hamiltonian Motifs in Sublinear Time Eden, Talya; Levi, Reut; Ron, Dana; Rubinfeld, Ronitt Counting small subgraphs, referred to as motifs, in large graphs is a fundamental task in graph analysis, extensively studied across various contexts and computational models. In the sublinear-time regime, the relaxed problem of approximate counting has been explored within two prominent query frameworks: the standard model, which permits degree, neighbor, and pair queries, and the strictly more powerful augmented model, which additionally allows for uniform edge sampling. Currently, in the standard model, (optimal) results have been established only for approximately counting edges, stars, and cliques, all of which have a radius of one. This contrasts sharply with the state of affairs in the augmented model, where algorithmic results (some of which are optimal) are known for any input motif, leading to a disparity which we term the “scope gap" between the two models. In this work, we make significant progress in bridging this gap. Our approach draws inspiration from recent advancements in the augmented model and utilizes a framework centered on counting by uniform sampling, thus allowing us to establish new results in the standard model and simplify on previous results. In particular, our first, and main, contribution is a new algorithm in the standard model for approximately counting any Hamiltonian motif in sublinear time, where the complexity of the algorithm is the sum of two terms. One term equals the complexity of the known algorithms by Assadi, Kapralov, and Khanna (ITCS 2019) and Fichtenberger and Peng (ICALP 2020) in the (strictly stronger) augmented model and the other is an additional, necessary, additive overhead. Our second contribution is a variant of our algorithm that enables nearly uniform sampling of these motifs, a capability previously limited in the standard model to edges and cliques. Our third contribution is to introduce even simpler algorithms for stars and cliques by exploiting their radius-one property. As a result, we simplify all previously known algorithms in the standard model for stars (Gonen, Ron, Shavitt (SODA 2010)), triangles (Eden, Levi, Ron Seshadhri (FOCS 2015)) and cliques (Eden, Ron, Seshadri (STOC 2018)). STOC ’25, Prague, Czechia

Sandwiching Random Geometric Graphs and Erdos-Renyi with Applications: Sharp Thresholds, Robust Testing, and Enumeration

2026-03-08T03:39:03Z

Sandwiching Random Geometric Graphs and Erdos-Renyi with Applications: Sharp Thresholds, Robust Testing, and Enumeration Bangachev, Kiril; Bresler, Guy The distribution RGG(n,Sd−1,p) is formed by sampling independent vectors {Vi}i = 1n uniformly on Sd−1 and placing an edge between pairs of vertices i and j for which ⟨ Vi,Vj⟩ ≥ τdp, where τdp is such that the expected density is p. Our main result is a poly-time implementable coupling between Erdős-Rényi and RGG such that G(n,p(1 − O(√np/d)))⊆ RGG(n,Sd−1,p)⊆ G(n,p(1 + O(√np/d))) edgewise with high probability when d≫ np. We apply the result to: 1) Sharp Thresholds: We show that for any monotone property having a sharp threshold with respect to the Erdős-Rényi distribution and critical probability pnc, random geometric graphs also exhibit a sharp threshold when d≫ npnc, thus partially answering a question of Perkins. 2) Robust Testing: The coupling shows that testing between G(n,p) and RGG(n,Sd−1,p) with є n2p adversarially corrupted edges for any constant є>0 is information-theoretically impossible when d≫ np. We match this lower bound with an efficient (constant degree SoS) spectral refutation algorithm when d≪ np. 3) Enumeration: We show that the number of geometric graphs in dimension d is at least exp(dnlog−7n), recovering (up to the log factors) the sharp result of Sauermann. STOC ’25, Prague, Czechia

Sample-Optimal Private Regression in Polynomial Time

2026-03-08T03:22:47Z

Sample-Optimal Private Regression in Polynomial Time Anderson, Prashanti; Bakshi, Ainesh; Majid, Mahbod; Tiegel, Stefan We consider the task of privately obtaining prediction error guarantees in ordinary least-squares regression problems with Gaussian covariates (with unknown covariance structure). We provide the first sample-optimal polynomial time algorithm for this task under both pure and approximate differential privacy. We show that any improvement to the sample complexity of our algorithm would violate either statistical-query or information-theoretic lower bounds. Additionally, our algorithm is robust to a small fraction of arbitrary outliers and achieves optimal error rates as a function of the fraction of outliers. In contrast, all prior efficient algorithms either incurred sample complexities with sub-optimal dimension dependence, scaling with the condition number of the covariates, or obtained a polynomially worse dependence on the privacy parameters. Our technical contributions are two-fold: first, we leverage resilience guarantees of Gaussians within the sum-of-squares framework. As a consequence, we obtain efficient sum-of-squares algorithms for regression with optimal robustness rates and sample complexity. Second, we generalize the recent robustness-to-privacy framework of Hopkins, Kamath, Majid, and Narayanan to account for the geometry induced by the covariance of the input samples. This framework crucially relies on the robust estimators to be sum-of-squares algorithms, and combining the two steps yields a sample-optimal private regression algorithm. We believe our techniques are of independent interest, and we demonstrate this by obtaining an efficient algorithm for covariance-aware mean estimation, with an optimal dependence on the privacy parameters. STOC ’25, Prague, Czechia

Faster Weighted and Unweighted Tree Edit Distance and APSP Equivalence

2026-03-08T03:22:38Z

Faster Weighted and Unweighted Tree Edit Distance and APSP Equivalence Nogler, Jakob; Polak, Adam; Saha, Barna; Vassilevska Williams, Virginia; Xu, Yinzhan; Ye, Christopher The tree edit distance (TED) between two rooted ordered trees with n nodes labeled from an alphabet Σ is the minimum cost of transforming one tree into the other by a sequence of valid operations consisting of insertions, deletions and relabeling of nodes. The tree edit distance is a well-known generalization of string edit distance and has been studied since the 1970s. Its running time has seen steady improvements starting with an O(n6) algorithm [Tai, J.ACM 1979], improved to O(n4) [Shasha, Zhang, SICOMP 1989] and to O(n3logn) [Klein, ESA 1998], and culminating in an O(n3) algorithm [Demaine, Mozes, Rossman, Weimann, ACM TALG 2010]. The latter is known to be optimal for any dynamic programming based algorithm that falls under a certain decomposition framework that captures all known sub-n4 time algorithms. Fine-grained complexity casts further light onto this hardness showing that a truly subcubic time algorithm for TED implies a truly subcubic time algorithm for All-Pairs Shortest Paths (APSP) [Bringmann, Gawrychowski, Mozes, Weimann, ACM TALG 2020]. Therefore, under the popular APSP hypothesis, a truly subcubic time algorithm for TED cannot exist. However, unlike many problems in fine-grained complexity for which conditional hardness based on APSP also comes with equivalence to APSP, whether TED can be reduced to APSP has remained unknown. In this paper, we resolve this. Not only we show that TED is fine-grained equivalent to APSP, our reduction is tight enough, so that combined with the fastest APSP algorithm to-date [Williams, SICOMP 2018] it gives the first ever subcubic time algorithm for TED running in n3/2Ω(√logn) time. We also consider the unweighted tree edit distance problem in which the cost of each edit (insertion, deletion, and relabeling) is one. For unweighted TED, a truly subcubic algorithm is known due to Mao [Mao, FOCS 2022], and later improved slightly by Dürr [Dürr, IPL 2023] to run in O(n2.9148) time. Since their algorithm uses bounded monotone min-plus product as a crucial subroutine, and the best running time for this product is Õ(n3+ω/2)≤ O(n2.6857) (where ω is the exponent of fast matrix multiplication), the much higher running time of unweighted TED remained unsatisfactory. In this work, we close this gap and give an algorithm for unweighted TED that runs in Õ(n3+ω/2) time. STOC ’25, Prague, Czechia

The Structure of Catalytic Space: Capturing Randomness and Time via Compression

2026-03-08T03:22:44Z

The Structure of Catalytic Space: Capturing Randomness and Time via Compression Cook, James; Li, Jiatu; Mertz, Ian; Pyne, Edward In the catalytic logspace (CL) model of (Buhrman et. al. STOC 2013), we are given a small work tape, and a larger catalytic tape that has an arbitrary initial configuration. We may edit this tape, but it must be exactly restored to its initial configuration at the completion of the computation. This model is of interest from a complexity-theoretic perspective as it gains surprising power over traditional space. However, many fundamental structural questions remain open. We substantially advance the understanding of the structure of CL, addressing several questions raised in prior work. Our main results are as follows. 1: We unconditionally derandomize catalytic logspace: CBPL = CL. 2: We show time and catalytic space bounds can be achieved separately if and only if they can be achieved simultaneously: any problem in both CL and P can be solved in polynomial time-bounded CL. 3: We characterize deterministic catalytic space by the intersection of randomness and time: CL is equivalent to polytime-bounded, zero-error randomized CL. Our results center around the compress--or--random framework. For the second result, we introduce a simple yet novel compress--or--compute algorithm which, for any catalytic tape, either compresses the tape or quickly and successfully computes the function at hand. For our first result, we further introduce a compress--or--compress--or--random algorithm that combines runtime compression with a second compress--or--random algorithm, building on recent work on distinguish-to-predict transformations and pseudorandom generators with small-space deterministic reconstruction. STOC ’25, Prague, Czechia

Rounding Large Independent Sets on Expanders

2026-03-08T03:22:48Z

Rounding Large Independent Sets on Expanders Bafna, Mitali; Hsieh, Jun-Ting; Kothari, Pravesh K. We develop a new approach for approximating large independent sets when the input graph is a one-sided spectral expander - that is, the uniform random walk matrix of the graph has its second eigenvalue bounded away from 1. Consequently, we obtain a polynomial time algorithm to find linear-sized independent sets in one-sided expanders that are almost 3-colorable or are promised to contain an independent set of size (1/2−є)n. Our second result above can be refined to require only a weaker vertex expansion property with an efficient certificate. In a surprising contrast to our algorithmic result, we observe that the analogous task of finding a linear-sized independent set in almost 4-colorable one-sided expanders (even when the second eigenvalue is on(1)) is NP-hard, assuming the Unique Games Conjecture. All prior algorithms that beat the worst-case guarantees for this problem rely on bottom eigenspace enumeration techniques (following the classical spectral methods of Alon and Kahale) and require two-sided expansion, meaning a bounded number of negative eigenvalues of magnitude Ω(1). Such techniques naturally extend to almost k-colorable graphs for any constant k, in contrast to analogous guarantees on one-sided expanders, which are Unique Games-hard to achieve for k ≥ 4. Our rounding scheme builds on the method of simulating multiple samples from a pseudo-distribution introduced in Bafna et. al. for rounding Unique Games instances. The key to our analysis is a new clustering property of large independent sets in expanding graphs - every large independent set has a larger-than-expected intersection with some member of a small list - and its formalization in the low-degree sum-of-squares proof system. STOC ’25, Prague, Czechia

Universal SNARGs for NP from Proofs of Correctness

2026-03-08T03:22:24Z

Universal SNARGs for NP from Proofs of Correctness Jin, Zhengzhong; Kalai, Yael Tauman; Lombardi, Alex; Mathialagan, Surya We give new constructions of succinct non-interactive arguments (SNARGs) for NP in the settings of both non-adaptive and adaptive soundness. Our construction of non-adaptive SNARG is universal assuming the security of a (leveled or unleveled) fully homomorphic encryption (FHE) scheme as well as a batch argument (BARG) scheme. Specifically, for any choice of parameters ℓ and L, we construct a candidate SNARG scheme for any NP language L with the following properties: (i) the proof length is ℓ· poly(λ), (ii) the common reference string crs has length L· poly(λ), and (iii) the setup is transparent (no private randomness). We prove that this SNARG has non-adaptive soundness assuming the existence of any SNARG where the proof size is ℓ, the crs size is L, and there is a size L Extended Frege (EF) proof of completeness for the SNARG. Moreover, we can relax the underlying SNARG to be any 2-message privately verifiable argument where the first message is of length L and the second message is of length ℓ. This yields new SNARG constructions based on any “EF-friendly” designated-verifier SNARG or witness encryption scheme. We emphasize that our SNARG is universal in the sense that it does not depend on the argument system. We show several new implications of this construction that do not reference proof complexity: (1) a non-adaptive SNARG for NP with transparent crs from LWE under the evasive LWE heuristic. This gives a candidate lattice-based SNARG for NP. (2) a non-adaptive SNARG for NP with transparent crs assuming the (non-explicit) existence of any iO and LWE. (3) a non-adaptive SNARG for NP with a short and transparent (i.e., uniform) crs assuming LWE, FHE and the (non-explicit) existence of any hash function that makes Micali’s SNARG construction sound. (4) a non-adaptive SNARG for languages such as QR and DCR assuming only LWE. In the setting of adaptive soundness, we show how to convert any designated verifier SNARG into publicly verifiable SNARG, assuming the underlying designated verifier SNARG has an EF proof of completeness. As a corollary, we construct an adaptive SNARG for UP with a transparent crs assuming subexponential LWE under the evasive LWE heuristic. We prove our results by extending the encrypt-hash-and-BARG paradigm of [Jin-Kalai-Lombardi-Vaikuntanathan, STOC ’24]. STOC ’25, Prague, Czechia

The Medium is the Message: How Non-Clinical Information Shapes Clinical Decisions in LLMs

2026-03-08T03:22:18Z

The Medium is the Message: How Non-Clinical Information Shapes Clinical Decisions in LLMs Gourabathina, Abinitha; Gerych, Walter; Pan, Eileen; Ghassemi, Marzyeh The integration of large language models (LLMs) into clinical diagnostics necessitates a careful understanding of how clinically irrelevant aspects of user inputs directly influence generated treatment recommendations and, consequently, clinical outcomes for end-users. Building on prior research that examines the impact of demographic attributes on clinical LLM reasoning, this study explores how non-clinically relevant attributes shape clinical decision-making by LLMs. Through the perturbation of patient messages, we evaluate whether LLM behavior remains consistent, accurate, and unbiased when non-clinical information is altered. These perturbations assess the brittleness of clinical LLM reasoning by replicating structural errors that may occur during electronic data processing patient questions and simulating interactions between patient-AI systems in diverse, vulnerable patient groups. Our findings reveal notable inconsistencies in LLM treatment recommendations and significant degradation of clinical accuracy in ways that reduce care allocation to patients. Additionally, there are significant disparities in treatment recommendations between gender subgroups as well as between model-inferred gender subgroups. We also apply our perturbation framework to a conversational clinical dataset to find that even in conversation, LLM clinical accuracy decreases post-perturbation, and disparities exist in how perturbations impact gender subgroups. By analyzing LLM outputs in response to realistic yet modified clinical contexts, our work deepens understanding of the sensitivity, inaccuracy, and biases inherent in medical LLMs, offering critical insights for the deployment of patient-AI systems. FAccT ’25, Athens, Greece

High-Performance Mixed-Precision Matrix Multiplication via Tile-Centric Design on Modern Architectures

2025-12-23T03:10:24Z

High-Performance Mixed-Precision Matrix Multiplication via Tile-Centric Design on Modern Architectures Zhang, Qiao; Alomairy, Rabab; Wang, Dali; Gu, Zhuowei; Cao, Qinglei General Matrix Multiplication (GEMM) is a critical operation underpinning a wide range of applications in high-performance computing (HPC) and artificial intelligence (AI). The emergence of hardware optimized for low-precision arithmetic necessitates a reevaluation of numerical algorithms to leverage mixed-precision computations, achieving improved performance and energy efficiency. This research presents an adaptive mixed-precision GEMM framework that enables support for various precision formats at fine-grained tile and block levels, offering a reliable foundation for trustworthy mixed-precision computations. Furthermore, we leverage the PaRSEC runtime system to effectively balance workloads across diverse architectures. The performance exhibits strong scalability across both homogeneous platforms (Intel CPU-based systems and the ARM CPU-based Fugaku supercomputer) and heterogeneous systems (Nvidia V100, A100, and H100 GPU-based platforms, as well as the AMD GPU-based Frontier supercomputer). This work aims to improve computational efficiency and accuracy by bridging algorithmic innovations with hardware capabilities, fostering transformative advancements across a wide range of applications.

Search for t-channel scalar and vector leptoquark exchange in the high-mass dimuon and dielectron spectra in proton-proton collisions at √s = 13 TeV

2026-03-08T03:38:51Z

Search for t-channel scalar and vector leptoquark exchange in the high-mass dimuon and dielectron spectra in proton-proton collisions at √s = 13 TeV Hayrapetyan, A.; Tumasyan, A.; Adam, W.; Andrejkovic, J. W.; Benato, L.; Bergauer, T.; Chatterjee, S.; Damanakis, K.; Dragicevic, M.; Hussain, P. S.; Jeitler, M.; Krammer, N.; Li, A.; Liko, D.; Mikulec, I.; Schieck, J.; Schöfbeck, R. A search for t-channel exchange of leptoquarks (LQs) is performed in dimuon and dielectron spectra using proton-proton collision data collected at √ s = 13 TeV with the CMS detector at the CERN LHC. The data correspond to an integrated luminosity of 138 fb−1 . Eight scenarios are considered, in which up or down quarks couple to muons or electrons via a scalar or vector LQ exchange, for dilepton invariant masses above 500 GeV. The LQ masses are probed up to 5 TeV, beyond a regime probed by previous pair-production and single-production searches. The differential distributions of dilepton events are fit to templates that model the nonresonant LQ exchange and various standard model background processes. Limits are set on LQ-fermion coupling strengths for scalar and vector LQ masses in the 1–5 TeV range at 95% confidence level, establishing stringent limits on first- and second-generation LQs.

Search for charged-lepton flavour violation in top quark interactions with an up-type quark, a muon, and a τ lepton in proton-proton collisions at √s = 13 TeV

2026-03-08T03:38:51Z

Search for charged-lepton flavour violation in top quark interactions with an up-type quark, a muon, and a τ lepton in proton-proton collisions at √s = 13 TeV Hayrapetyan, A.; Tumasyan, A.; Adam, W.; Andrejkovic, J. W.; Benato, L.; Bergauer, T.; Chatterjee, S.; Damanakis, K.; Dragicevic, M.; Hussain, P. S.; Jeitler, M.; Krammer, N.; Li, A.; Liko, D.; Mikulec, I.; Schieck, J. A search for charged-lepton flavour violation (CLFV) in top quark (t) production and decay is presented. The search uses proton-proton collision data corresponding to 138 fb−1 collected with the CMS experiment at √ s = 13 TeV. The signal consists of the production of a single top quark via a CLFV interaction or top quark pair production followed by a CLFV decay. The analysis selects events containing a hadronically decaying τ lepton and a muon of opposite electric charge, as well as at least three jets, one of which is identified as originating from the fragmentation of a bottom quark. Machine learning classification techniques are used to distinguish signal from standard model background events. The results of this search are consistent with the standard model expectations. The upper limits at 95% confidence level on the branching fraction B for CLFV top quark decays to a muon, a τ lepton, and an up or a charm quark are set at B(t → µτu) < (0.04, 0.08, and 0.12) × 10−6 , and B(t → µτ c) < (0.81, 1.71, and 2.05) × 10−6 for scalar, vector, and tensor-like operators, respectively.

Anyon delocalization transitions out of a disordered fractional quantum anomalous Hall insulator

2026-03-08T03:39:17Z

Anyon delocalization transitions out of a disordered fractional quantum anomalous Hall insulator Shi, Zhengyan Darius; Todadri, Senthil Motivated by the experimental discovery of the fractional quantum anomalous Hall effect, we develop a theory of doping-induced transitions out of the = 2/3 lattice Jain state in the presence of quenched disorder. We show that disorder strongly affects the evolution into the conducting phases described in our previous work. The delocalization of charge 2/3 anyons leads to a chiral superconductor through a direct second-order transition for a smooth random potential with long-wavelength modulations. The longitudinal resistance has a universal peak at the associated quantum critical point. Close to the transition, we show that the superconducting ground state is an “Anomalous Vortex Glass” stabilized in the absence of an external magnetic field. For short-wavelength disorder, this transition generically splits into three distinct ones with intermediate insulating topological phases. If instead, the charge 1/3 anyon delocalizes, then at low doping the resulting phase is a Reentrant Integer Quantum Hall state with xy = h/e 2 . At higher doping this undergoes a second transition to a Fermi liquid metal. We show that this framework provides a plausible explanation for the complex phase diagram recently observed in twisted MoTe2 near = 2/3 and discuss future experiments that can test our theory in more detail.

From Data to Transformative Change: Designing Interactive Systems for Citizen Science Empowerment

2025-12-20T03:09:51Z

From Data to Transformative Change: Designing Interactive Systems for Citizen Science Empowerment Prandi, Catia; Herodotou, Christothea; Dionisio, Mara; Reeves, Neal; Reitsma, Lizette; Mora, Simone Citizen Science (CS) is a research approach in which scientists and everyday people collaborate to address a research problem. Advancements in digital technologies have significantly expanded the reach of Citizen Science by enabling large-scale data collection and collaboration. In addition to its scientific benefits, citizen science enhances participants’ science literacy, fosters public engagement, and promotes collaborative problem-solving. Despite this being true, we believe that the true potential of CS has not yet been fully explored as a collaborative practice for transformative change. With this in mind, we planned a one-day workshop as a forum for critical discussions and reflections on the role of HCI researchers, designers, and practitioners in designing CS-empowered interactive systems for increasing awareness about social good and societal issues and promoting concrete actions and behavioural change, from data to sustainable futures. Participants will have the possibility to reflect on and discuss the main open challenges still affecting the design of CS-empowered interactive systems, and to prototype, exploiting data physicalization and co-design, solutions that focus on a specific real-world challenge as presented by experts of the Madeira Island that offers a unique ecosystem to spark reflections on the interplay between sustainability, technology and CS. DIS ’25 Companion, Funchal, Portugal

Lost in Transplantation: Characterizing Racial Gaps in Physician Organ Offer Acceptance

2025-12-20T03:09:52Z

Lost in Transplantation: Characterizing Racial Gaps in Physician Organ Offer Acceptance Adam, Hammaad; Bermea, Rene; Yang, Ming Ying; Celi, Leo Anthony; Ghassemi, Marzyeh There are known racial disparities in the organ transplant allocation system in the United States. While recent research has focused on designing scores and matching algorithms for organ allocation, prior work has yet to study how transplant center physician decisions on offer acceptance—the final step in the allocation process—contribute to the observed disparities. In this paper, we use data from the Scientific Registry of Transplant Recipients to examine the role of candidate race in the acceptance of heart, liver, and lung transplant offers. We find that Black race was associated with significantly lower odds of offer acceptance for livers and lungs. Further, existing allocation scores such as MELD and LAS did not account for clinical factors that made Black patients harder to match. Our analysis also revealed that donor candidate race-match was associated with significantly higher odds of offer acceptance for hearts, livers, and lungs. Finally, we found that rejecting an offer was associated with lower survival times for all three organs. Our findings demonstrate the additional barriers that Black patients face in accessing organ transplants and the consequences of these barriers on patient survival. Overall, our work highlights the limitations of technical solutions to socio-technical problems; new allocation scores and other algorithmic updates will not improve equity if they do not explicitly account for gaps in the ensuing human decisions. FAccT ’25, Athens, Greece

Coboundary Expansion of Coset Complexes

2025-12-20T03:09:54Z

Coboundary Expansion of Coset Complexes Kaufman, Tali; Oppenheim, Izhar; Weinberger, Shmuel Coboundary expansion is a high dimensional generalization of the Cheeger constant to simplicial complexes. Originally, this notion was motivated by the fact that it implies topological expansion, but nowadays a significant part of the motivation stems from its deep connection to problems in theoretical computer science such as list agreement expansion and agreement expansion in the low soundness regime. In this paper, we prove coboundary expansion with non-Abelian coefficients for the coset complex construction of Kaufman and Oppenheim. Our proof uses a novel global argument, as opposed to the local-to-global arguments that are used to prove cosystolic expansion. STOC ’25, Prague, Czechia

Disclosure without Engagement: An Empirical Review of Positionality Statements at FAccT

2025-12-20T03:09:55Z

Disclosure without Engagement: An Empirical Review of Positionality Statements at FAccT Schroeder, Hope; Pareek, Akshansh; Barocas, Solon Positionality statements have become more common in engineering fields in recent years, despite ongoing debates across many fields about the merits of the practice. In 2024, the Program Chairs of FAccT recommended that authors include positionality statements with their paper submissions, dramatically increasing their use at the conference. In this paper, we analyze all positionality statements at FAccT from 2018 to 2024, highlighting the different aspects of identity commonly disclosed by authors and the degree to which authors explore the potential impact of these aspects of their positionality on their research. While we encountered and highlight a number of thoughtful positionality statements, we also identified and describe several concerning trends, including patterns of identity disclosure without discussion of corresponding impacts, a notable lack of reflection on the potential impacts of industry affiliation, and cases where identity is invoked to excuse what are really methodological choices, among others. We raise particular concerns about the possibility that disclosure without engagement may cause readers to rely on stereotypes to make guesses about the perspectives that individuals from certain groups bring to their work. We conclude by considering potential mechanisms for encouraging reflexivity in the FAccT community, with a focus on setting policies that protect researchers from risks, supporting researchers from backgrounds without existing traditions of reflexive practice, and empirically evaluating the efficacy of interventions designed to foster reflexivity. FAccT ’25, Athens, Greece

LEAD: Towards Learning-Based Equity-Aware Decarbonization in Ridesharing Platforms

2025-12-20T03:09:43Z

LEAD: Towards Learning-Based Equity-Aware Decarbonization in Ridesharing Platforms Sahebdel, Mahsa; Zeynali, Ali; Bashir, Noman; Shenoy, Prashant; Hajiesmaili, Mohammad Ridesharing platforms such as Uber, Lyft, and DiDi have grown in popularity due to their on-demand availability, ease of use, and commute cost reductions, among other benefits. However, not all ridesharing promises have panned out. Recent studies demonstrate that the expected drop in traffic congestion and reduction in greenhouse gas (GHG) emissions have not materialized. This is primarily due to the substantial distances traveled by the ridesharing vehicles without passengers between rides, known as deadhead miles. Recent work has focused on reducing the impact of deadhead miles while considering additional metrics such as rider waiting time, GHG emissions from deadhead miles, or driver earnings. However, most prior studies consider these environmental and equity-based metrics individually despite them being interrelated. In this paper, we propose a Learning-based Equity-Aware Decarabonization approach, LEAD, for ridesharing platforms. LEAD targets minimizing emissions while ensuring that the driver’s utility, defined as the difference between the trip distance and the deadhead miles, is fairly distributed. LEAD uses reinforcement learning to match riders with drivers based on the expected future utility of drivers and the expected carbon emissions of the platform without increasing the rider waiting times. Extensive experiments based on a real-world ridesharing dataset show that LEAD improves the defined notion of fairness by 150% when compared to emission-aware ride-assignment and reduces emissions by 14.6% while ensuring fairness within 28–52% of the fairness-focused baseline. It also reduces the rider wait time, by at least 32.1%, compared to a fairness-focused baseline. FAccT ’25, Athens, Greece

LuciEntry: Towards Understanding the Design of Lucid Dream Induction

2025-12-19T05:31:15Z

LuciEntry: Towards Understanding the Design of Lucid Dream Induction Wang, Po-Yao (Cosmos); Fang, Xiao Zoe; Ducos, Gabriel; Lee, Nathaniel Yung Xiang; Loose, Antony; Rajesh, Rohit; Botheju, Nethmini; Chen, Eric; Montoya, Maria; Kitson, Alexandra; Konkoly, Karen; Sagi, Rohan; Patibanda, Rakesh; Whitmore, Nathan; Jafarzadeh Esfahani, Mahdad; Deng, Jialin; Bu, Jiajun; Dresler, Martin; Elvitigala, Don Samitha; Semertzidis, Nathan; Mueller, Florian Lucid dreaming, a state in which people become aware that they are dreaming, is known for its many mental and physical health benefits. However, most lucid dream induction techniques, such as reality testing, require significant time and effort to master, creating a barrier for people seeking these experiences. We designed LuciEntry, a portable interactive prototype aimed at helping people induce lucid dreaming through well-timed visual and auditory cues. We conducted a lab and a field study to understand LuciEntry’s user experience. The interview data allowed us to identify three themes. Building on these findings and our design practice, we derived seven considerations to guide the design of future lucid dream systems. Ultimately, this work aims to inspire further research into interactive technologies for altered states of consciousness. DIS ’25, Funchal, Portugal

The Reality of AI and Biorisk

2025-12-19T05:31:08Z

The Reality of AI and Biorisk Peppin, Aidan; Reuel, Anka; Casper, Stephen; Jones, Elliot; Strait, Andrew; Anwar, Usman; Agrawal, Anurag; Kapoor, Sayash; Koyejo, Sanmi; Pellat, Marie; Bommasani, Rishi; Frosst, Nick; Hooker, Sara To accurately and confidently answer the question “could an AI model or system increase biorisk”, it is necessary to have both a sound theoretical threat model for how AI models or systems could increase biorisk and a robust method for testing that threat model. This paper provides an analysis of existing available research surrounding two AI and biorisk threat models: 1) access to information and planning via large language models (LLMs), and 2) the use of AI-enabled biological tools (BTs) in synthesizing novel biological artifacts. We find that existing studies around AI-related biorisk are nascent, often speculative in nature, or limited in terms of their methodological maturity and transparency. The available literature suggests that current LLMs and BTs do not pose an immediate risk, and more work is needed to develop rigorous approaches to understanding how future models could increase biorisks. We end with recommendations about how empirical work can be expanded to more precisely target biorisk and ensure rigor and validity of findings. FAccT ’25, Athens, Greece

SoS Certifiability of Subgaussian Distributions and Its Algorithmic Applications

2025-12-19T05:31:17Z

SoS Certifiability of Subgaussian Distributions and Its Algorithmic Applications Diakonikolas, Ilias; Hopkins, Samuel; Pensia, Ankit; Tiegel, Stefan We prove that there is a universal constant C>0 so that for every d ∈ ℕ, every centered subgaussian distribution D on ℝd, and every even p ∈ ℕ, the d-variate polynomial (Cp)p/2 · ||v||2p − EX ∼ D ⟨ v,X⟩p is a sum of square polynomials. This establishes that every subgaussian distribution is SoS-certifiably subgaussian—a condition that yields efficient learning algorithms for a wide variety of high-dimensional statistical tasks. As a direct corollary, we obtain computationally efficient algorithms with near-optimal guarantees for the following tasks, when given samples from an arbitrary subgaussian distribution: robust mean estimation, list-decodable mean estimation, clustering mean-separated mixture models, robust covariance-aware mean estimation, robust covariance estimation, and robust linear regression. Our proof makes essential use of Talagrand’s generic chaining/majorizing measures theorem. STOC ’25, Prague, Czechia

Using collective dialogues and AI to find common ground between Israeli and Palestinian peacebuilders

2025-12-19T05:31:06Z

Using collective dialogues and AI to find common ground between Israeli and Palestinian peacebuilders Konya, Andrew; Thorburn, Luke; Almasri, Wasim; Leshem, Oded Adomi; Procaccia, Ariel; Schirch, Lisa; Bakker, Michiel A growing body of work has shown that AI-assisted methods — leveraging large language models, social choice methods, and collective dialogues — can help navigate polarization and surface common ground in controlled lab settings. But what can these approaches contribute in real-world contexts? We present a case study applying these techniques to find common ground between Israeli and Palestinian peacebuilders in the period following October 7th, 2023. From April to July 2024 an iterative deliberative process combining LLMs, bridging-based ranking, and collective dialogues was conducted in partnership with the Alliance for Middle East Peace. Around 138 civil society peacebuilders participated including Israeli Jews, Palestinian citizens of Israel, and Palestinians from the West Bank and Gaza. The process resulted in a set of collective statements, including demands to world leaders, with at least 84% agreement from participants on each side. In this paper, we document the process, results, challenges, and important open questions. FAccT ’25, Athens, Greece

Recourse, Repair, Reparation, & Prevention: A Stakeholder Analysis of AI Supply Chains

2025-12-19T05:31:11Z

Recourse, Repair, Reparation, & Prevention: A Stakeholder Analysis of AI Supply Chains Hopkins, Aspen; Struckman, Isabella; Klyman, Kevin; Silbey, Susan S. The AI industry is exploding in popularity, with increasing attention to potential harms and unwanted consequences. In the current digital ecosystem, AI deployments are often the product of AI supply chains (AISC): networks of outsourced models, data, and tooling through which multiple entities contribute to AI development and distribution. AI supply chains lack the modularity, redundancies, or conventional supply chain practices that enable identification, isolation, and easy correction of failures, exacerbating the already difficult processes of responding to ML-generated harms. As the stakeholders participating in and impacted by AISCs have scaled and diversified, so too have the risks they face. In this stakeholder analysis of AI supply chains, we consider who participates in AISCs, what harms they face, where sources of harm lie, and how market dynamics and power differentials inform the type and probability of remedies. Because AI supply chains are purposely invented and implemented, they may be designed to account for, rather than ignore, the complexities, consequences, and risks of deploying AI systems. To enable responsible design and management of AISCs, we offer a typology of responses to AISC-induced harms: recourse, repair, reparation or prevention. We apply this typology to stakeholders participating in a health-care AISC across three stylized markets—vertical integration, horizontal integration, free market—to illustrate how stakeholder positioning and power within an AISC may shape responses to an experienced harm. FAccT ’25, Athens, Greece

When to Ask a Question: Understanding Communication Strategies in Generative AI Tools

2025-12-19T05:31:04Z

When to Ask a Question: Understanding Communication Strategies in Generative AI Tools Park, Charlotte; Donahue, Kate; Raghavan, Manish Generative AI tools (GAITs) fundamentally differ from traditional machine learning tools in that they allow users to provide as much or as little information as they choose in their inputs. This flexibility often leads users to omit certain details, relying on the GAIT to infer and fill in less critical information based on distributional knowledge of user preferences. Inferences about preferences lead to natural questions about fairness, since a GAIT’s “best guess” may skew towards the preferences of larger groups at the expense of smaller ones. Unlike more traditional recommender systems, GAITs can acquire additional information about a user’s preferences through feedback or by explicitly soliciting it. This creates an interesting communication challenge: the user is aware of their specific preference, while the GAIT has knowledge of the overall distribution of preferences, and both parties can only exchange a limited amount of information. In this work, we present a mathematical model to describe human-AI co-creation of content under information asymmetry. Our results suggest that GAITs can use distributional information about overall preferences to determine the “right” questions to ask to maximize both welfare and fairness, opening up a rich design space in human-AI collaboration. UMAP Adjunct ’25, New York City, NY, USA

The Cloud Next Door: Investigating the Environmental and Socioeconomic Strain of Datacenters on Local Communities

2025-12-19T05:31:54Z

The Cloud Next Door: Investigating the Environmental and Socioeconomic Strain of Datacenters on Local Communities Ngata, Wacuka M; Bashir, Noman; Westerlaken, Michelle; Liote, Laurent; Chandio, Yasra; Olivetti, Elsa Datacenters have become the backbone of modern digital infrastructure, powering the rapid rise of artificial intelligence and promising economic growth and technological progress. However, this expansion has brought growing tensions in the local communities where datacenters are already situated or being proposed. While the mainstream discourse often focuses on energy usage and carbon footprint of the computing sector at a global scale, the local socio-environmental consequences—such as health impacts, water usage, noise pollution, infrastructural strain, and economic burden—remain largely underexplored and poorly addressed. In this work1, we surface these community-level consequences through a mixed-methods study that combines quantitative data with qualitative insights. Focusing on Northern Virginia’s “Data Center Alley,” we highlight how datacenter growth reshapes local environments and everyday life, and examine the power dynamics that determine who benefits and who bears the costs. Our goal is to bring visibility to these impacts and prompt more equitable and informed decisions about the future of digital infrastructure. COMPASS ’25, Toronto, ON, Canada

SuperSONIC: Cloud-Native Infrastructure for ML Inferencing

2025-12-19T05:31:03Z

SuperSONIC: Cloud-Native Infrastructure for ML Inferencing Kondratyev, Dmitry; Riedel, Benedikt; Chou, Yuan-Tang; Cochran-Branson, Miles; Paladino, Noah; Schultz, David; Liu, Mia; Duarte, Javier; Harris, Philip; Hsu, Shih-Chieh The increasing computational demand from growing data rates and complex machine learning (ML) algorithms in large-scale scientific experiments has driven the adoption of the Services for Optimized Network Inference on Coprocessors (SONIC) approach. SONIC accelerates ML inference by offloading it to local or remote coprocessors to optimize resource utilization. Leveraging its portability to different types of coprocessors, SONIC enhances data processing and model deployment efficiency for cutting-edge research in high energy physics (HEP) and multi-messenger astrophysics (MMA). We developed the SuperSONIC project, a scalable server infrastructure for SONIC, enabling the deployment of computationally intensive tasks to Kubernetes clusters equipped with graphics processing units (GPUs). Using NVIDIA Triton Inference Server, SuperSONIC decouples client workflows from server infrastructure, standardizing communication, optimizing throughput, load balancing, and monitoring. SuperSONIC has been successfully deployed for the CMS and ATLAS experiments at the CERN Large Hadron Collider (LHC), the IceCube Neutrino Observatory (IceCube), and the Laser Interferometer Gravitational-Wave Observatory (LIGO) and tested on Kubernetes clusters at Purdue University, the National Research Platform (NRP), and the University of Chicago. SuperSONIC addresses the challenges of the Cloud-native era by providing a reusable, configurable framework that enhances the efficiency of accelerator-based inference deployment across diverse scientific domains and industries. PEARC ’25, Columbus, OH, USA

Systematic engineering of controlled, localized oligonucleotide delivery systems for wound angiogenesis

2025-12-19T04:11:01Z

Systematic engineering of controlled, localized oligonucleotide delivery systems for wound angiogenesis Berger, Adam G. The standard of care for diabetic wounds has remained relatively unchanged for decades, resulting in patients with wounds that do not heal on meaningful time scales, referred to as ulcers, and high rates of recurrence for patients whose wounds do heal. This common complication of diabetes decreases quality of life, increases mortality, and raises health care costs. New paradigms to treat these wounds remains a formidable but critical challenge. Addressing diabetic ulcers at the molecular level may decrease healing time and prevent recurrence. Impaired blood vessel formation, or angiogenesis, in diabetic ulcers is an important target pathway. Angiogenesis is needed to bring oxygen, nutrients, signaling cues, and cells to newly formed tissue while removing waste. Nucleic acid oligonucleotide therapies, such as small interfering RNAs (siRNAs) or microRNA inhibitors (anti-miRs), that regulate gene expression at the post-transcriptional level, hold particular promise for promoting angiogenesis and wound healing; however, the large size and negative charge of these therapies require drug carriers to mediate their biological effect. In this thesis, we leverage sequential electrostatic adsorption of oligonucleotide therapy and polyelectrolytes into thin film coatings on commercial wound dressings through the layer-by-layer (LbL) process. These dressings package oligonucleotide, enhance its transfection efficacy, and control its temporal release locally to the wound bed. After initial validation experiments, we sought to systematically understand our drug carrier system and use this insight to engineer better wound dressings. First, we developed a proof-of-concept anti-miR-coated dressing and showed its efficacy in promoting both wound closure and sex-dependent angiogenesis. We found that therapy released from coated dressings had a preferential association with different wound cell types, particularly endothelial cells. We then sought to uncover how changes in the oligonucleotide structure itself may alter interactions with transfection polymers in thin film coatings. We found that binding with certain polyelectrolytes differed based on whether the therapy was a flexible single stranded anti-miR or a more rigid double stranded helix siRNA. We also showed how chemically modified nucleotides, such as locked nucleic acid and 2’-O-methyl RNA, can modulate affinity to polyelectrolytes and ultimately impact transfection efficacy. We also elucidated how physicochemical properties of the hydrolysable transfection-enhancing poly(β-aminoester) polymer mediate its efficiency in transfecting oligonucleotide therapy. We demonstrated that a more hydrophobic polymer enhanced transfection efficacy through its ability to facilitate permeation of biological barriers. Finally, we identified how modulation of the anionic excipients contained in these thin film coatings can be leveraged to vary the release kinetics from coated wound dressings. We engineered formulations that released on a fast or slow time scale. We observed that while both release time scales promoted efficacy in wound closure, they did so through potentially different mechanisms despite the same putative pro-angiogenic anti-miR therapy. In sum, this thesis elucidates how physicochemical properties and formulation of coated wound dressings alter their interfacial effects with biological systems. We use this knowledge to rationally design better drug carriers that can deliver pro-angiogenic oligonucleotide therapeutics to the wound bed. The findings have broad applications in the delivery of nucleic acid therapies for a wide host of diseases where local delivery to the injured tissue could prove beneficial. Ultimately, we also advance our pro-angiogenic coated wound dressing strategy towards clinical translation. Our strategy has the potential to provide a new, targeted therapeutic paradigm to help those suffering from diabetic ulcers.

What I Don’t Get About AI . . .

2025-12-19T05:31:43Z

What I Don’t Get About AI . . . Wright, Randall S. In a recent MIT News article titled “Explained: Generative AI,” Adam Zewe (Citation2023) writes But what do people really mean when they say ‘generative AI?’ Before the generative AI boom of the past few years, when people talked about AI, typically they were talking about machine-learning models that can learn to make a prediction based on data. For instance, such models are trained, using millions of examples, to predict whether a certain X-ray shows signs of a tumor or if a particular borrower is likely to default on a loan. Generative AI can be thought of as a machine-learning model that is trained to create new data, rather than making a prediction about a specific dataset. A generative AI system is one that learns to generate more objects that look like the data it was trained on.

Bringing a Working-Class Archive Online: Multimodal Storytelling in a Post-Industrial City

2025-12-19T05:31:50Z

Bringing a Working-Class Archive Online: Multimodal Storytelling in a Post-Industrial City Walley, Christine We find it familiar to consider objects as useful or aesthetic, as necessities or vain indulgences. We are on less familiar ground when we consider objects as companions to our emotional lives or as provocations to thought. The notion of evocative objects brings together these two less familiar ideas, underscoring the inseparability of thought and feeling in our relationship to things. We think with the objects we love; we love the objects we think with.

What is a Right?

2025-12-19T05:31:47Z

What is a Right? Setiya, Kieran This paper argues for a theory of natural rights on which they are explained in terms of reasons supplied by rational consent. When B has a claim-right against A that A φ, A’s non-consent is not a reason for B not to simply make A φ. This theory solves a puzzle that defeats alternative views, including standard will and interest theories, the demand theory of rights, and the view that rights are irreducible or primitive.

The Afterlife of Energy: Post-carbon and Feminist Post-work Politics

2025-12-19T05:31:45Z

The Afterlife of Energy: Post-carbon and Feminist Post-work Politics Ghosn, Rania; Vronskaya, Alla; Jia, Ruo; Pohl, Ethel Baraona; Dharia, Namita Vijay; Aidoo, Fallon Samuels; Wolff, Ilze In the conclusion to her book The Birth of Energy: Fossil Fuels, Thermodynamics, and the Politics of Work, political scientist Cara Daggett considers “A Post-Work Energy Politics” in which she examines the historical coupling of energy and work—meaning human, waged work—in an invitation to disassociate their values and futures. The exponential power of fossil fuels animated the pipedream that powerful, inorganic slaves could substitute unfree human labor, ideas that have driven European imperialism. Fossil fuel systems did not lead, however, to a world beyond work. Rather, today’s “patriarchal slave states” continue to manage the project of putting the world to work through the maximization of productivity, and the subordination of racialized, immigrant, and gendered bodies—who would work for lower, or for no, wages. “The project of work,” Daggett argues, “is in tension with the project of life.” 1 And the rise of “work–life balance” is a mere tactic of governance in which the enemy is fatigue, exhaustion, and burn-out. She suggests, in turn, an alliance between post-carbon and feminist post-work politics and asks: what might it mean for energy politics to refer to the politics of ensuring public vitality? In order to advance a feminist revaluation of work, Daggett draws on Kathi Weeks’s The Problem with Work to outline a project that makes two utopian demands. One demand articulates a paradoxical relationship between the pragmatism of (present) demands and the speculative seeds of possibility; a second demand outlines a utopian form for such politics: partial, fragmented kin to the genre of the manifesto. Daggett concludes with an invitation that “a radical planet politics, if it seeks to contest ecomodernist claims, needs its own politics of pleasure.” 2 In an echo to Daggett’s invitation, the authors of this Educators’ Roundtable were invited to contribute a short text that picks up on the possibilities of a post-carbon, post-work politics.

‘May Our Egos Die So That the World May Live’

2025-12-19T05:31:51Z

‘May Our Egos Die So That the World May Live’ Gupta, Huma This image-based essay reflects upon the author’s experience of running an experimental filmmaking workshop titled Climate Futures, Cities Past in the spring of 2023 at MIT’s School of Architecture featuring stills from four student films set in Greece, Italy, Pakistan, and Syria. It explores how architectural pedagogy can intersect with filmmaking to offer a critical space outside the studio or seminar paper. Engaging eco-critical and narrative approaches of Stefanie K. Dunning, Jennifer Fay, Ursula K. Le Guin, Donna Harraway, Saidiya Hartman, Adrian J. Ivakhiv, and Ousmane Sembène, it explores how ‘cinema might teach us to die’ or rather, embrace a different eschatological paradigm that moves beyond individual authorship, accomplishment, and post-mortem legacy towards more mutualist, collectivist, and anarchic models of existence. It argues that filmmaking as inquiry can offer a way to collect different kinds of stories that help facilitate the messy, uncomfortable, and wildly creative processes of unworlding and reworlding.

From Oil to Information: Caudill, Rowlett, Scott and Architectures of the Energy Crisis

2025-12-19T05:31:53Z

From Oil to Information: Caudill, Rowlett, Scott and Architectures of the Energy Crisis Hanly, B. Jack This paper traces the history of architecture-engineering firm Caudill Rowlett Scott (CRS), roughly 1948–1983, in the context of the postwar oil economy and the 1973 energy crisis. The paper examines CRS’s transformation from a design firm into an energy conglomerate over the course of three decades, as it both concretized the fossil economy between Houston and Saudi Arabia and modeled its own corporate structure after its oil clientele. Analyzing numerous CRS projects designed and built for the oil industry, from corporate office towers to industrial training colleges, the paper looks at a moment in which energy systems and the architectural profession were coproduced through the discourses, practices, and institutions of oil at its most vulnerable historical inflection points. CRS thereby epitomized an energy transition from oil as a substance to oil as information, where a growing postindustrial society would leverage the immaterial dimensions of energy as a foundation for building.

Demonstrating Xstrings: 3D Printing Cable-driven Mechanism for Actuation, Deformation, and Manipulation

2025-12-19T05:30:56Z

Demonstrating Xstrings: 3D Printing Cable-driven Mechanism for Actuation, Deformation, and Manipulation Li, Jiaji; Feng, Shuyue; Perroni-Scharf, Maxine; Liu, Yujia; Guan, Emily; Mueller, Stefanie In this Demo, we present Xstrings, a method for designing and fabricating 3D printed objects with integrated cable-driven mechanisms that can be printed in one go without the need for manual assembly. Xstrings supports four types of cable-driven interactions—bend, coil, screw and compress—which are activated by applying an input force to the cables. To facilitate the design of Xstrings objects, we developed a design tool that allows users to embed cable-driven mechanisms into the object geometry based on the desired interaction by automatically placing joints and cables at the respective locations. The application potential of Xstrings is demonstrated through examples such as manipulable gripping, bionic robot manufacturing, and dynamic prototyping. CHI EA ’25, Yokohama, Japan

My CXL Pool Obviates Your PCIe Switch

2025-12-19T05:31:09Z

My CXL Pool Obviates Your PCIe Switch Zhong, Yuhong; Berger, Daniel; Zardoshti, Pantea; Saurez, Enrique; Nelson, Jacob; Psistakis, Antonis; Fried, Joshua; Cidon, Asaf Pooling PCIe devices across multiple hosts offers a promising solution to mitigate stranded I/O resources, enhance device utilization, address device failures, and reduce total cost of ownership. The only viable option today are PCIe switches, which decouple PCIe devices from hosts by connecting them through a hardware switch. However, the high cost and limited flexibility of PCIe switches hinder their widespread adoption beyond specialized datacenter use cases. This paper argues that PCIe device pooling can be effectively implemented in software using CXL memory pools. CXL memory pools improve memory utilization and already have positive return on investment. We find that, once CXL pools are in place, they can serve as a building block for pooling any kind of PCIe device. We demonstrate that PCIe devices can directly use CXL memory as I/O buffers without device modifications, which enables routing PCIe traffic through CXL pool memory. This software-based approach is deployable on today's hardware and is more flexible than hardware PCIe switches. In particular, we explore how disaggregating devices such as NICs can transform datacenter infrastructure. HOTOS 25, May 14–16, 2025, Banff, AB, Canada

PolyMOF nanoparticles constructed from intrinsically microporous polymer ligand towards scalable composite membranes for CO2 separation

2025-12-18T05:49:51Z

PolyMOF nanoparticles constructed from intrinsically microporous polymer ligand towards scalable composite membranes for CO2 separation Lee, Tae Hoon; Lee, Byung Kwan; Yoo, Seung Yeon; Lee, Hyunhee; Wu, Wan-Ni; Smith, Zachary P; Park, Ho Bum Integrating different modification strategies into a single step to achieve the desired properties of metal–organic frameworks (MOFs) has been very synthetically challenging, especially in developing advanced MOF/polymer mixed matrix membranes (MMMs). Herein, we report a polymer–MOF (polyMOF) system constructed from a carboxylated polymer with intrinsic microporosity (cPIM-1) ligand. This intrinsically microporous ligand could coordinate with metals, leading to ~100 nm-sized polyMOF nanoparticles. Compared to control MOFs, these polyMOFs exhibit enhanced ultramicroporosity for efficient molecular sieving, and they have better dispersion properties in casting solutions to prepare MMMs. Ultimately, integrating coordination chemistries through the cPIM-1 and polymer-based functionality into porous materials results in polyMOF/PIM-1 MMMs that display excellent CO2 separation performance (surpassing the CO2/N2 and CO2/CH4 upper bounds). In addition to exploring the physicochemical and transport properties of this polyMOF system, scalability has been demonstrated by converting the developed MMM material into large-area (400 cm2) thin-film nanocomposite (TFN) membranes.

Single Layer Silk and Cotton Woven Fabrics for Acoustic Emission and Active Sound Suppression

2025-12-18T05:49:53Z

Single Layer Silk and Cotton Woven Fabrics for Acoustic Emission and Active Sound Suppression Yang, Grace H; Lin, Jinuan; Cheung, Henry; Rui, Guanchun; Zhao, Yongyi; Balachander, Latika; Joo, Taigyu; Lee, Hyunhee; Smith, Zachary P; Zhu, Lei; Ma, Chu; Fink, Yoel Whether intentionally generating acoustic waves or attempting to mitigate unwanted noise, sound control is an area of challenge and opportunity. This study investigates traditional fabrics as emitters and suppressors of sound. When attached to a single strand of a piezoelectric fiber actuator, a silk fabric emits up to 70 dB of sound. Despite the complex fabric structure, vibrometer measurements reveal behavior reminiscent of a classical thin plate. Fabric pore size relative to the viscous boundary layer thickness is found—through comparative fabric analysis—to influence acoustic‐emission efficiency. Sound suppression is demonstrated using two distinct mechanisms. In the first, direct acoustic interference is shown to reduce sound by up to 37 dB. The second relies on pacifying the fabric vibrations by the piezoelectric fiber, reducing the amplitude of vibration waves by 95% and attenuating the transmitted sound by up to 75%. Interestingly, this vibration‐mediated suppression in principle reduces sound in an unlimited volume. It also allows the acoustic reflectivity of the fabric to be dynamically controlled, increasing by up to 68%. The sound emission and suppression efficiency of a 130 µm silk fabric presents opportunities for sound control in a variety of applications ranging from apparel to transportation to architecture.

Implications of changing the base raw material – the case of license plate manufacturing

2025-12-18T05:49:47Z

Implications of changing the base raw material – the case of license plate manufacturing Uygun, Yilmaz; Mohammadian, Noushin; Un Nisa, Mehr License plates to uniquely identify vehicles mainly use aluminum as the base material. Currently, there is no distinction of different use cases of license plates, such as short-time usage for test drive and transportation purposes that do not need such long-lasting materials not only from cost but also from sustainability perspectives. This paper presents a methodology to select the best material for different use cases under the holistic consideration of specifically defined criteria as to material properties, sustainability aspects, and supply chain implications. We show that there are several candidate materials for different use cases that stick out by changing the importance of these numerous criteria. In addition, the paper delves deeper into the sustainability aspect by means of a comprehensive System Dynamics model. We show that a scenario in which the company picks up used license plates by relying on a logistics service provider to get them delivered to an external recycling service provider yields the best results.

Metalite, a new class of composite laminates with unique properties

2025-12-18T05:49:40Z

Metalite, a new class of composite laminates with unique properties Miravete, Antonio Metalite is a new class of antisymmetric composite laminates composed of angle-plies, 0-degree, and 90-degree plies, presenting unique properties. These include extremely thin laminates suitable for minimum gauge applications, remarkable weight savings compared to conventional quads, adjustable zero and negative coefficients of thermal expansion (CTE), ease of manufacturing, excellent ability to adjust mode frequency, change sound radiation characteristics, and high tunability. In this study, Metalite laminates ranging from 3 to 8 plies are described using their feasible spaces and compared with quads, detailing the weight savings achieved for hard, soft, and neutral laminates. Through an experimental study, the CTE value of a hybrid Metalite is correlated with theory, demonstrating how to tune zero and negative CTE values. The proposed work offers significant benefits through practical solutions for designing and manufacturing lightweight composite laminates with unique properties.

Tertiary-Amine-Functional Poly(arylene ether)s for Acid-Gas Separations

2025-12-18T05:49:50Z

Tertiary-Amine-Functional Poly(arylene ether)s for Acid-Gas Separations Dean, Pablo A; Wu, Yifan; Guo, Sheng; Swager, Timothy M; Smith, Zachary P Competitive sorption enables the emergent phenomenon of enhanced CO₂-based selectivities for gas separation membranes when using microporous polymers with primary amines. However, strong secondary forces in these polymers through hydrogen bonding results in low solvent solubility, precluding standard solution processing approaches to form these polymers into membrane films. Herein, we circumvent these manufacturing constraints while maintaining competitive-sorption enhancements by synthesizing eight representative microporous poly(arylene ether)s (PAEs) with tertiary amines. High-pressure H₂S, CO₂, and CH₄ sorption isotherms were collected for these samples to demonstrate enhanced affinity for acid gases relative to the unfunctional control polymer. Although competitive sorption was observed for all samples, improvements were less pronounced than for primary-amine-functional analogs. For H₂S-based separations, the benefits of competitive sorption offset decreases in selectivity due to plasticization. This detailed study helps to elucidate the role of tertiary amines for acid gas separations in solution-processable microporous PAEs.

Weathering the storm: examining how organisations navigate the sea of cybersecurity regulations

2025-12-18T05:49:49Z

Weathering the storm: examining how organisations navigate the sea of cybersecurity regulations Proudfoot, Jeffrey G; Cram, W Alec; Madnick, Stuart Governments around the world routinely regulate the activities of private enterprises to guide the behaviour of individuals and organisations towards acceptable norms. This holds true in a cybersecurity context. However, practitioners report that cybersecurity regulations are often out of date and compliance is confusing, expensive, and time consuming. As a result, organisational leaders are often uncertain about the practicalities of adopting and implementing the various rules, which can lead to trickle-down effects on the robustness of lower-level cybersecurity controls and compliance activities. In this research, we aim to clarify how cybersecurity regulations are operationalised in organisations, as well as reveal the compliance and performance consequences of cybersecurity regulations. To do so, we interviewed 22 senior leaders with expertise in cybersecurity regulations. Our analysis reveals 7 distinct themes (i.e., concept groupings) that are ordered within four phases (i.e., temporal stages), which we use to create the Institutional Cybersecurity Regulations Model (ICRM). The results provide a holistic view of the cybersecurity regulations process in organisations that can serve to clarify current theory relationships and inform future research. As well, the ICRM can provide a practical roadmap for managers to navigate regulatory cybersecurity challenges in their own companies.

Assessment of Technoeconomic Opportunities in Automation for Nuclear Microreactors

2025-12-18T05:49:45Z

Assessment of Technoeconomic Opportunities in Automation for Nuclear Microreactors Naranjo de Candido, Isabel; Al Rashdan, Ahmad; Abou Jaoude, Abdalla; Buongiorno, Jacopo Achieving full decarbonization of all economic sectors remains a challenge, especially in niche markets. For example, remote communities and industrial or mining activities detached from the main electric grid heavily rely on fossil fuels, similar to urban and industrial microgrids with combined heat and power needs. A combination of renewables and energy storage is often not suitable due to cost, reliability, intermittency, and large storage requirements. Small nuclear reactors with a flexible purpose could serve these applications. Microreactors (MR) are a class of reactors that are compact, factory manufactured, transportable, and self-regulating. Typically, they generate much less power than their large reactor counterparts. The main advantages of microreactors include the versatile nature of the energy produced, the reliability of supply, and freedom from having to transport and store large quantities of fuels on-site, coupled with the absence of dependence on an electrical grid. A strong business case is needed to move from the microreactor prototype to the commercialization phase. In fact, fossil fuels are still relatively inexpensive, and in the near term, carbon credits will be available to virtually compensate for emissions. For microreactors, one of the main costs in operation and maintenance (O&M) is their staffing levels. In this study, we investigate how to optimize the number (and thus the cost) of workers, moving from a traditional, fully manned, on-site personnel approach to an unmanned, remote personnel approach. We examine four different staffing models that can be implemented as the technology matures and evolves. We estimate the staffing needs of each model and build a business case to justify the substitution of on-site personnel with adequate technologies. To do so, we propose a cost model to quantify potential cost reductions from automating O&M activities. The model accounts for both the reduction in cost derived from the reduced number of full-time-equivalent (FTE) employees and the increase in cost derived from the need to buy new control hardware as needed. Applying the cost model that we created to different scenarios, an on-site O&M cost reduction exceeding 80% can be expected. Additionally, we found that it is more impactful to focus on automating routine O&M tasks rather than attempting to automate transient management (shutdowns, restarts, monitoring condition deviations). In fact, transients typically account for less than 1% of the total FTE time spent on the reactors.

Enhancing acid–gas separations using free volume manipulation for microporous poly(arylene ether)s

2025-12-18T05:49:37Z

Enhancing acid–gas separations using free volume manipulation for microporous poly(arylene ether)s Joo, Taigyu; Wu, Yifan; Lee, Tae Hoon; Dean, Pablo A; Wu, Wan-Ni; Swager, Timothy M; Smith, Zachary P To address global energy needs, traditional and renewable natural gas will likely be key energy sources for years to come. However, raw feeds require removal of impurities like hydrogen sulfide (H2S) and carbon dioxide (CO2) before use. In this study, we illustrate the key challenges of using traditional post-synthetic modification approaches to simultaneously enhance H2S/CH4 and CO2/CH4 selectivities in microporous polymer membranes, while also demonstrating how free volume manipulation (FVM) can overcome some of these challenges. By integrating tert-butoxycarbonyl-protected piperazinyl (PIP-tBOC) groups into a microporous poly(arylene ether) (PAE-1) and applying thermal treatment with oxygen to degrade the incorporated units in solid-state films, we successfully increased sorption capacity and diffusion selectivity. This modification enhanced the mixed-gas selectivity of H2S/CH4 and CO2/CH4 by 88% and 114%, respectively, compared to the original PAE-1 films. Consequently, the films achieved a combined acid gas (CAG) selectivity of 48, which approached the CAG upper bound for glassy polymers. The FVM process not only improved the selectivity of these membrane films but also markedly increased their resistance to plasticization, making them more suitable for industrial applications in acid–gas separation. This post-synthetic modification strategy, applicable to any glassy polymer containing a nucleophilic aromatic unit, provides a means to leverage the competitive sorption of H2S molecules and the molecular sieving properties of the polymer.

Knowledge and ignorance in forensic identification: the origins of a contested human rights fact

2025-12-18T05:49:39Z

Knowledge and ignorance in forensic identification: the origins of a contested human rights fact Medina, Eden In 2006, DNA testing revealed that the Chilean Medical Legal Service had misidentified at least half of the 96 human rights victims whose remains had been exhumed in 1991 from a lot in the Santiago General Cemetery known as Patio 29. Years earlier the government had returned those remains to the victims' families. This examination of the history of that forensic misidentification uncovers the role played by the shifting relations of knowledge and ignorance in establishing the legal facts of those identities. Building on the growing literature in agnotology, the article demonstrates the ways in which the context of dictatorship created varied and overlapping forms of ignorance that continued to shape the outcome of the forensic work even after Chile returned to democracy. By detailing different examples of ignorance production by the state, a human rights organization, and a university department under military surveillance, the article illuminates the diverse ways that the civil–military dictatorship worked against knowledge production in the domains of science and human rights.

Tweeting during the Pandemic in New York City: Unveiling the Evolving Sentiment Landscape of NYC through a Spatiotemporal Analysis of Geolocated Tweets

2025-12-18T05:50:00Z

Tweeting during the Pandemic in New York City: Unveiling the Evolving Sentiment Landscape of NYC through a Spatiotemporal Analysis of Geolocated Tweets Ignaccolo, Carmelo; Wibisono, Kevin; Sutto, Maria Paola; Plunz, Richard A. This article explores the relationship between spatial factors, socioeconomic conditions, and Twitter (now called X) sentiment in New York City (NYC) during the COVID-19 pandemic. Using Twitter data, the study investigates how sentiment varied across different geographies. It examines whether sentiment scores, unemployment rates, and COVID-19 hospitalization rates in NYC zip codes revealed spatial associations. The research employs sentiment analysis, a natural language processing technique used to algorithmically determine the emotional tone of a text, on a database of geo-located tweets spanning January to December 2020. The findings reveal a shift towards more negative sentiment during the initial year of the pandemic. Moreover, the study uncovers variations in sentiment trends across boroughs and zip codes. Additionally, a zip code-level fixed-effects model demonstrates a statistically significant relationship between sentiment scores and unemployment rates. In summary, this article makes a two-fold contribution: firstly, it adds a spatial lens to the scholarly debate regarding the use of Twitter data as an indicator of publicly expressed sentiment; secondly, it provides empirical evidence on the spatial interconnectedness of sentiment, health (hospitalization), and socioeconomic factors (unemployment). Overall, this research sheds light on the nuanced relationship between sentiment and space during the COVID-19 pandemic in NYC.

Solution‐Processable, Ladder‐Branched Polyimides of Intrinsic Microporosity by [4+4] Cycloaddition for Membrane Gas Separation

2025-12-18T05:49:54Z

Solution‐Processable, Ladder‐Branched Polyimides of Intrinsic Microporosity by [4+4] Cycloaddition for Membrane Gas Separation Lee, Tae Hoon; Dean, Pablo A; Yeo, Jing Ying; Smith, Zachary P Advancements in membrane-based gas separation have the potential to address global challenges related to energy and the environment. However, new membrane materials must have excellent separation performance, stability, and processability, and simultaneously achieving all three metrics is extremely challenging. To circumvent these issues, a post-synthetic modification of polyimides of intrinsic microporosity (PIM-PIs) synthesized with a UV light (UV)-reactive anthracene co-monomer is reported. UV irradiation on the PIM-PI solution converts the anthracene units into dianthracene linkages by [4+4] cycloaddition, while the resultant PIM-PI is still solution-processable due to the branched structure. The ladder-like dianthracene moieties significantly increased both microporosity (<20 Å) and ultramicroporosity (<7 Å) of the precursor PIM-PI. Notably, the UV-treated PIM-PI membrane exhibits a large boost in pure-gas CO2 permeability by up to 260%, reaching 376 barrer, while maintaining CO2/CH4 ideal selectivity of 35 at 1 bar. Moreover, the developed membrane material has enhanced stability against physical aging and plasticization and showcases excellent CO2/CH4 mixed-gas selectivity (>30 up to 31 bar feed pressure), which surpasses the 2018 mixed-gas upper bound.

Interactive infill topology optimisation guided by user drawn patterns

2025-12-18T05:50:02Z

Interactive infill topology optimisation guided by user drawn patterns Schiffer, Gillian; Schmidt, Martin-Pierre; Pedersen, Claus BW; Carstensen, Josephine V Widespread use of topology optimisation as a design tool for additive manufacturing faces major inhibiting obstacles, such as high computational costs and complexity, concern for other failure modes, and manufacturability. Interactive infill topology optimisation presents an alternative approach to circumvent some of these barriers. The novel contribution of the present work prompts the user to draw a tailored infill pattern, specify regions of interest to locate the infill, and control how strictly the pattern is replicated in the material layout of the design using appearance constraints. This approach improves engineering metrics not directly included in the optimisation formulation by incorporating the user’s engineering experience, thereby avoiding increased computational costs, parameter tuning, and numerical artifacts associated with complex objective functions and constraints. Two 2D benchmark examples increase the linear buckling resistance and energy absorption, respectively, and a 2.5D example minimises compliance while reducing the quantity of overhang supports for additive manufacturing.

Minimum Bucket and Car Battery Problems

2025-12-18T05:49:43Z

Minimum Bucket and Car Battery Problems Feng, Raymond A solar car needs 5 fully charged batteries to run, and it depletes those batteries in 5 hours. The batteries are rechargeable, and solar panels on the car are able to charge 3 batteries simultaneously. It takes 3 hours for the solar panels to finish charging 3 batteries. Furthermore, batteries cannot be charging and in use at the same time. If the car always starts running as soon as 5 full batteries are available, and the solar panels can only operate if 3 empty batteries are available, how many batteries are needed so that the car can eventually run without stopping? We investigate this resource optimization problem and its different variations.

Developers Grappling with Flood Risks: Evaluating Boston’s Climate Resiliency Checklist

2025-12-18T05:49:56Z

Developers Grappling with Flood Risks: Evaluating Boston’s Climate Resiliency Checklist Loescher-Montal, Angela; Mazereeuw, Miho; Shen, Kairos Ongoing waterfront development in risky areas across the globe raises the continued paradox between resilience initiatives and broader market mechanisms. Even as flood risk increases, existing development patterns do not often adequately account for future flood risk. This research examines using resiliency checklists as a growing regulatory tool to improve predevelopment flood resilience standards. The research employs mixed quantitative and qualitative methods to evaluate how four large-scale developments interacted with Boston’s Climate Resiliency Checklist in the last decade and how its current design criteria influenced design decisions. The checklist’s format, design, and time horizon considerations are evaluated. Increased format and smaller-scale tools are considered.

The “content” of intergroup contact: lessons from the Denton Women’s Interracial Fellowship

2025-12-18T05:49:42Z

The “content” of intergroup contact: lessons from the Denton Women’s Interracial Fellowship English, Jasmine Does the content of intergroup contact matter? Despite extensive research on the benefits of contact for intergroup relations, we know little about what happens during contact-based programs and interventions. This article addresses this gap by inductively building theory about the desired content of contact. My analysis draws on oral history interviews and archival data from the Denton Women’s Interracial Fellowship: a real-world case of intergroup contact that emerged to ease the process of school desegregation in Denton, Texas. My analysis of these data moves beyond the scope conditions suggested by (Allport, Gordon W. 1954. The Nature of Prejudice. 25th ed. Cambridge, MA: Perseus Books) to highlight the role of conversations about outgroup experiences. I illuminate how these conversations produce positive impacts on intergroup relations and draw out the implications for research on intergroup contact: namely, that forms of intergroup contact that incorporate these conversations are more likely to improve intergroup relations, and that intergroup contact interventions should explicitly encourage or incorporate these kinds of conversations.

Roadmap of Graphite Moderator and Graphite-Matrix TRISO Fuel Management Options

2025-12-18T05:50:03Z

Roadmap of Graphite Moderator and Graphite-Matrix TRISO Fuel Management Options Forsberg, CW Most high-temperature reactors use graphite as a moderator and structural material. This includes high-temperature gas-cooled reactors with helium cooling and TRi-structural ISOtropic (TRISO) fuel particles embedded in graphite, as well as fluoride salt–cooled high-temperature reactors with clean salt coolant and TRISO fuel particles embedded in graphite and thermal spectrum molten salt reactors with a graphite moderator and fuel dissolved in the salt. The largest volume radioactive waste stream from these reactors is the irradiated graphite. We describe herein a roadmap for management of these graphite wastes that contain radioactive 14C, tritium, and other radionuclides. There may be some graphite wastes with sufficiently low radioactivity levels that can be treated as nonradioactive waste and managed like other graphite waste. Management options for the graphite include (1) direct disposal, (2) recycled back to the reactor or other nuclear applications, and (3) oxidizing the graphite with release as an effluent or underground sequestration of the carbon dioxide. Cosequestration of this carbon dioxide with carbon dioxide from industrial, biological, and cement production processes can isotopically dilute the 14C before sequestration to eliminate the possibility of exceeding individual radiation exposure limits. We also describe options for processing graphite-matrix TRISO fuel, including separating the bulk graphite to reduce the volumes of used fuel for disposal or processing to recover fissile materials. The inventories of radioactive isotopes in different carbon wastes vary by many orders of magnitude; thus, there is no single economic option for the management of all graphite waste.

Adaptive Model Reduction of High-Order Solutions of Compressible Flows via Optimal Transport

2025-12-18T05:49:58Z

Adaptive Model Reduction of High-Order Solutions of Compressible Flows via Optimal Transport Van Heyningen, Robert Loek; Nguyen, Ngoc Cuong; Blonigan, Patrick; Peraire, Jaime The solution of conservation laws with parametrised shock waves presents challenges for both high-order numerical methods and model reduction techniques. We introduce an r-adaptivity scheme based on optimal transport and apply it to develop reduced order models for compressible flows. The optimal transport theory allows us to compute high-order r-adaptive meshes from a starting reference mesh by solving the Monge–Ampère equation. A high-order discretization of the conservation laws enables high-order solutions to be computed on the resulting r-adaptive meshes. Furthermore, the Monge–Ampère solutions contain mappings that are used to reduce the spatial locality of the resulting solutions and make them more amenable to model reduction. We use a non-intrusive model reduction method to construct reduced order models of both the mesh and the solution. The procedure is demonstrated on three supersonic and hypersonic test cases, with the hybridisable discontinuous Galerkin method being used as the full order model.

Demonstrating Thermochromorph: Dynamic Relief Printing with Thermochromic Inks

2025-12-18T05:49:35Z

Demonstrating Thermochromorph: Dynamic Relief Printing with Thermochromic Inks Sethapakdi, Ticha; Myers, Paris; Yu, Tianyu; Covarrubias, Juliana; Leake, Mackenzie; Mueller, Stefanie We demonstrate Thermochromorph, a novel relief printing technique that produces multicolored images that transition into each other through changes in temperature. Our process utilizes two sets of CMYK thermochromic inks that exhibit complementary color-changing behaviors: one shifting from color to transparency, the other from transparency to color at the same activation temperature. We describe our printmaking workflow, provide an open-source software toolkit, showcase prints made with our system, and explore how our system can be used in creative practice through an artist workshop. By incorporating new materials and technology with the rich history of printmaking, our work extends the expressive capabilities of relief printing as the medium continues to evolve. CHI EA ’25, Yokohama, Japan

Demonstrating OpenEarable 2.0: An AI-Powered Ear Sensing Platform

2025-12-18T05:49:04Z

Demonstrating OpenEarable 2.0: An AI-Powered Ear Sensing Platform R?ddiger, Tobias; Zitz, Valeria; Hummel, Jonas; K?ttner, Michael; Lepold, Philipp; King, Tobias; Paradiso, Joseph; Clarke, Christopher; Beigl, Michael In this demo, we present OpenEarable 2.0, an open-source earphone platform designed to provide an interactive exploration of physiological ear sensing and the development of AI applications. Attendees will have the opportunity to explore real-time sensor data and understand the capabilities of OpenEarable 2.0’s sensing components. OpenEarable 2.0 integrates a rich set of sensors, including two ultrasound-capable microphones (inward/outward), a 3-axis ear canal accelerometer/bone conduction microphone, a 9-axis head inertial measurement unit, a pulse oximeter, an optical temperature sensor, an ear canal pressure sensor, a microSD slot, and a microcontroller. Participants will be able to try out the web-based dashboard and mobile app for real-time control and data visualization. Furthermore, the demo will show different applications and real-time data based on OpenEarable 2.0 across physiological sensing and health monitoring, movement and activity tracking, and human-computer interaction. CHI EA ’25, Yokohama, Japan

Conductive Ceramics: Embedding Electronics in Everyday Ceramic Objects

2025-12-18T05:49:02Z

Conductive Ceramics: Embedding Electronics in Everyday Ceramic Objects Chin, Sam; Kim, Keunwook; An, Audrey; Kuang, Quincy; Zhang, Kai We present a method for integrating conductive traces into ceramic objects using a silver-based glaze compatible with traditional firing processes. Our glaze combines silver powder with a glass former and xanthan gum, enabling application through standard ceramic techniques while maintaining the durability of conventional ceramics. Through a material-driven experimentation approach, we characterized how glaze composition and post-processing methods affect conductivity and surface quality. We demonstrate this technique through functional prototypes including a temperature-responsive heating vessel, a touch-sensitive musical controller utilizing kintsugi repair, and an interactive marble machine. This work bridges traditional ceramic craft with interactive technology, offering ceramicists a way to incorporate electronic functionality while preserving traditional methods. CHI EA ’25, Yokohama, Japan

Towards Resource-Efficient Compound AI Systems

2025-12-18T05:48:25Z

Towards Resource-Efficient Compound AI Systems Chaudhry, Gohar Irfan; Choukse, Esha; Goiri, ??igo; Fonseca, Rodrigo; Belay, Adam; Bianchini, Ricardo Compound AI Systems, integrating multiple interacting components like models, retrievers, and external tools, have emerged as essential for addressing complex AI tasks. However, current implementations suffer from inefficient resource utilization due to tight coupling between application logic and execution details, a disconnect between orchestration and resource management layers, and the perceived exclusiveness between efficiency and quality. We propose a vision for resource-efficient Compound AI Systems through a declarative workflow programming model and an adaptive runtime system for dynamic scheduling and resource-aware decision-making. Decoupling application logic from low-level details exposes levers for the runtime to flexibly configure the execution environment and resources, without compromising on quality. Enabling collaboration between the workflow orchestration and cluster manager enables higher efficiency through better scheduling and resource management. We are building a prototype system, called Murakkab, to realize this vision. Our preliminary evaluation demonstrates speedups up to ~ 3.4× in workflow completion times while delivering ~ 4.5× higher energy efficiency, showing promise in optimizing resources and advancing AI system design. HOTOS 25, May 14–16, 2025, Banff, AB, Canada

Fits like a Flex-Glove: Automatic Design of Personalized FPCB-Based Tactile Sensing Gloves

2025-12-18T05:48:58Z

Fits like a Flex-Glove: Automatic Design of Personalized FPCB-Based Tactile Sensing Gloves Murphy, Devin; Li, Yichen; Owens, Crystal; Stanton, Layla; Liang, Paul Pu; Luo, Yiyue; Torralba, Antonio; Matusik, Wojciech Resistive tactile sensing gloves have captured the interest of researchers spanning diverse domains, such as robotics, healthcare, and human-computer interaction. However, existing fabrication methods often require labor-intensive assembly or costly equipment, limiting accessibility. Leveraging flexible printed circuit board (FPCB) technology, we present an automated pipeline for generating resistive tactile sensing glove design files solely from a simple hand photo on legal-size paper, which can be readily supplied to commercial board houses for manufacturing. Our method enables cost-effective, accessible production at under $130 per glove with sensor assembly times under 15 minutes. Sensor performance was characterized under varying pressure loads, and a preliminary user evaluation showcases four unique automatically manufactured designs, evaluated for their reliability and comfort. CHI EA ’25, Yokohama, Japan

Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMs

2025-12-18T05:49:05Z

Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMs Khan, Ariba; Casper, Stephen; Hadfield-Menell, Dylan Research on the ‘cultural alignment’ of Large Language Models (LLMs) has emerged in response to growing interest in understanding representation across diverse stakeholders. Current approaches to evaluating cultural alignment through survey-based assessments that borrow from social science methodologies often overlook systematic robustness checks. We identify and test three assumptions behind current survey-based evaluation methods: (1) Stability: that cultural alignment is a property of LLMs rather than an artifact of evaluation design, (2) Extrapolability: that alignment with one culture on a narrow set of issues predicts alignment with that culture on others, and (3) Steerability: that LLMs can be reliably prompted to represent specific cultural perspectives. Through experiments examining both explicit and implicit preferences of leading LLMs, we find a high level of instability across presentation formats, incoherence between evaluated versus held-out cultural dimensions, and erratic behavior under prompt steering. We show that these inconsistencies can cause the results of an evaluation to be very sensitive to minor variations in methodology. Finally, we demonstrate in a case study on evaluation design that narrow experiments and a selective assessment of evidence can be used to paint an incomplete picture of LLMs’ cultural alignment properties. Overall, these results highlight significant limitations of current survey-based approaches to evaluating the cultural alignment of LLMs and highlight a need for systematic robustness checks and red-teaming for evaluation results. Data and code are available at https://doi.org/akhan02/cultural-dimension-cover-letters and https://doi.org/ariba-k/llm-cultural-alignment-evaluation, respectively. FAccT ’25, Athens, Greece

Aptly: Making Mobile Apps from Natural Language

2025-12-18T05:48:46Z

Aptly: Making Mobile Apps from Natural Language Patton, Evan; Kim, David; Granquist, Ashley; Liu, Robin; Scott, Arianna; Zamanova, Jennet; Abelson, Harold This paper introduces Aptly, a platform designed to democratize mobile app development, particularly for young learners. Aptly integrates a Large Language Model (LLM) with App Inventor, enabling users to create apps using their natural language. User’s description is translated into a programming language that corresponds with App Inventor’s visual blocks. A preliminary study with high school students demonstrated the usability and potential of the platform. Prior programming experience influenced how users interact with Aptly. Participants identified areas for improvement and expressed a shift in perspective regarding programming accessibility and AI’s role in creative endeavors. CHI EA ’25, Yokohama, Japan

AcceloPrint: Fabricating Customizable Accelerometers with Multi-Material 3D Printing

2025-12-18T05:48:43Z

AcceloPrint: Fabricating Customizable Accelerometers with Multi-Material 3D Printing Ozbek, Doga; AlAlawi, Marwa; Wessely, Michael We introduce AcceloPrint, 3D-printed acceleration sensors that can be fabricated in one pass alongside a 3D object and report on its angular orientation or acceleration. AcceloPrint utilizes capacitive sensing to track the deflection of a 3D printed cantilever beam to a sensor patch. Our AcceloPrint tool integrated into a 3D editor generates a sensor with a user-defined sensing range generated by our computational model. We also propose a novel sensor design with an adjustable sensing range post-fabrication. Our technical evaluation shows our sensor can detect acceleration up to 50 m/s2, with a root mean squared error of 0.35 m/s2 (%3.57) in the range up to 10 m/s2. We demonstrate AcceloPrint with three application examples on sports performance tracking and tangible tools. CHI EA ’25, Yokohama, Japan

"How can we learn and use AI at the same time?": Participatory Design of GenAI with High School Students

2025-12-18T05:48:40Z

"How can we learn and use AI at the same time?": Participatory Design of GenAI with High School Students Pu, Isabella; Ravi, Prerna; Dinh, Linh; Joe, Chelsea; Ogoe, Caitlin; Li, Zixuan; Breazeal, Cynthia; Ostrowski, Anastasia As generative AI (GenAI) emerges as a transformative force, clear understanding of high school students’ perspectives is essential for GenAI’s meaningful integration in high school environments. In this work, we draw insights from a participatory design workshop where we engaged 17 high school students—a group rarely involved in prior research in this area—through the design of novel GenAI tools and school policies addressing their key concerns. Students identified challenges and developed solutions outlining their ideal features in GenAI tools, appropriate school use, and regulations. These centered around the problem spaces of combating bias & misinformation, tackling crime & plagiarism, preventing over-reliance on AI, and handling false accusations of academic dishonesty. Building on our participants’ underrepresented perspectives, we propose new guidelines targeted at educational technology designers for development of GenAI technologies in high schools. We also argue for further incorporation of student voices in development of AI policies in their schools. IDC ’25, Reykjavik, Iceland

Bird: A Point Cursor for Virtual Immersive Environments

2025-12-18T05:48:42Z

Bird: A Point Cursor for Virtual Immersive Environments Simonson, Aubrey; Gretton, Dana; Harteveld, Casper This paper introduces the Bird, a novel point cursor for immersive virtual environments (IVEs) that enables precise, one-handed control over a point in 3D space beyond arm’s reach. Interaction techniques commonly used in VR today lack this functionality. While direct manipulation allows for control of the position of an object in 3D space, it is limited to arm’s reach. Ray-casting enables interaction at a distance but specifies a line rather than a point, making it impossible to move objects closer or farther without additional mechanics. The Bird overcomes these limitations by allowing users to select any visible object and place it anywhere within view, with one hand and without requiring a controller. We explore a range of use cases that highlight the Bird’s potential to expand the design space for spatial computing. CHI EA ’25, Yokohama, Japan

On Approximability of Satisfiable 𝑘-CSPs: V

2025-12-18T05:48:59Z

On Approximability of Satisfiable 𝑘-CSPs: V Bhangale, Amey; Khot, Subhash; Minzer, Dor We propose a framework of algorithm vs. hardness for all Max-CSPs and demonstrate it for a large class of predicates. This framework extends the work of Raghavendra [STOC, 2008], who showed a similar result for almost satisfiable Max-CSPs. Our framework is based on a new hybrid approximation algorithm, which uses a combination of the Gaussian elimination technique (i.e., solving a system of linear equations over an Abelian group) and the semidefinite programming relaxation. We complement our algorithm with a matching dictator vs. quasirandom test that has perfect completeness. The analysis of our dictator vs. quasirandom test is based on a novel invariance principle, which we call the mixed invariance principle. Our mixed invariance principle is an extension of the invariance principle of Mossel, O’Donnell and Oleszkiewicz [Annals of Mathematics, 2010] which plays a crucial role in Raghavendra’s work. The mixed invariance principle allows one to relate 3-wise correlations over discrete probability spaces with expectations over spaces that are a mixture of Guassian spaces and Abelian groups, and may be of independent interest. STOC ’25, Prague, Czechia

Fewer Than 1% of Explainable AI Papers Validate Explainability with Humans

2025-12-18T05:48:45Z

Fewer Than 1% of Explainable AI Papers Validate Explainability with Humans Suh, Ashley; Hurley, Isabelle; Smith, Nora; Siu, Ho Chit This late-breaking work presents a large-scale analysis of explainable AI (XAI) literature to evaluate claims of human explainability. We collaborated with a professional librarian to identify 18,254 papers containing keywords related to explainability and interpretability. Of these, we find that only 253 papers included terms suggesting human involvement in evaluating an XAI technique, and just 128 of those conducted some form of a human study. In other words, fewer than 1% of XAI papers (0.7%) provide empirical evidence of human explainability when compared to the broader body of XAI literature. Our findings underscore a critical gap between claims of human explainability and evidence-based validation, raising concerns about the rigor of XAI research. We call for increased emphasis on human evaluations in XAI studies and provide our literature search methodology to enable both reproducibility and further investigation into this widespread issue. CHI EA ’25, Yokohama, Japan

Characteristics of Driver Peripheral Vision: How Drivers Respond to Ubiquitous Information on Wide-Area In-Vehicle Displays

2025-12-18T05:48:47Z

Characteristics of Driver Peripheral Vision: How Drivers Respond to Ubiquitous Information on Wide-Area In-Vehicle Displays Huang, Hongwei; Li, Jiateng; Feng, Xuejing; Ma, Jun; Mehler, Bruce Despite advancements in In-Vehicle Information Systems (IVIS) and extensive research on screen layouts, the influence of drivers’ peripheral vision on interactions with evolving multi-screen and large display technologies remains poorly understood. This study examines drivers’ responses to in-vehicle interactive information through peripheral vision, aiming to optimize visual interaction efficiency and enhance driving safety. Analyzing data from 216 participants in a driving simulator, we explored how horizontal eccentricity, screen type, cognitive load, visual crowding, and stimulus type affect perception rates and reaction times. Our findings highlight the significance of these factors and the need for driver-centered design. The results suggest designing IVIS that align with natural visual tendencies to improve interaction efficiency and driving safety. CHI EA ’25, Yokohama, Japan

Guiding Evolutionary AutoEncoder Training with Activation-Based Pruning Operators

2025-12-18T05:48:34Z

Guiding Evolutionary AutoEncoder Training with Activation-Based Pruning Operators Jorgensen, Steven; Hemberg, Erik; Toutouh, Jamal; O'Reilly, Una-May This study explores a novel approach to neural network pruning using evolutionary computation, focusing on simultaneously pruning the encoder and decoder of an autoencoder. We introduce two new mutation operators that use layer activations to guide weight pruning. Our findings reveal that one of these activation-informed operators outperforms random pruning, resulting in more efficient autoencoders with comparable performance to canonically trained models. Prior work has established that autoencoder training is effective and scalable with a spatial coevolutionary algorithm that cooperatively coevolves a population of encoders with a population of decoders, rather than one autoencoder. We evaluate how the same activity-guided mutation operators transfer to this context. We find that random pruning is better than guided pruning, in the coevolutionary setting. This suggests activation-based guidance proves more effective in low-dimensional pruning environments, where constrained sample spaces can lead to deviations from true uniformity in randomization. Conversely, population-driven strategies enhance robustness by expanding the total pruning dimensionality, achieving statistically uniform randomness that better preserves system dynamics. We experiment with pruning according to different schedules and present best combinations of operator and schedule for the canonical and coevolving populations cases. GECCO ’25, July 14–18, 2025, Malaga, Spain

Mirai: A Wearable Proactive AI "Inner-Voice" for Contextual Nudging

2025-12-18T05:48:37Z

Mirai: A Wearable Proactive AI "Inner-Voice" for Contextual Nudging Fang, Cathy Mengying; Samaradivakara, Yasith; Maes, Pattie; Nanayakkara, Suranga People often find it difficult to turn their intentions into real actions—a challenge that affects both personal growth and mental well-being. While established methods like cognitive-behavioral therapy and mindfulness training help people become more aware of their behaviors and set clear goals, these approaches cannot provide immediate guidance when people fall into automatic reactions or habits. We introduce Mirai, a novel wearable AI system with an integrated camera, real-time speech processing, and personalized voice-cloning to provide proactive and contextual nudges for positive behavior change. Mirai continuously monitors and analyzes the user’s environment to anticipate their intentions, generating contextually-appropriate responses delivered in the user’s own cloned voice. We demonstrate the application of Mirai through three scenarios focusing on dietary choices, work productivity, and communication skills. We also discuss future work on improving the proactive agent via human feedback and the need for a longitudinal study in naturalistic settings. CHI EA ’25, Yokohama, Japan

9.301J Neural Plasticity in Learning and Development, Spring 2002

2025-12-17T00:09:23Z

9.301J Neural Plasticity in Learning and Development, Spring 2002 Miller, Earl K; Liu, Guosong; Wilson, Matthew A.; Tonegawa, Susumu; Quinn, William Roles of neural plasticity in learning and memory and in development of invertebrates and mammals. An in-depth critical analysis of current literature of molecular, cellular, genetic, electrophysiological, and behavioral studies. Discussion of original papers supplemented by introductory lectures.

9.110J Neurology, Neuropsychology, and Neurobiology of Aging. Spring 2005

2025-12-17T00:04:31Z

9.110J Neurology, Neuropsychology, and Neurobiology of Aging. Spring 2005 Corkin, Suzanne Hammond; Ingram, Vernon M. Lectures and discussions in this course cover the clinical, behavioral, and molecular aspects of the brain aging processes in humans. Topics include the loss of memory and other cognitive abilities in normal aging, as well as neurodegenerative conditions such as Parkinson’s and Alzheimer’s diseases. Discussions based on readings taken from primary literature explore the current research in this field.

9.18 Developmental Neurobiology, Spring 2005

2025-12-17T00:00:42Z

9.18 Developmental Neurobiology, Spring 2005 Nedivi, Elly This course considers molecular control of neural specification, formation of neuronal connections, construction of neural systems, and the contributions of experience to shaping brain structure and function. Topics include: neural induction and pattern formation, cell lineage and fate determination, neuronal migration, axon guidance, synapse formation and stabilization, activity-dependent development and critical periods, development of behavior.

9.013J Cell and Molecular Neurobiology, Spring 2008

2025-12-17T17:54:04Z

9.013J Cell and Molecular Neurobiology, Spring 2008 Constantine-Paton, Martha; Sheng, Morgan Hwa-Tze; Quinn, William This course explores the major areas of cellular and molecular neurobiology, including excitable cells and membranes, ion channels and receptors, synaptic transmission, cell-type determination, axon guidance, neuronal cell biology, neurotrophin signaling and cell survival, synapse formation and neural plasticity. Material includes lectures and exams, and involves presentation and discussion of primary literature. It focuses on major concepts and recent advances in experimental neuroscience.

9.322J Genetic Neurobiology, Fall 2005

2025-12-16T23:50:34Z

9.322J Genetic Neurobiology, Fall 2005 Littleton, J. Troy; Quinn, William This course deals with the specific functions of neurons, the interactions of neurons in development, and the organization of neuronal ensembles to produce behavior. Topics covered include the analysis of mutations, and molecular analysis of the genes required for nervous system function. In particular, this course focuses on research work done with nematodes, fruit flies, mice, and humans.

9.19J Cognitive & Behavioral Genetics, Spring 2001

2025-12-16T23:46:07Z

9.19J Cognitive & Behavioral Genetics, Spring 2001 Nedivi, Elly; Pinker, Steven How genetics can add to our understanding of cognition, language, emotion, personality, and behavior. Use of gene mapping to estimate risk factors for psychological disorders and variation in behavioral and personality traits. Mendelian genetics, genetic mapping techniques, and statistical analysis of large populations and their application to particular studies in behavioral genetics. Topics also include environmental influence on genetic programs, evolutionary genetics, and the larger scientific, social, ethical, and philosophical implications

5.95J Teaching College-Level Science and Engineering, Spring 2009

2025-12-16T23:41:01Z

5.95J Teaching College-Level Science and Engineering, Spring 2009 Mahajan, Sanjoy This participatory seminar focuses on the knowledge and skills necessary for teaching science and engineering in higher education. This course is designed for graduate students interested in an academic career, and anyone else interested in teaching. Readings and discussions include: teaching equations for understanding, designing exam and homework questions, incorporating histories of science, creating absorbing lectures, teaching for transfer, the evils of PowerPoint, and planning a course. The subject is appropriate for both novices and those with teaching experience.

7.51 Biochemistry, Fall 2001

2025-12-17T17:55:19Z

7.51 Biochemistry, Fall 2001 Sauer, Robert T; Solomon, Frank; Baker, Tania The tools and analytical methods that biochemists use to dissect biological problems. Analysis of the mode of action and structure of regulatory, binding, and catalytic proteins.

7.391 Concept-Centered Teaching, Fall 2005

2025-12-16T23:19:27Z

7.391 Concept-Centered Teaching, Fall 2005 Khodor, Julia Do you like teaching, but find yourself frustrated by how little students seem to learn? Would you like to try teaching, but are nervous about whether you will be any good at it? Are you interested in new research on science education? Research in science education shows that the greatest obstacle to student learning is the failure to identify and confront the misconceptions with which the students enter the class or those that they acquire during their studies. This weekly seminar course focuses on developing the participants’ ability to uncover and confront student misconceptions and to foster student understanding and retention of key concepts. Participants read primary literature on science education, uncover basic concepts often overlooked when teaching biology, and lead a small weekly discussion session for students currently enrolled in introductory biology classes. The instructor for this course, Dr. Julia Khodor, is a member of the HHMI Education Group.

7.391 Concept-Centered Teaching, Spring 2006

2025-12-16T23:14:32Z

7.391 Concept-Centered Teaching, Spring 2006 Kosinski-Collins, Melissa Do you like teaching, but find yourself frustrated by how little students seem to learn? Would you like to try teaching, but are nervous about whether you will be any good at it? Are you interested in new research on science education? Research in science education shows that the greatest obstacle to student learning is the failure to identify and confront the misconceptions with which the students enter the class or those that they acquire during their studies. This weekly seminar course focuses on developing the participants’ ability to uncover and confront student misconceptions and to foster student understanding and retention of key concepts. Participants read primary literature on science education, uncover basic concepts often overlooked when teaching biology, and lead a small weekly discussion session for students currently enrolled in introductory biology classes. The instructor for this course, Dr. Kosinski-Collins, is a member of the HHMI Education Group.

20.010J Introduction to Bioengineering (BE.010J), Spring 2006

2025-12-16T23:10:25Z

20.010J Introduction to Bioengineering (BE.010J), Spring 2006 Lauffenburger, Douglas A; Matsudaira, Paul T.; Belcher, Angela M Bioengineering at MIT is represented by the diverse curricula offered by most Departments in the School of Engineering. This course samples the wide variety of bioengineering options for students who plan to major in one of the undergraduate Engineering degree programs. The beginning lectures describe the science basis for bioengineering with particular emphasis on molecular cell biology and systems biology. Bioengineering faculty will then describe the bioengineering options in a particular engineering course as well as the type of research conducted by faculty in the department,

7.347 Epigenetic Regulation of Stem Cells, Spring 2014

2025-12-16T22:53:33Z

7.347 Epigenetic Regulation of Stem Cells, Spring 2014 Williams, Eric O; Subramanian, Vidya During development a single totipotent cell gives rise to the vast array of cell types present in the adult human body, yet each cell has essentially the same DNA sequence. As cells differentiate, distinct sets of genes must be coordinately activated and repressed, ultimately leading to a cell-type specific pattern of gene expression and a particular cell fate. In eukaryotic organisms, DNA is packaged in a complex protein super structure known as chromatin. Modification and reorganization of chromatin play a critical role in coordinating the cell-type specific gene expression programs that are required as a cell transitions from a pluripotent stem cell to a fully differentiated cell type. Epigenetics refers to such heritable changes that occur in chromatin without altering the primary DNA sequence. This class will focus on the role of epigenetic regulation with respect to developmental fate and also consider the fact that the epigenetic mechanisms discussed have broad implications, including how seemingly normal cells can be transformed into cancerous cells. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

Structural Tuning of Self‐Conductive Polymer as Gas Diffusion Layer for Electrocatalytic Reactions at High Current

2026-03-08T03:38:50Z

Structural Tuning of Self‐Conductive Polymer as Gas Diffusion Layer for Electrocatalytic Reactions at High Current Noh, Hwiyoon; Lee, Tae Hoon; Ahn, Sang Hyun; Davis, Jonathan T; Jeong, Daecheol; Gounder, Rajamani; Smith, Zachary P; Boudouris, Bryan W; Tackett, Brian M Electrocatalytic conversions offer a promising route for sustainable chemical production using renewable energy. Gas diffusion layers (GDLs) enable selective product formation at high current densities but suffer from electrolyte flooding, and polytetrafluoroethylene (PTFE)-based GDLs typically require metal conductive layers, which constrain catalyst development. A recently developed GDL configuration, electropolymerized poly(3,4-ethylenedioxythiophene) (PEDOT)-coated PTFE, demonstrates notable flooding resistance, but suffers from gas diffusion limitations at elevated currents due to limited gas diffusion through the PEDOT layer. Here, different dopants in PEDOT are exploited to modify the physical properties and enhance gas transport. ClO4−-doped PEDOT exhibits superior performance due to optimized physical structure, leading to increased gas permeance and faradaic efficiency (FE) for CO production during electrocatalytic CO2 reduction. Further optimization of coverage and thickness achieved by adjusting charge density led to an optimal configuration at 33 mC cm−2. This GDL supports various metal electrocatalysts and demonstrates FECO of > 90% for over 150 h at −200 mA cm−2 using a commercial silver electrocatalyst. This work highlights the importance of GDL engineering in enhancing performance and durability for long-term electrocatalytic processes.

Navigating Emotions Through Art

2026-03-08T03:22:26Z

Navigating Emotions Through Art Wu, Christine; Kumar, Ila; Picard, Rosalind In this study, we design and deploy a novel system to examine the safety and efficacy of using a chatbot to conduct aspects of art therapy with youth who have experienced developmental trauma, focusing on supporting emotion identification, processing, and expression. This publication describes phase one, gathering feedback on the system from practicing art therapists (n = 17) and making recommendations for how to evolve such work in beneficial ways to meet the needs of trauma-impacted youth. Our findings highlight the potential value of chatbots for trauma-impacted youth as well as important reflection questions these chatbots should ask. Additionally, the study discusses the risk of harm associated with chatbot interventions, particularly if the conversation brings up negative emotions that the chatbot fails to help process. Finally, we end by presenting a set of practitioner-driven recommendations for chatbot designers who are interested in helping trauma-impacted youth understand and cope with their emotions, leveraging art therapy techniques. CHI EA ’25, Yokohama, Japan

Bridging Tradition and Technology: Human-AI Interface for Exploration and Co-Creation of Classical Dance Heritage

2026-03-08T03:22:20Z

Bridging Tradition and Technology: Human-AI Interface for Exploration and Co-Creation of Classical Dance Heritage Pataranutaporn, Pat; Archiwaranguprok, Chayapatr; Bhongse-tong, Piyaporn; Maes, Pattie; Klunchun, Pichet This paper introduces Text2Tradition, a system designed to bridge the epistemological gap between modern language processing and traditional dance knowledge by translating user-generated prompts into Thai classical dance repertoire. Our system interprets user prompts through the lens of Mae Bot Yai—the 59 foundational movements constituting the vocabulary of traditional Thai dance—and incorporates six choreographic elements that encode centuries of cultural knowledge. This research explores the fertile tension between two knowledge systems: the embodied, culturally-specific wisdom of traditional dance and the data-driven, statistically-derived, and often Western-centric intelligence of LLMs. By mediating between these epistemologies, we highlight the potential of AI-mediated systems not only to preserve traditional forms but also to foster new cultural co-creations, suggesting that these tensions can be harnessed to stimulate cultural dialogue and innovation. CHI EA ’25, Yokohama, Japan

A Write-Optimized Distributed B+Tree Index on Disaggregated Memory

2026-03-08T03:21:59Z

A Write-Optimized Distributed B+Tree Index on Disaggregated Memory Kraska, Tim If it were possible to scale memory independently from compute, it would be feasible to dynamically adjust the amount of memory based on the workload. It would further enable better resource utilization. Consider a dynamic workload regarding the number of queries but with very strict response time requirements, which can only be met, if data is kept in-memory. In this case, the separation of compute and memory would enable to scale the compute with the number of queries while keeping all the data constantly in-memory. This design principle is already used by services such as Google, which keeps the entire web-index in-memory.

Data-Driven AI Avatars for Valuation in Dating Scenarios

2026-03-08T03:22:10Z

Data-Driven AI Avatars for Valuation in Dating Scenarios Baradari, D?nya; Polimetla, Tejaswi; Maes, Pattie Dating applications facilitate partner selection by presenting curated information about potential matches. However, traditional dating profiles often fail to convey the depth of a person’s personality, communication style, and lived experience, leading to inefficiencies in the match-finding process. This work-in-progress study introduces and evaluates two novel, data-driven dating interfaces: (1) a Data Dashboard, which aggregates and visualizes insights from a user’s digital footprint, and (2) an AI Avatar, an interactive, voice-enabled model using personal data to simulate real-world interactions. A user study with nine participants comparing these interfaces against traditional dating profiles reveals that the Data Dashboard enables more accurate personality assessments but imposes a high cognitive load. Meanwhile, the AI Avatar enhances engagement and enjoyability but raises concerns about trust and emotional investment. Our findings highlight the challenge of maintaining authenticity in AI-mediated interactions and bridging the gap between digital and real-life personas. CHI EA ’25, Yokohama, Japan

Cultivating a Supportive Sphere: Designing Technology to Increase Social Support for Foster-Involved Youth

2026-03-08T03:22:15Z

Cultivating a Supportive Sphere: Designing Technology to Increase Social Support for Foster-Involved Youth Kumar, Ila; Ferguson, Craig; Wu, Jiayi; Picard, Rosalind Approximately 400,000 youth in the US are living in foster care due to experiences with abuse or neglect at home [17]. For multiple reasons, these youth often don’t receive adequate social support from those around them. Despite technology’s potential, very little work has explored how these tools can provide more support to foster-involved youth. To begin to fill this gap, we worked with current and former foster-involved youth to develop the first digital tool that aims to increase social support for this population, creating a novel system in which users complete reflective check-ins in an online community setting. We then conducted a pilot study with 15 current and former foster-involved youth, comparing the effect of using the app for two weeks to two weeks of no intervention. We collected qualitative and quantitative data, which demonstrated that this type of interface can provide youth with types of social support that are often not provided by foster care services and other digital interventions. The paper details the motivation behind the app, the trauma-informed design process, and insights gained from this initial evaluation study. Finally, the paper concludes with recommendations for designing digital tools that effectively provide social support to foster-involved youth.

Allocation Multiplicity: Evaluating the Promises of the Rashomon Set

2026-03-08T03:22:46Z

Allocation Multiplicity: Evaluating the Promises of the Rashomon Set Jain, Shomik; Wang, Margaret; Creel, Kathleen; Wilson, Ashia The Rashomon set of equally-good models promises less discriminatory algorithms, reduced outcome homogenization, and fairer decisions through model ensembles or reconciliation. However, we argue from the perspective of allocation multiplicity that these promises may remain unfulfilled. When there are more qualified candidates than resources available, many different allocations of scarce resources can achieve the same utility. This space of equal-utility allocations may not be faithfully reflected by the Rashomon set, as we show in a case study of healthcare allocations. We attribute these unfulfilled promises to several factors: limitations in empirical methods for sampling from the Rashomon set, the standard practice of deterministically selecting individuals with the lowest risk, and structural biases that cause all equally-good models to view some qualified individuals as inherently risky. FAccT ’25, Athens, Greece

Interactive Sketchpad: A Multimodal Tutoring System for Collaborative, Visual Problem-Solving

2026-03-08T03:22:21Z

Interactive Sketchpad: A Multimodal Tutoring System for Collaborative, Visual Problem-Solving Lee, Jimin; Chen, Steven-Shine; Liang, Paul Pu Humans have long relied on visual aids like sketches and diagrams to support reasoning and problem-solving. Visual tools, like auxiliary lines in geometry or graphs in calculus, are essential for understanding complex ideas. However, many tutoring systems remain text-based, providing feedback only through natural language. Leveraging recent advances in Large Multimodal Models (LMMs), this paper introduces Interactive Sketchpad, a tutoring system that combines language-based explanations with interactive visualizations to enhance learning. Built on a pre-trained LMM, Interactive Sketchpad is fine-tuned to provide step-by-step guidance in both text and visuals, enabling natural multimodal interaction with the student. Accurate and robust diagrams are generated by incorporating code execution into the reasoning process. User studies conducted on math problems such as geometry, calculus, and trigonometry demonstrate that Interactive Sketchpad leads to improved task comprehension, problem-solving accuracy, and engagement levels, highlighting its potential for transforming educational technologies. All code is available at: https://stevenshinechen.github.io/interactivesketchpad/. CHI EA ’25, Yokohama, Japan

Meeting at Crossroads: An exploration of playful listening through a co-creative AI game

2026-03-08T03:22:07Z

Meeting at Crossroads: An exploration of playful listening through a co-creative AI game Lee, Cassandra; Dimitrakopoulou, Dimitra; Roy, Deb Active listening is a well-established cornerstone of empathetic communication and a hallmark of “civic competence”, but is a challenging and energy consuming skill. Games offer a provocative lens to consider how active listening could be explored playfully. In this paper, we present Crossroads, an interactive social game which makes active listening fun by inviting players to co-create images about one another’s personal experiences. Deployed through a tablet-mobile web app, players take turns acting in ‘listener roles’ to generate AI images, and eventually uncover a collective picture along a “crossroad” shaped map. An initial mixed-method evaluation with 36 users demonstrates that players find the experience highly engaging and feel especially heard during in-game conversations. This work contributes a novel game which uses AI to mediate empathetic dialogue, and surfaces questions about the trade-offs of gamifying listening. CHI EA ’25, Yokohama, Japan

Atmospheric Impacts of Hydrogen as an Aviation Fuel

2025-12-17T03:06:42Z

Atmospheric Impacts of Hydrogen as an Aviation Fuel Gibney, Evan M. Hydrogen is being investigated as a promising zero-carbon aviation fuel, offering the potential to eliminate direct CO₂ emissions while being produced with low lifecycle greenhouse gas emissions. Despite these benefits, there are additional indirect climate and air quality costs associated with direct hydrogen emissions which are often overlooked. We quantify the perturbation in the atmospheric composition associated with the introduction of hydrogen-fueled aircraft, broadening the current understanding of the non-CO₂ effects of these fleets. We use the GEOS-Chem High Performance (GCHP) global chemistry-transport model to conduct a spatially discretized, multi-year impact assessment of the atmospheric impacts of hydrogen-fueled aviation. We implement a flux surface boundary condition for hydrogen to provide an improved representation of the soil sink, relative to the default fixed boundary condition. This results in a net surface exchange of-16.7 Tg H₂ per year. Two hydrogen scenarios are evaluated using the updated GCHP implementation, which are representative of a high and low mitigation scenario for direct hydrogen emission rates. For the two scenarios, respectively, we observe increases in the mean atmospheric methane mixing ratio of 3.34 ppbv and 10.7 ppbv, corresponding to an increased methane lifetime of between 0.24% and 0.77%, respectively. The increased methane lifetime as well as in-situ oxidation of stratospheric hydrogen results in an increased stratospheric water vapor burden of 0.42 Tg and 2.3 Tg (or 0.052% and 0.28%) for the high and low mitigation scenarios, respectively. Additionally, we show the perturbation to tropospheric ozone levels to be between-0.047% and +0.30%, where the decreased ozone results from the removal of NOₓ emissions associated with fuel cells and low hydrogen emission rates. This analysis provides the foundation for understanding the implications of potential future hydrogen-based aviation fleets on climate and air quality.

A Changing Climate Beneath Our Feet: How plant and microbial life in tropical soils are shifting and what that could mean for the future of our warming planet

2025-12-17T03:06:34Z

A Changing Climate Beneath Our Feet: How plant and microbial life in tropical soils are shifting and what that could mean for the future of our warming planet Ocharoenchai, Nanticha Discussions about climate change and carbon sequestration have largely revolved around plant structures we can easily see, like leaves that absorb CO₂ for photosynthesis and woody trunks that store carbon as biomass. Carbon credits that companies and consumers buy to compensate for emissions they’ve produced are primarily calculated based on these parts, as are models that predict climate change impacts. But researchers are now beginning to understand that what we see aboveground is only part of the equation. The other part lies beneath our feet in an intricate, expansive, covert realm where plant roots, microbial communities and soil dynamics interact. These belowground systems are crucial for cycling carbon through the Earth and regulating the climate, but relatively little is known about them compared to aboveground systems. This is especially true in tropical regions, where one-third of the world’s terrestrial carbon storage lies. However, these systems are evolving quickly with climate change, contradicting what models have previously projected. With so many global decisions based on such models, these uncertainties hold planetary significance for our future. A group of scientists is climbing an uphill battle, racing against time to understand this understudied field.

Engineering Winter

2025-12-17T03:06:43Z

Engineering Winter White, Mackenzie As winters warm and snowfall becomes less reliable, ski resorts worldwide increasingly depend on artificial snow to stay open. Snowmaking, once a stopgap, has become the backbone of entire seasons in a sprawling choreography of pumps and pressurized mist designed to hold trails together. At resorts like Vermont’s Bromley Mountain, snowmakers work through the night, drawing millions of gallons from limited reservoirs and operating within narrowing windows of cold air. What emerges is a portrait of winter in transition: less predictable, more expensive, increasingly manufactured. The efforts to preserve winter recreation carry growing costs in energy, water, and equitable access. Many smaller, independent ski areas struggle to meet the demands of climate adaptation, while larger resorts expand their operations, widening the divide in who can afford to sustain operations. In the American West, where rivers depend heavily on snowpack melt, the spread of snowmaking ties winter recreation to a water system already under immense strain. As artificial snow becomes the norm, winter is increasingly a season bought, built, and rationed, raising the question of whether attempts to keep the season alive are accelerating the changes that threaten to erase it.

Influence of Electronic Structure and Lattice Dynamics on Oxygen Ion Transport in Solid-State Ionic Conductors

2025-12-17T03:03:54Z

Influence of Electronic Structure and Lattice Dynamics on Oxygen Ion Transport in Solid-State Ionic Conductors Vivona, Daniele Solid-state oxygen ion conductors are crucial for electrochemical devices such as separation membranes, solid-oxide electrolyzers, fuel cells, and sensors, serving as a technological link between renewable energy generation and consumption. Currently, these conductors are limited by slow transport rates and high operational temperatures, which pose challenges and increase costs. Developing faster conductors that operate at lower temperatures requires reducing activation energy and enhancing the pre-exponential factor in the Arrhenius equation of conductivity. However, our understanding of the fundamental processes in oxygen ion transport and methods to improve oxygen ion conductivity remain limited. This thesis focuses on understanding the fundamental mechanisms that regulate oxygen ion transport. First, the migration energy barrier in perovskite oxides is linked to an electronic energy penalty from local charge screening near the hopping ion. The energy of local electronic states is identified as a fundamental descriptor of the migration barrier. Next, migration entropy and phonon density of states (DOS) are highlighted as the main factors regulating the pre-exponential factor of oxygen ion conductivity across different materials. The phonons of oxygen ions near the hopping ion significantly contribute to migration entropy, suggesting that migration entropy can be tuned by designing the phonon dynamics of these atoms. These results imply that a widely observed correlation between increasing pre-exponential factors and activation energy arises from coupling local electronic energy states and phonons. The results are extended to the formation of oxygen vacancies and interstitials in perovskite and RuddlesdenPopper oxides. We find that defect formation energy rises with defect formation entropy, which is linked to electronic energy states interacting with phonons. In perovskite oxides, lower vacancy formation entropy is correlated with increasing oxygen phonon band center and shortening bond lengths with oxygen vacancy formation. In Ruddlesden-Popper oxides, lower interstitial formation entropy is associated with reduced octahedral tilting and local phonon changes. This thesis establishes a theoretical foundation for treating migration entropy and defect formation entropy as design variables in next-generation ionic conductors. By highlighting the impact of electronic structure and lattice dynamics on energy barriers and entropic drivers, the findings suggest new pathways for material design through the strategic separation of these factors and the intelligent design of lattice moieties in oxygen ion transport environments.

7.344 Antibiotics, Toxins, and Protein Engineering, Spring 2007

2025-12-16T22:59:01Z

7.344 Antibiotics, Toxins, and Protein Engineering, Spring 2007 Koehrer, Caroline; Sassanfar, Mandana The lethal poison Ricin (best known as a weapon of bioterrorism), Diphtheria toxin (the causative agent of a highly contagious bacterial disease), and the widely used antibiotic tetracycline have one thing in common: They specifically target the cell’s translational apparatus and disrupt protein synthesis. In this course, we will explore the mechanisms of action of toxins and antibiotics, their roles in everyday medicine, and the emergence and spread of drug resistance. We will also discuss the identification of new drug targets and how we can manipulate the protein synthesis machinery to provide powerful tools for protein engineering and potential new treatments for patients with devastating diseases, such as cystic fibrosis and muscular dystrophy. This course is one of many Advanced Undergraduate Seminars offered by the Biology Department at MIT. These seminars are tailored for students with an interest in using primary research literature to discuss and learn about current biological research in a highly interactive setting. Many instructors of the Advanced Undergraduate Seminars are postdoctoral scientists with a strong interest in teaching.

OpenMC Interpretation of FNS SINBAD Shielding Benchmark Experiments

2026-03-08T03:38:47Z

OpenMC Interpretation of FNS SINBAD Shielding Benchmark Experiments Ebiwonjumi, Bamidele; Segantin, Stefano; Peterson, Ethan The Fusion Neutron Source (FNS) clean benchmark experiments on tungsten, vanadium, and beryllium assemblies from the SINBAD (Shielding Integral Benchmark Archive and Database) are analyzed to experimentally validate OpenMC (version 0.14.1-dev) fusion neutronics capabilities. The assemblies were irradiated with a 14-MeV deuterium-tritium neutron source. Neutron spectra, photon spectra, reaction rates, gamma heating rates (GHRs), and tritium production rates (TPRs) are compared to measured data in the experimental assemblies and MCNP-6.2 results. In the tungsten case, slight overestimations of the experimental data were observed in the neutron spectra, and the photon spectra agreed well with the experiments. Most of the GHRs agreed with the measured data within the range of experimental uncertainty in the tungsten and vanadium assemblies. In the vanadium assembly, the calculated neutron spectra underestimated the experiments in the low energy region while the photon spectra were well calculated when compared to experiments. The most noticeable discrepancies with experimental data in the gamma heating were observed at detector positions closest to the source. For the reaction rates, notable discrepancies with experimental data were seen at the front and rear of the assemblies. Compared to experiments, the OpenMC neutron spectra were well predicted in the beryllium assembly, whereas the calculated fission reaction rate and TPRs overestimated the experiments, an observation similar to that which has been reported by other authors. The average, overall calculation-to-experiment ratio (C/E) over nine TPR and seven GHR measurements were 1.03 ± 0.20 and 0.95 ± 0.14, respectively. In the case of verification, the OpenMC results of the benchmark calculations indicated comparable accuracy to MCNP-6.2. In general, the validation exercise showed that OpenMC can be used to analyze the fusion neutronics shielding benchmark problems.

SoK: Acoustic Side Channels

2026-03-08T03:31:57Z

SoK: Acoustic Side Channels Wang, Ping; Nagaraja, Shishir; Bourquard, Aur?lien; Gao, Haichang; Yan, Jeff Acoustic side channels (ASCs) have been discovered for several decades, highlighting the tangible security risks posed by unintended sound emissions from computing and electronic systems. Their existence has drawn considerable attention from researchers, driving rapid progress in both attack methodologies and defense mechanisms across a wide range of scenarios. In this paper, we provide a state-of-the-art analysis of ASCs, covering all the significant academic research in the area. First, we clarify existing ambiguities and conceptual confusion, proposing a clear definition of ASC. Second, we analyse the characteristics of known ASCs, discuss their security implications, and propose the first taxonomy. Next, we summarize attack techniques, discuss countermeasures, and identify areas for future research. We also link side channels and inverse problems, two fields that appear to be completely isolated from each other but have deep connections.

Umm Kamel’s Affair: How Infidelity Liberated the Night Sky in Jabal ‘Amil

2026-03-08T03:38:46Z

Umm Kamel’s Affair: How Infidelity Liberated the Night Sky in Jabal ‘Amil Nahleh, Mohamad Weakened by the expansion of several imperial and colonial projects, the inhabitants of Jabal ‘Amil survived as second-class citizens, severed from the urban expression of Lebanese nationalism, and having to formulate their identity amid countless transgressions on their scholarship and literary production. It is thus in the spectacles of the universe and the mysteries of the cosmos that they inscribed fragments of their oral legacy, turning the night sky into an archive that no empire could burn or colonize. And yet it is light pollution, leaking from the same cities they were once forced to nourish, that quickly established itself as the main transgressor, clearing the faintest stories in their celestial library. Although distant manifestations of Islamic cosmology could no longer animate their rural nights, new alterations in the sky after dark, no matter how violent, have proven worthy carri-ers of their modern myths and legends. And it is onto the loudest object in their polluted sky, the Israeli reconnaissance drone IAI Searcher MK, that they grafted the tale of their legendary matriarch Umm Kamel. I argue that Umm Kamel’s physical and symbolic ascent into the sky was orchestrated by a modern generation of ‘Amilis whose infidelity to the celestial stories authored by their ancestors fortified their ability to transform the combined pressures of pollution and colonization. United by their efforts to forge new imaginaries around a starless night, they invite reflection on the possibility (and responsibility) of confronting the sky we have together inherited rather than lamenting the one we have lost. In tracing Umm Kamel’s transformation from figure to constellation, I contend that their cosmic interventions set the stage for new alliances between design and darkness, and ultimately, for a more expanded imagination of night design, particularly within the context of the climate crisis.

Ophthalmology Optical Coherence Tomography Databases for Artificial Intelligence Algorithm: A Review

2026-03-08T03:38:47Z

Ophthalmology Optical Coherence Tomography Databases for Artificial Intelligence Algorithm: A Review Restrepo, David; Quion, Justin Michael; Do Carmo Novaes, Frederico; Azevedo Costa, Iago Diogenes; Vasquez, Constanza; Bautista, Alyssa Nicole; Quiminiano, Ellaine; Lim, Patricia Abigail; Mwavu, Roger; Celi, Leo Anthony; Nakayama, Luis Filipe BACKGROUND: Imaging plays a pivotal role in eye assessment. With the introduction of advanced machine learning and artificial intelligence (AI), the focus has shifted to imaging datasets in ophthalmology. While disparities and health inequalities hidden within data are well-documented, the ophthalmology field faces specific challenges to the creation and maintenance of datasets. Optical Coherence Tomography (OCT) is useful for the diagnosis and monitoring of retinal pathologies, making it valuable for AI applications. This review aims to identify and compare the landscape of publicly available optical coherence tomography databases for AI applications. METHODS: We conducted a literature review on OCT and AI articles with publicly accessible datasets, using PubMed, Scopus, and Web of Science databases. The review retrieved 183 articles, and after full-text analysis, 50 articles were included. From the included articles were identified 8 publicly available OCT datasets, focusing on patient demographics and clinical details for thorough assessment and comparison. RESULTS: The resulting datasets encompass 154,313 images collected from Spectralis, Cirrus HD, Topcon 3D, and Bioptigen devices. These datasets included normal exams, age-related macular degeneration, and diabetic maculopathy, among others. Comprehensive demographic information is available in one dataset and the USA is the most represented population. DISCUSSION: Current publicly available OCT databases for AI applications exhibit limitations, stemming from their non-representative nature and the lack of comprehensive demographic information. Limited datasets hamper research and equitable AI development. To promote equitable AI algorithmic development in ophthalmology, there is a need for the creation and dissemination of more representative datasets.

Market Design for Capacity Sharing in Networks

2026-03-08T03:31:54Z

Market Design for Capacity Sharing in Networks Amin, Saurabh; Jaillet, Patrick; Pulyassary, Haripriya; Wu, Manxi We study a market mechanism that sets edge prices to incentivize strategic agents to efficiently share limited network capacity. In this market, agents form coalitions, with each coalition sharing a unit capacity of a selected route and making payments to cover edge prices. Our focus is on the existence and computation of market equilibrium, where challenges arise from the interdependence between coalition formation among strategic agents with heterogeneous preferences and route selection that induces a network flow under integral capacity constraints. To address this interplay between coalition formation and network capacity utilization, we introduce a novel approach based on combinatorial auction theory and network flow theory. We establish sufficient conditions on the network topology and agents' preferences that guarantee both the existence and polynomial-time computation of a market equilibrium. Additionally, we identify a particular market equilibrium that maximizes utilities for all agents and is equivalent to the classical Vickrey-Clarke-Groves mechanism. Furthermore, we extend our results to multi-period settings and general networks, showing that when the sufficient conditions are not met, an equilibrium may still exist but requires more complex, path-based pricing mechanisms that set differentiated prices based on agents' preference parameters.

Forage: Understanding RAG-based Sensemaking for Community Conversations

2026-03-08T03:22:24Z

Forage: Understanding RAG-based Sensemaking for Community Conversations Schroeder, Hope; Beeferman, Doug; Detwiller, Maya; Dimitrakopoulou, Dimitra; Roy, Deb We introduce Forage, a RAG-based and LLM-augmented search engine, which we apply to the problem of sensemaking for community conversation data. We report on formative user studies introducing Forage to two distinct user groups: NPR journalists and municipal staff in the city of Durham, North Carolina. We taxonomize the query types users make with the tool, use cases that include synthesizing insights across conversations and finding content about a particular subject. We find that users tend to gravitate towards using the system for synthesis more than for pure search. We report on challenges and opportunities surfaced by performing sensemaking with an open-ended interface like Forage, such as the benefits of finding content quickly, but also the challenges users face interacting with a system in natural language. Insights from this formative study confirm the usefulness of Forage for sensemaking, but also make follow-up work, such as systematically evaluating system performance and developing appropriate design, urgent. CHI EA ’25, Yokohama, Japan

Generative artificial intelligence in supply chain and operations management: a capability-based framework for analysis and implementation

2026-03-08T03:38:45Z

Generative artificial intelligence in supply chain and operations management: a capability-based framework for analysis and implementation Jackson, Ilya; Ivanov, Dmitry; Dolgui, Alexandre; Namdar, Jafar This research examines the transformative potential of artificial intelligence (AI) in general and Generative AI (GAI) in particular in supply chain and operations management (SCOM). Through the lens of the resource-based view and based on key AI capabilities such as learning, perception, prediction, interaction, adaptation, and reasoning, we explore how AI and GAI can impact 13 distinct SCOM decision-making areas. These areas include but are not limited to demand forecasting, inventory management, supply chain design, and risk management. With its outcomes, this study provides a comprehensive understanding of AI and GAI's functionality and applications in the SCOM context, offering a practical framework for both practitioners and researchers. The proposed framework systematically identifies where and how AI and GAI can be applied in SCOM, focussing on decision-making enhancement, process optimisation, investment prioritisation, and skills development. Managers can use it as a guidance to evaluate their operational processes and identify areas where AI and GAI can deliver improved efficiency, accuracy, resilience, and overall effectiveness. The research underscores that AI and GAI, with their multifaceted capabilities and applications, open a revolutionary potential and substantial implications for future SCOM practices, innovations, and research.

Remote Direct Code Execution

2026-03-08T03:31:55Z

Remote Direct Code Execution Huang, Yibo; Qiu, Yiming; Ding, Daqian; Kon, Patrick Tser Jern; Zhang, Yiwen; Mao, Yuzhou; Bhatnagar, Archit; Chowdhury, Mosharaf; Devadas, Srinivas; Xing, Jiarong; Chen, Ang We propose remote direct code execution (RDX), which elevates the power of RDMA from memory access to code execution. We target runtime extension frameworks such as Wasm filters, BPF programs, and UDF functions, where RDX enables an agentless architecture that unlocks capabilities such as fast extension injection, update consistency guarantees, and minimal resource contention. We outline the roadmap for RDX around a new CodeFlow abstraction, encompassing programming remote extensions, exposing management stubs, remotely validating and JIT compiling code, seamlessly linking code to local context, managing remote extension state, and synchronizing code to targets. The case studies and initial results demonstrate the feasibility of RDX and its potential to spark the next wave of RDMA innovations. HotNets ’25, College Park, MD, USA

User Adoption of Intelligent Environments: A Review of Technology Adoption Models, Challenges, and Prospects

2026-03-08T03:38:44Z

User Adoption of Intelligent Environments: A Review of Technology Adoption Models, Challenges, and Prospects FakhrHosseini, Shabnam; Chan, Kathryn; Lee, Chaiwoo; Jeon, Myounghoon; Son, Heesuk; Rudnik, John; Coughlin, Joseph Recent technological advancements have enabled the development of smarter (more automated) and more intelligent (adaptable) environments. To understand what factors lead users to reject or adopt Intelligent Environments (IEs), we reviewed nine prominent technology adoption theories. We conducted a literature review to investigate the acceptance and adoption of different types of IEs. We found that perceived usefulness, ease of use, perceived control or self-efficacy, affect and enjoyment, and perceived risks are the common factors across the studies explaining the adoption of IEs. However, shortcomings in the design and methods of the reviewed studies present major concerns in the generalizability and application of existing theories to emerging IEs. We identify eight lacunae in the existing literature and propose a new conceptual model for explaining the adoption of IEs. Through this study, we contribute to the formulation of the theoretical background for the successful introduction of IEs and their integration into users’ everyday life.

Safe and Secure Control of Connected and Automated Vehicles: An Event-Triggered Control Approach using Trust-Aware Robust Control Barrier Functions

2026-03-08T03:31:56Z

Safe and Secure Control of Connected and Automated Vehicles: An Event-Triggered Control Approach using Trust-Aware Robust Control Barrier Functions Ahmad, H M SABBIR; Sabouni, Ehsan; Xiao, Wei; Cassandras, Christos; Li, Wenchao We address the security of a network of Connected and Automated Vehicles (CAVs) cooperating to safely navigate through a conflict area (e.g., traffic intersections, merging roadways, roundabouts). Previous studies have shown that such a network can be targeted by adversarial attacks causing traffic jams or safety violations resulting in collisions. We focus on attacks targeting the V2X communication network used to share vehicle data and consider as well uncertainties due to noise in sensor measurements and communication channels. To combat these, motivated by recent work on the safe control of CAVs, we propose a trust-aware robust event-triggered decentralized control and coordination framework that can provably guarantee safety. We maintain a trust metric for each vehicle in the network computed based on their behavior and used to balance the tradeoff between conservativeness (when deeming every vehicle as untrustworthy) while guaranteeing safety and performance. It is important to highlight that our framework is invariant to the specific choice of the trust framework. Based on this framework, we propose an attack detection and mitigation scheme which has twofold benefits: (i) the trust framework is immune to false positives, and (ii) it provably guarantees safety against false positive cases which may arise from a poor choice of trust framework. We use extensive simulations in SUMO and CARLA to validate the theoretical guarantees and demonstrate the efficacy of our proposed scheme to detect and mitigate adversarial attacks. The code for the simulated scenarios can be found in this \href{https://github.com/SabbirAhmad26/Trust_based_CBF}{\textit{\underline{link}}}.

Three-dimensional, soft magnetic-cored solenoids via multi-material extrusion

2026-03-08T03:38:45Z

Three-dimensional, soft magnetic-cored solenoids via multi-material extrusion Cañada, Jorge; Kim, Hyeonseok; Velásquez-García, Luis Fernando This study reports fully 3D-printed, three-dimensional, soft magnetic-cored solenoids that generate three times the largest magnetic fields previously reported from 3D-printed solenoids. The devices are fabricated on a customised, multi-material 3D printer that can extrude both filaments and pellets. Three different kinds of materials are employed to manufacture the reported soft magnetic-cored solenoids: pure PLA (dielectric portions), PLA doped with copper particles (electrically conductive structures), and nylon or PLA doped with metallic particles (soft magnetic cores). Via manufacturing optimisation, the reported devices are 33% smaller and can withstand about twice the current, generating three times more magnetic field. The 3D-printed solenoids generate Gauss-level magnetic fields while drawing tens-of-milliamps currents and can be readily used to implement fully 3D-printed induction sensors. The results of this work extend the state of the art in 3D-printed electronics, enabling the creation of more complex and capable solenoids for in-situ manufactured and in-space manufactured electromagnetic systems.

The Churns and Turns of HCI: Which CHI Papers Make the Most Impact in an Ever-growing Sea of HCI Publications

2026-03-08T03:22:23Z

The Churns and Turns of HCI: Which CHI Papers Make the Most Impact in an Ever-growing Sea of HCI Publications Kaltenhauser, Annika; Sch?ning, Johannes; Churchill, Elizabeth; Ishii, Hiroshi; Mekler, Elisa; Shneiderman, Ben The ACM Conference on Human Factors in Computing Systems (CHI) is the premier venue for research in Human-Computer Interaction (HCI). 11,290 full papers have been published and collectively cited almost one million times. Highly cited papers undoubtedly represent influential work, affecting the creation of review standards and conference submission and acceptance practices within and beyond CHI. However, the factors contributing to high citation counts and what constitutes a highly cited CHI paper remain largely unclear. In this panel discussion, we will engage the CHI community in exploring the relationship between paper characteristics, citation numbers, and effective impact on HCI as a discipline, and on HCI as an influential endeavour in technology design and development. To ground this discussion, we present findings from a literature review of the 100 most cited CHI full papers, looking at past and present fields and subfields of influence. We will also share insights from HCI experts. Our goals are to shed light on the meaning of impactful work at CHI and in HCI more broadly, to reflect on key trends in HCI over the years, and to discuss themes that have driven pivotal shifts in HCI research. We will lead the conversation toward a deeper understanding of citation practices, the role of citations in focusing and driving HCI research, and the implications of citation when it comes to shaping what is considered impactful HCI. CHI EA ’25, Yokohama, Japan

Safeguards and Security for High-Burnup TRISO Pebble Bed Spent Fuel and Reactors

2026-03-08T03:38:49Z

Safeguards and Security for High-Burnup TRISO Pebble Bed Spent Fuel and Reactors Forsberg, Charles; Kadak, Andrew Several high-temperature thermal neutron–spectrum pebble bed reactors are being commercialized. China has started up two helium-cooled pebble bed high-temperature reactors. In the United States, the X-Energy helium-cooled and the Kairos Power salt-cooled pebble bed high-temperature reactors will produce spent nuclear fuel (SNF) with burnups exceeding 150 000 MWd per tonne. The reactor fuel in each case consists of small spherical graphite pebbles (4 to 6 cm in diameter) containing thousands of small TRISO (microspheric tri-structural isotropic) fuel particles embedded in the fuel of zone these pebbles. The unique isotopic, chemical, and physical characteristics of this high-burnup SNF create a technical case to eliminate safeguards based on the low risk for use in nuclear weapons, while maintaining safeguards in terms of risk for use in radiological weapons. These safeguards could be reduced to the simple counting and monitoring of pebbles in storage. Alternatively, there is the option to create a special category with reduced requirements for this SNF in storage, transport, and disposal. No safeguards would be required for a repository with only this type of SNF. Reactor safeguards are required for fresh fuel, partly burnt fuel, and to identify unconventional pebbles with depleted uranium or other materials that might be used to create weapons-useable materials.

Constrained Tabular Diffusion for Finance

2026-03-08T03:31:54Z

Constrained Tabular Diffusion for Finance Cardei, Michael; Munoz, Jose; Barrera, Oscar; Chandrahas, Shreyas; Saha, Partha Generative models in finance face the dual challenge of producing realistic data while satisfying strict regulatory and economic objectives, a requirement that standard tabular diffusion models cannot provide. To address this difficulty, we introduce Constrained Tabular Diffusion for Finance (CTDF), a novel integration of sampling-time feasibility operations with mixed-type tabular diffusion in financial applications. By incorporating a training-free feasibility operator into the reverse‑diffusion sampling loop, CTDF enforces hard constraints for applications such as simulation, legal compliance, and extrapolation. Extensive experiments on large-scale financial datasets demonstrate zero constraint violations and improvement in scarce data utility. CTDF establishes a robust method for generating trustworthy and compliant synthetic data, opening new avenues for rigorous generative modeling and analysis in the financial domain. 6th ACM International Conference on AI in Finance (ICAIF ’25), November 15–18, 2025, Singapore, Singapore

A Bayesian sampling framework for constrained optimisation of build layouts in additive manufacturing

2026-03-08T03:38:48Z

A Bayesian sampling framework for constrained optimisation of build layouts in additive manufacturing Kim, Suh In; Gee, Kaitlyn; Hart, A John In additive manufacturing processes such as laser powder bed fusion, the build orientation and packing of components affect the required support structures, the number of parts in each build, and the surface roughness of the printed parts, among other factors. Maximising the packing density while minimising the build height can increase effective machine utilisation and decrease per-part cost. Yet, the build layout optimisation problem is highly nonlinear and difficult to solve using human intuition, so a systematic algorithm approach is required. Here, we present and demonstrate a voxel-based analysis method with Bayesian optimisation for determining component build orientation in additive manufacturing. We introduce selected case studies incorporating exemplary process attributes of laser powder bed fusion, including the determination of orientation and packing configurations based on support removal and tool-accessibility constraints.

18.05 Introduction to Probability and Statistics, Spring 2014

2025-12-16T19:16:26Z

18.05 Introduction to Probability and Statistics, Spring 2014 Orloff, Jeremy; Bloom, Jonathan This course provides an elementary introduction to probability and statistics with applications. Topics include: basic combinatorics, random variables, probability distributions, Bayesian inference, hypothesis testing, confidence intervals, and linear regression. The Spring 2014 version of this subject employed the residential MITx system, which enables on-campus subjects to provide MIT students with learning and assessment tools such as online problem sets, lecture videos, reading questions, pre-lecture questions, problem set assistance, tutorial videos, exam review content, and even online exams.

RES.STR-001 Geographic Information System (GIS) Tutorial, January IAP 2016

2025-12-16T05:02:38Z

RES.STR-001 Geographic Information System (GIS) Tutorial, January IAP 2016 MIT GIS Services Group The MIT GIS Services Group at the MIT Libraries hosts a number of tutorial workshops throughout the year. This resource gathers together some of those introductory workshop materials designed to accustom GIS novices to the various available software packages and introduce them to some of the many features included in GIS systems. Topics include an introduction to two GIS applications, spatial data analysis, and spatial statistics.

RES.18-001 Calculus Online Textbook, Spring 2005

2025-12-16T05:03:40Z

RES.18-001 Calculus Online Textbook, Spring 2005 Strang, Gilbert Published in 1991 by Wellesley-Cambridge Press, the book is a useful resource for educators and self-learners alike. It is well organized, covers single variable and multivariable calculus in depth, and is rich with applications. In addition to the Textbook, there is also an online Instructor's Manual and a student Study Guide. Prof. Strang has also developed a related series of videos, Highlights of Calculus, on the basic ideas of calculus.The 2010 second edition of the Calculus textbook includes a new chapter on "Highlights of Calculus" that connects to the video series of the same name. The new chapter has summaries and practice questions for all of the videos. It also introduces The Exponential Function (e^x) as presented in Prof. Strang's video on this topic.

CMS.930 / 21G.034 Media, Education, and the Marketplace, Fall 2001

2025-12-16T03:37:19Z

CMS.930 / 21G.034 Media, Education, and the Marketplace, Fall 2001 Miyagawa, Shigeru How can we harness the emerging forms of interactive media to enhance the learning process? Professor Miyagawa and prominent guest speakers will explore a broad range of issues on new media and learning - technical, social, and business. Concrete examples of use of media will be presented as case studies. One major theme, though not the only one, is that today's youth, influenced by video games and other emerging interactive media forms, are acquiring a fundamentally different attitude towards media. Media is, for them, not something to be consumed, but also to be created. This has broad consequences for how we design media, how the young are taught in schools, and how mass media markets will need to adjust.

21W.765J / 21L.489J / CMS.845J Interactive and Non-Linear Narrative: Theory and Practice, Spring 2006

2025-12-16T01:03:38Z

21W.765J / 21L.489J / CMS.845J Interactive and Non-Linear Narrative: Theory and Practice, Spring 2006 Coleman, Beth This course covers techniques of creating narratives that take advantage of the flexibility of form offered by the computer. The course studies the structural properties of book-based narratives that experiment with digression, multiple points of view, disruptions of time and of storyline. The class analyzes the structure and evaluates the literary qualities of computer-based narratives including hypertexts, adventure games, and classic artificial intelligence programs like Eliza. With this base, students use authoring systems to model a variety of narrative techniques and to create their own fictions. Knowledge of programming is helpful but not necessary.

CMS.610 / CMS.922 Media Industries and Systems, Spring 2006

2025-12-16T00:51:34Z

CMS.610 / CMS.922 Media Industries and Systems, Spring 2006 Weaver, Christopher This course examines the interplay of art, science, and commerce shaping the production, marketing, distribution, and consumption of contemporary media. It combines perspectives on media industries and systems with an awareness of the creative process, the audience, and trends shaping content. There will be invited discussions with industry experts in various subject areas. Class projects will encourage students to think through the challenges of producing media in an industry context. CMS.610 is for undergraduate credit, whereas CMS.922 is for graduate credit. Though the requirements for graduates are more stringent, the course is intended for both undergraduate and graduate students.

A rapid simple point-of-care assay for the detection of SARS-CoV-2 neutralizing antibodies

2026-03-08T03:34:07Z

A rapid simple point-of-care assay for the detection of SARS-CoV-2 neutralizing antibodies Kongsuphol, Patthara; Jia, Huan; Cheng, Hoi Lok; Gu, Yue; Shunmuganathan, Bhuvaneshwari DO; Chen, Ming Wei; Lim, Sing Mei; Ng, Say Yong; Tambyah, Paul Ananth; Nasir, Haziq; Gao, Xiaohong; Tay, Dousabel; Kim, Seunghyeon; Gupta, Rashi; Qian, Xinlei; Kozma, Mary M; Purushotorman, Kiren; McBee, Megan E; MacAry, Paul A; Sikes, Hadley D; Preiser, Peter R Background Neutralizing antibodies (NAbs) prevent pathogens from infecting host cells. Detection of SARS-CoV-2 NAbs is critical to evaluate herd immunity and monitor vaccine efficacy against SARS-CoV-2, the virus that causes COVID-19. All currently available NAb tests are lab-based and time-intensive. Method We develop a 10 min cellulose pull-down test to detect NAbs against SARS-CoV-2 from human plasma. The test evaluates the ability of antibodies to disrupt ACE2 receptor— RBD complex formation. The simple, portable, and rapid testing process relies on two key technologies: (i) the vertical-flow paper-based assay format and (ii) the rapid interaction of cellulose binding domain to cellulose paper. Results Here we show the construction of a cellulose-based vertical-flow test. The developed test gives above 80% sensitivity and specificity and up to 93% accuracy as compared to two current lab-based methods using COVID-19 convalescent plasma. Conclusions A rapid 10 min cellulose based test has been developed for detection of NAb against SARS-CoV-2. The test demonstrates comparable performance to the lab-based tests and can be used at Point-of-Care. Importantly, the approach used for this test can be easily extended to test RBD variants or to evaluate NAbs against other pathogens.

Developing a SARS-CoV-2 Antigen Test Using Engineered Affinity Proteins

2026-03-08T03:34:03Z

Developing a SARS-CoV-2 Antigen Test Using Engineered Affinity Proteins Kim, Seunghyeon; Yee, Emma; Miller, Eric A; Hao, Yining; Tay, Dousabel MY; Sung, Ki-Joo; Jia, Huan; Johnson, Joseph M; Saeed, Mohsan; Mace, Charles R; Yüksel Yurt, Deniz; Sikes, Hadley D The ongoing COVID-19 pandemic has clearly established how vital rapid, widely accessible diagnostic tests are in controlling infectious diseases and how difficult and slow it is to scale existing technologies. Here, we demonstrate the use of the rapid affinity pair identification via directed selection (RAPIDS) method to discover multiple affinity pairs for SARS-CoV-2 nucleocapsid protein (N-protein), a biomarker of COVID-19, from in vitro libraries in 10 weeks. The pair with the highest biomarker sensitivity was then integrated into a 10 min, vertical-flow cellulose paper test. Notably, the as-identified affinity proteins were compatible with a roll-to-roll printing process for large-scale manufacturing of tests. The test achieved 40 and 80 pM limits of detection in 1× phosphate-buffered saline (mock swab) and saliva matrices spiked with cell-culture-generated SARS-CoV-2 viruses and is also capable of detection of N-protein from characterized clinical swab samples. Hence, this work paves the way toward the mass production of cellulose paper-based assays which can address the shortages faced due to dependence on nitrocellulose and current manufacturing techniques. Further, the results reported herein indicate the promise of RAPIDS and engineered binder proteins for the timely and flexible development of clinically relevant diagnostic tests in response to emerging infectious diseases.

Archaeology of Self: Reflexivity in Data Activism to Address Systemic Injustices

2026-03-08T03:31:59Z

Archaeology of Self: Reflexivity in Data Activism to Address Systemic Injustices Walker, Raechel; Cruse, Brady; Cora, Aisha; Rogers, Kantwon; D'Ignazio, Catherine; Brion-Meisels, Gretchen; Breazeal, Cynthia Traditional data science education often neglects the importance of identity and sociopolitical context—especially for African American students whose lived experiences and cultural insights are essential for building justice centered technologies. This paper presents findings from the Data Activism Program, which integrated Dr. Yolanda Sealey-Ruiz’s Archaeology of Self™ framework to foster critical self-reflection and racial identity development among African American high school and college students. Through technical training in data science, art-based learning, and partnerships with social justice organizations, students engaged in reflexive practices that positioned them as active agents in challenging systemic oppression. Interviews reveal that the Archaeology of Self™ deepened students’ reflexivity skills and strengthened their sound racial identity, enabling them to interrogate bias within themselves and the data science process. We argue that embedding frameworks such as the Archaeology of Self™ into algorithmic design offers a concrete, transferable method for operationalizing reflexivity in data science and AI. This study contributes to the AI and data science community by offering actionable strategies to center identity and power in AI development. EAAMO ’25, Pittsburgh, PA, USA

Advances in Financial AI: Innovations, Risk, and Responsibility in the Era of LLMs

2026-03-08T03:31:53Z

Advances in Financial AI: Innovations, Risk, and Responsibility in the Era of LLMs Lee, Yongjae; Mehrasa, Nazanin; Choi, Chanyeol; Chen, Chung-Chi; Mehta, Dhagash; Zohren, Stefan; Kim, Yoon; Lee, Chulheum; Lee, Yeonhee; Oh, Eunsook The finance sector is seeing a rapid increase in the application of machine learning and AI, with Large Language Models (LLMs), ESG (Environmental, Social, and Governance) investing, and AI Safety significantly reshaping the field. This workshop focuses on how these advancements intersect with core financial AI applications. We will foster interdisciplinary discussion on applying LLMs to finance, addressing challenges in multilingual and non-English markets like Korea. The event will also highlight the integration of ESG signals into algorithmic decision-making and explore AI Safety, emphasizing reliability, fairness, and explainability for AI systems in regulated financial environments. By bringing together experts from academia, industry, and regulatory bodies, the workshop aims to stimulate discussions on practical issues, ethical dilemmas, and cutting-edge research shaping financial AI's future. We welcome submissions that combine technical rigor with societal relevance in AI-driven financial decisions. CIKM ’25, Seoul, Republic of Korea

Design and Fabrication of Hybrid Functional Identities for Mechanical Elements

2026-03-08T03:22:13Z

Design and Fabrication of Hybrid Functional Identities for Mechanical Elements AlAlawi, Marwa My PhD research explores the simultaneous integration of mechanical and electrical functionalities in mechanical components such as gears, linkages, and springs, which I define as "hybrid functional identities." The focus is on transforming these components into non-intrusive sensors and active elements that maintain structural integrity while providing electrical capabilities like sensing, energy harvesting, and communication. I establish a framework for hybrid functional identities by examining common mechanical elements and their associated motions—rotational, linear, and reciprocal—along with force-based interactions like stretching, compression, and torsion. This analysis identifies essential electrical functionalities that complement these mechanical behaviors. Building on this foundation, I investigate modular mechanical building blocks that support diverse mechanical and electrical interaction primitives using a unified geometric structure. Ultimately, I aim to create an interconnected system where hybrid mechanical-electrical components function autonomously and communicate through an embedded wireless network. CHI EA ’25, Yokohama, Japan

Connecting through Comics: Design and Evaluation of Cube, an Arts-Based Digital Platform for Trauma-Impacted Youth

2026-03-08T03:22:13Z

Connecting through Comics: Design and Evaluation of Cube, an Arts-Based Digital Platform for Trauma-Impacted Youth Kumar, Ila; Shen, Jocelyn; Ferguson, Craig; Picard, Rosalind This paper explores the design, development and evaluation of a digital platform that aims to assist young people who have experienced trauma in understanding and expressing their emotions and fostering social connections. Integrating principles from expressive arts and narrative-based therapies, we collaborate with lived experts to iteratively design a novel, user-centered digital tool for young people to create and share comics that represent their experiences. Specifically, we conduct a series of nine workshops with N=54 trauma-impacted youth and young adults to test and refine our tool, beginning with three workshops using low-fidelity prototypes, followed by six workshops with Cube, a web version of the tool. A qualitative analysis of workshop feedback and empathic relations analysis of artifacts provides valuable insights into the usability and potential impact of the tool, as well as the specific needs of young people who have experienced trauma. Our findings suggest that the integration of expressive and narrative therapy principles into Cube can offer a unique avenue for trauma-impacted young people to process their experiences, more easily communicate their emotions, and connect with supportive communities. We end by presenting implications for the design of social technologies that aim to support the emotional well-being and social integration of youth and young adults who have faced trauma.

The Structure of Cross-National Collaboration in Open-Source Software Development

2026-03-08T03:31:58Z

The Structure of Cross-National Collaboration in Open-Source Software Development Xu, Henry; Yu, Katy; He, Hao; Fang, Hongbo; Vasilescu, Bogdan; Park, Patrick Open-source software (OSS) development platforms, such as GitHub, expand the potential for cross-national collaboration among developers by lowering the geographic, temporal, and coordination barriers that limited software innovation in the past. However, research has shown that the technological affordances that facilitate cross-national collaboration do not uniformly benefit all countries. Using the GitHub Innovation Graph dataset, which aggregates the complete cross-country collaborations among the entire population of GitHub developers, we present quantitative evidence of deep-seated religious and cultural affinities, shared colonial histories, and geopolitical factors structuring the collaborations between non-U.S. country pairs that become visible when the overarching dominance of the U.S. is removed from the data. This study highlights the opportunities to develop decentralizing strategies to facilitate new collaborations between developers in non-U.S. countries, thereby fostering the development of novel, innovative solutions. More generally, this study also underscores the importance of contextualizing user behavior and knowledge management in information systems with long-term, macro-social conditions in which these systems are inextricably embedded. CIKM ’25, Seoul, Republic of Korea

Partition–diffusion–reaction bounds for thin-film membrane formation kinetics

2026-01-05T17:12:17Z

Partition–diffusion–reaction bounds for thin-film membrane formation kinetics Deshmukh, Akshay; Elimelech, Menachem; Lienhard, John H. New membrane chemistries and structures have rapidly developed over the last ten years, driven by applications ranging from critical metals separations and carbon capture to highly chlorine-resistant reverse-osmosis membranes. The thin selective layer at the heart of reverse osmosis and nanofiltration membranes is typically fabricated using interfacial synthesis, with multifunctional aqueous-phase monomers and organicphase monomers. Here, we develop a physics-based model of partition, diffusion, and reaction dynamics during the early stages of interfacial synthesis. These processes critically impact membrane structure and performance. By solving the resulting partial differential equations numerically and with analytical approximations, we demonstrate that the planar reaction rate is initially limited by the partitioning and diffusion of the aqueous-phase reactant into the organic phase. Later, finite reactant availability and aqueous-phase diffusion become limiting. Through a combination of nondimensionalization, parameter mapping, and property prediction, we develop a framework that spans a wide parameter space in reactant chemistry, solvent and support layer choice, and initial reactant concentrations. We demonstrate that the planar reaction rate and dynamics are strongly affected by the partition coefficient of the aqueous reactant, which varies rapidly with changes in reactant and solvent chemistry. The influence of diffusion variations is more limited. This tractable, physics-based model enables the rapid quantification of monomer and solvent impact on interfacial synthesis, which is essential for the rational development of new high-performance thin-film composite membranes.

Finger stick blood test to assess postvaccination SARS-CoV-2 neutralizing antibody response against variants

2025-12-13T03:10:37Z

Finger stick blood test to assess postvaccination SARS-CoV-2 neutralizing antibody response against variants Lim, Sing Mei; Cheng, Hoi Lok; Jia, Huan; Kongsuphol, Patthara; D/O Shunmuganathan, Bhuvaneshwari; Chen, Ming Wei; Ng, Say Yong; Gao, Xiaohong; Turaga, Shuvan Prashant; Heussler, Sascha P; Somani, Jyoti; Sengupta, Sharmila; Tay, Dousabel MY; McBee, Megan E; Young, Barnaby E; MacAry, Paul A; Sikes, Hadley D; Preiser, Peter R There is clinical need for a quantifiable point-of-care (PoC) SARS-CoV-2 neutralizing antibody (nAb) test that is adaptable with the pandemic's changing landscape. Here, we present a rapid and semi-quantitative nAb test that uses finger stick or venous blood to assess the nAb response of vaccinated population against wild-type (WT), alpha, beta, gamma, and delta variant RBDs. It captures a clinically relevant range of nAb levels, and effectively differentiates prevaccination, post first dose, and post second dose vaccination samples within 10 min. The data observed against alpha, beta, gamma, and delta variants agrees with published results evaluated in established serology tests. Finally, our test revealed a substantial reduction in nAb level for beta, gamma, and delta variants between early BNT162b2 vaccination group (within 3 months) and later vaccination group (post 3 months). This test is highly suited for PoC settings and provides an insightful nAb response in a postvaccinated population.

Rapid Evaluation of Vaccine Booster Effectiveness against SARS-CoV-2 Variants

2025-12-13T03:10:36Z

Rapid Evaluation of Vaccine Booster Effectiveness against SARS-CoV-2 Variants Cheng, Hoi Lok; Lim, Sing Mei; Jia, Huan; Chen, Ming Wei; Ng, Say Yong; Gao, Xiaohong; Somani, Jyoti; Sengupta, Sharmila; Tay, Dousabel MY; Chua, Patrina WL; R., Abirami; Ling, Sharon YH; McBee, Megan E; Young, Barnaby E; Sikes, Hadley D; Preiser, Peter R As the COVID-19 pandemic continues, countries around the world are switching toward vaccinations and boosters to combat the pandemic. However, waning immunity against SARS-CoV-2 wild-type (WT) and variants have been widely reported. Booster vaccinations have shown to be able to increase immunological protection against new variants; however, the protection observed appears to decrease quickly over time suggesting a second booster shot may be appropriate. Moreover, heterogeneity and waning of the immune response at the individual level was observed suggesting a more personalized vaccination approach should be considered. To evaluate such a personalized strategy, it is important to have the ability to rapidly evaluate the level of neutralizing antibody (nAbs) response against variants at the individual level and ideally at a point of care setting. Here, we applied the recently developed cellulose pulled-down virus neutralization test (cpVNT) to rapidly assess individual nAb levels to WT and variants of concerns in response to booster vaccination. Our findings confirmed significant heterogeneity of nAb responses against a panel of SARS-CoV-2 variants, and indicated a strong increase in nAb response against variants of concern (VOCs) upon booster vaccination. For instance, the nAb response against current predominant omicron variant was observed with medians of 88.1% (n = 6, 95% CI = 73.2% to 96.2%) within 1-month postbooster and 70.7% (n = 22, 95% CI = 66.4% to 81.8%) 3 months postbooster. Our data show a point of care (POC) test focusing on nAb response levels against VOCs can guide decisions on the potential need for booster vaccinations at individual level. Importantly, it also suggests the current booster vaccines only give a transient protective response against some VOC and new more targeted formulations of a booster vaccine against specific VOC may need to be developed in the future. IMPORTANCE Vaccination against SARS-CoV-2 induces protection through production of neutralization antibodies (nAb). The level of nAb is a major indicator of immunity against SARS-CoV-2 infection. We developed a rapid point-of-care test that can monitor the nAb level from a drop of finger stick blood. Here, we have implemented the test to monitor individual nAb level against wild-type and variants of SARS-CoV-2 at various time points of vaccination, including post-second-dose vaccination and postbooster vaccination. Huge diversity of nAb levels were observed among individuals as well as increment in nAb levels especially against Omicron variant after booster vaccination. This study evaluated the performance of this point-of-care test for personalized nAb response tracking. It verifies the potential of using a rapid nAb test to guide future vaccination regimens at both the individual and population level.

Tumor-localized catalases can fail to alter tumor growth and transcriptional profiles in subcutaneous syngeneic mouse tumor models

2025-12-13T03:10:51Z

Tumor-localized catalases can fail to alter tumor growth and transcriptional profiles in subcutaneous syngeneic mouse tumor models Sheen, Allison; Agarwal, Yash; Cheah, Keith M; Cowles, Sarah C; Stinson, Jordan A; Palmeri, Joseph R; Sikes, Hadley D; Wittrup, K Dane Catalase is an antioxidant enzyme that catalyzes the rapid conversion of hydrogen peroxide to water and oxygen. Use of catalase as a cancer therapeutic has been proposed to reduce oxidative stress and hypoxia in the tumor microenvironment, both activities which are hypothesized to reduce tumor growth. Furthermore, exposing murine tumors to exogenous catalase was previously reported to have therapeutic benefit. We studied the therapeutic effect of tumor-localized catalases with the aim to further elucidate the mechanism of action. To do this, we engineered two approaches to maximize intratumoral catalase exposure: 1) an injected extracellular catalase with enhanced tumor retention, and 2) tumor cell lines that over-express intracellular catalase. Both approaches were characterized for functionality and tested for therapeutic efficacy and mechanism in 4T1 and CT26 murine syngeneic tumor models. The injected catalase was confirmed to have enzyme activity >30,000 U/mg and was retained at the injection site for more than one week in vivo. The engineered cell lines exhibited increased catalase activity and antioxidant capacity, with catalase over-expression that was maintained for at least one week after gene expression was induced in vivo. We did not observe a significant difference in tumor growth or survival between catalase-treated and untreated mice when either approach was used. Finally, bulk RNA sequencing of tumors was performed, comparing the gene expression of catalase-treated and untreated tumors. Gene expression analysis revealed very few differentially expressed genes as a result of exposure to catalase and notably, we did not observe changes consistent with an altered state of hypoxia or oxidative stress. In conclusion, we observe that sustained intratumoral catalase neither has therapeutic benefit nor triggers significant differential expression of genes associated with the anticipated therapeutic mechanism in the subcutaneous syngeneic tumor models used. Given the lack of effect observed, we propose that further development of catalase as a cancer therapeutic should take these findings into consideration.

Optical Detection of Interleukin-6 Using Liquid Janus Emulsions Using Hyperthermophilic Affinity Proteins

2025-12-13T03:10:31Z

Optical Detection of Interleukin-6 Using Liquid Janus Emulsions Using Hyperthermophilic Affinity Proteins Chen, Michelle; Corless, Elliot I; Engelward, Bevin P; Swager, Timothy M; Sikes, Hadley D. When equal volumes of two immiscible liquids are mixed (e.g., a hydrocarbon and a fluorocarbon), Janus droplets can form in an aqueous solution. In a gravity-aligned Janus droplet, the boundary between the two phases is flat and thus optically transparent when viewed from above. When tipped due to interactions with an analyte (i.e., agglutination), the resulting change in refraction and reflection yields an optical signal that can be detected and quantified. This study reports the detection and quantitation of interleukin-6 (IL-6) using emulsions functionalized at the hydrocarbon:aqueous interface with engineered proteins that bind IL-6 at high affinity and specificity. Hyperthermophilic affinity proteins (rcSso7d) are derived from thermophiles, giving them excellent thermal stability. Two rcSso7d affinity protein variants were synthesized with a noncanonical azide-functionalized amino acid to enable click chemistry to novel polymeric anchors embedded in the hydrocarbon phase. The two binding proteins recognize different epitopes, enabling the detection of both monomeric and dimeric IL-6 via agglutination. It is noteworthy that the rsSso7d protein variants, in addition to having superior thermal stability and facile recombinant synthesis in E. coli, show superior performance when compared to commercial antibodies for IL-6.

Point-of-need diagnostics in a post-Covid world: an opportunity for paper-based microfluidics to serve during syndemics

2025-12-13T03:10:44Z

Point-of-need diagnostics in a post-Covid world: an opportunity for paper-based microfluidics to serve during syndemics Tsaloglou, Maria-Nefeli; Christodouleas, Dionysios C; Milette, Jonathan; Milkey, Kendall; Romine, Isabelle C; Im, Judy; Lathwal, Shefali; Selvam, Duraipandian Thava; Sikes, Hadley D; Whitesides, George M Zoonotic outbreaks present with unpredictable threats to human health, food production, biodiversity, national security and disrupt the global economy. The COVID-19 pandemic—caused by zoonotic coronavirus, SARS-CoV2— is the most recent upsurge of an increasing trend in outbreaks for the past 100 years. This year, emergence of avian influenza (H5N1) is a stark reminder of the need for national and international pandemic preparedness. Tools for threat reduction include consistent practices in reporting pandemics, and widespread availability of accurate detection technologies. Wars and extreme climate events redouble the need for fast, adaptable and affordable diagnostics at the point of need. During the recent pandemic, rapid home tests for SARS-CoV-2 proved to be a viable functional model that leverages simplicity. In this perspective, we introduce the concept of syndemnicity in the context of infectious diseases and point-of-need healthcare diagnostics. We also provide a brief state-of-the-art for paper-based microfluidics. We illustrate our arguments with a case study for detecting brucellosis in cows. Finally, we conclude with lessons learned, challenges and opportunities for paper-based microfluidics to serve point-of-need healthcare diagnostics during syndemics.

Faster search for tensor decomposition over finite fields

2025-12-13T03:10:04Z

Faster search for tensor decomposition over finite fields Yang, Jason We present an 𝑂 ∗ (|F| min{𝑅, Í 𝑑≥2 𝑛𝑑 }+(𝑅−𝑛0 ) (Í 𝑑≠0 𝑛𝑑 ) )-time algorithm for determining whether the rank of a concise tensor 𝑇 ∈ F 𝑛0×···×𝑛𝐷−1 is ≤ 𝑅, assuming 𝑛0 ≥ · · · ≥ 𝑛𝐷−1 and 𝑅 ≥ 𝑛0. For 3-dimensional tensors, we have a second algorithm running in 𝑂 ∗ (|F| 𝑛0+𝑛2+(𝑅−𝑛0+1−𝑟∗ ) (𝑛1+𝑛2 )+𝑟 2 ∗ ) time, where 𝑟∗ := j 𝑅 𝑛0 k + 1. Both algorithms use polynomial space and improve on our previous work, which achieved running time 𝑂 ∗ (|F| 𝑛0+(𝑅−𝑛0 ) (Í 𝑑 𝑛𝑑 ) ). ISSAC ’25, Guanajuato, Mexico

Human-AI Interaction for Augmented Reasoning: Improving Human Reflective and Critical Thinking with Artificial Intelligence

2025-12-13T03:09:46Z

Human-AI Interaction for Augmented Reasoning: Improving Human Reflective and Critical Thinking with Artificial Intelligence Danry, Valdemar; Pataranutaporn, Pat; Cui, Christopher; Hung, Jui-Tse; Blanchard, Lancelot; Bu?inca, Zana; Tan, Chenhao; Starner, Thad; Maes, Pattie AI-Augmented Reasoning systems are cognitive assistants that support human reasoning by providing AI-based feedback that can help users improve their critical reasoning skills. Made possible with new techniques like argumentation mining, fact-checking, crowdsourcing, attention nudging, and large language models, AI augmented reasoning systems can provide real-time feedback on logical reasoning, help users identify and avoid flawed arguments and misinformation, suggest counter-arguments, provide evidence-based explanations, and foster deeper reflection. The goal of this workshop is to bring together researchers from AI, HCI, cognitive and social science to discuss recent advances in AI-augmented reasoning, to identify open problems in this area, and to cultivate an emerging community on this important topic. CHI EA ’25, Yokohama, Japan

Speech to Reality: On-Demand Production using Natural Language, 3D Generative AI, and Discrete Robotic Assembly

2025-12-13T03:10:08Z

Speech to Reality: On-Demand Production using Natural Language, 3D Generative AI, and Discrete Robotic Assembly Kyaw, Alexander Htet; Smith, Miana; Jeon, Se Hwan; Gershenfeld, Neil We present a system that transforms speech into physical objects using 3D generative AI and discrete robotic assembly. By leveraging natural language, the system makes design and manufacturing more accessible to people without expertise in 3D modeling or robotic programming. While generative AI models can produce a wide range of 3D meshes, AI-generated meshes are not directly suitable for robotic assembly or account for fabrication constraints. To address this, we contribute a workflow that integrates natural language, 3D generative AI, geometric processing, and discrete robotic assembly. The system discretizes the AI-generated geometry and modifies it to meet fabrication constraints such as component count, overhangs, and connectivity to ensure feasible physical assembly. The results are demonstrated through the assembly of various objects, ranging from chairs to shelves, which are prompted via speech and realized within 5 minutes using a robotic arm. SCF ’25, Cambridge, MA, USA

MechStyle: Augmenting Generative AI with Mechanical Simulation to Create Stylized and Structurally Viable 3D Models

2025-12-13T03:10:14Z

MechStyle: Augmenting Generative AI with Mechanical Simulation to Create Stylized and Structurally Viable 3D Models Faruqi, Faraz; Abdel-Rahman, Amira; Tejedor, Leandra; Nisser, Martin; Li, Jiaji; Phadnis, Vrushank; Jampani, Varun; Gershenfeld, Neil; Hofmann, Megan; Mueller, Stefanie Recent developments in Generative AI enable creators to stylize 3D models based on text prompts. These methods change the 3D model geometry, which can compromise the model’s structural integrity once fabricated. We present MechStyle, a system that enables creators to stylize 3D printable models while preserving their structural integrity. MechStyle accomplishes this by augmenting the Generative AI-based stylization process with feedback from a Finite Element Analysis (FEA) simulation. As the stylization process modifies the geometry to approximate the desired style, feedback from the FEA simulation reduces modifications to regions with increased stress. We evaluate the effectiveness of FEA simulation feedback in the augmented stylization process by comparing three stylization control strategies. We also investigate the time efficiency of our approach by comparing three adaptive scheduling strategies. Finally, we demonstrate MechStyle’s user interface that allows users to generate stylized and structurally viable 3D models and provide five example applications. SCF ’25, Cambridge, MA, USA

Technology-Policy Handbook for Trans-Atlantic Nuclear Maritime Corridors: Ports, Infrastructure, and Safety

2026-01-21T15:08:23Z

Technology-Policy Handbook for Trans-Atlantic Nuclear Maritime Corridors: Ports, Infrastructure, and Safety Valiaveedu, Anthony; Edmonds, Nat On September 18, 2025, the United States and the United Kingdom published a Memorandum of Understanding (MoU) on scientific and technological advancement. This new partnership focuses on understanding and deploying disruptive technologies in Artificial Intelligence, quantum, and civil nuclear energy. Less highlighted was a single sentence within the MoU outlining efforts to "explore opportunities" for establishing a "maritime shipping corridor" between the US and UK. So far, research on civilian nuclear ships has generally prioritized ship design and operation analysis. This paper will instead analyze port, regulatory, and infrastructural issues within this space and provide a path forward for technology policy solutions supporting systems safety. Advised by Jacopo Buongiorno, Eric Forrest, Fotini Christia, Koroush Shirvanm and Themis Sapsis.; Contact information: Anthony Valiaveedu (arv7@mit.edu); Nat Edmonds (edmondsn@mit.edu)

AI-assisted sensemaking: Human-AI collaboration for the analysis and interpretation of recorded facilitated conversations

2025-12-13T03:09:44Z

AI-assisted sensemaking: Human-AI collaboration for the analysis and interpretation of recorded facilitated conversations Kabbara, Jad; Phan, Thanh-Mai; Rakhilin, Marina; Detwiller, Maya; Dimitrakopoulou, Dimitra; Roy, Deb In light of growing toxic polarization and societal fragmentation often fueled by social media, we are designing alternative communication spaces we refer to as dialogue networks—networks of people engaged in recorded small-group prompted dialogue. We introduce the dialogue network framework and our use of tools powered by large language models that assist humans in the analysis and interpretation of themes and patterns across conversations which we refer to as sensemaking. We pilot case studies in collaboration with community partners using a prototype AI-assisted sensemaking tool. Insights from these pilots can inform the use of AI for human-led community engagement processes. CHI EA ’25, Yokohama, Japan

$HealthGenie:$ A Knowledge-Driven LLM Framework for Tailored Dietary Guidance

2025-12-13T03:10:10Z

$HealthGenie:$ A Knowledge-Driven LLM Framework for Tailored Dietary Guidance Gao, Fan; Zhao, Xinjie; Xia, Ding; Zhou, Zhongyi; Yang, Rui; Lu, Jinghui; Jiang, Hang; Park, Chanjun; Li, Irene Seeking dietary guidance often requires navigating complex nutritional knowledge while considering individual health needs. To address this, we present HealthGenie, an interactive platform that leverages the interpretability of knowledge graphs (KGs) and the conversational power of large language models (LLMs) to deliver tailored dietary recommendations alongside integrated nutritional visualizations for fast, intuitive insights. Upon receiving a user query, HealthGenie performs intent refinement and maps user's needs to a curated nutritional knowledge graph. The system then retrieves and visualizes relevant subgraphs, while offering detailed, explainable recommendations. Users can interactively adjust preferences to further tailor results. A within-subject study and quantitative analysis show that HealthGenie reduces cognitive load and interaction effort while supporting personalized, health-aware decision-making. CIKM ’25, Seoul, Republic of Korea

TinkerXR: In-Situ, Reality-Aware CAD and 3D Printing Interface for Novices

2025-12-13T03:10:17Z

TinkerXR: In-Situ, Reality-Aware CAD and 3D Printing Interface for Novices Arslan, O?uz; Akdo?an, Artun; Dogan, Mustafa Doga Despite the growing accessibility of augmented reality (AR) for visualization, existing computer-aided design (CAD) systems remain confined to traditional screens or require complex setups or predefined parameters, limiting immersion and accessibility for novices. We present TinkerXR, an open-source AR interface enabling in-situ design and fabrication through Constructive Solid Geometry (CSG) modeling. TinkerXR operates solely with a headset and 3D printer, allowing users to design directly in and for their physical environments. By leveraging spatial awareness, depth occlusion, recognition of physical constraints, reference objects, and hand movement controls, TinkerXR enhances realism, precision, and ease of use. Its AR-based workflow integrates design and 3D printing with a drag-and-drop interface for printers’ virtual twins. A user study comparing TinkerXR with Tinkercad shows that TinkerXR offers novices higher accessibility, engagement, and ease of use. Participants highlighted how designing directly in physical space made the process more intuitive. By bridging the gap between digital creation and physical output, TinkerXR aims to transform everyday spaces into expressive creative studios. We release TinkerXR as open source1 to encourage further exploration of accessible, spatially grounded CAD tools. SCF ’25, Cambridge, MA, USA

Hierarchical Discrete Lattice Assembly: An Approach for the Digital Fabrication of Scalable Macroscale Structures

2025-12-13T03:10:19Z

Hierarchical Discrete Lattice Assembly: An Approach for the Digital Fabrication of Scalable Macroscale Structures Smith, Miana; Richard, Paul; Kyaw, Alexander; Gershenfeld, Neil Although digital fabrication processes at the desktop scale have become proficient and prolific, systems aimed at producing larger-scale structures are still typically complex, expensive, and unreliable. In this work, we present an approach for the fabrication of scalable macroscale structures using simple robots and interlocking lattice building blocks. A target structure is first voxelized so that it can be populated with an architected lattice. These voxels are then grouped into larger interconnected blocks, which are produced using standard digital fabrication processes, leveraging their capability to produce highly complex geometries at a small scale. These blocks, on the size scale of tens of centimeters, are then fed to mobile relative robots that are able to traverse over the structure and place new blocks to form structures on the meter scale. To facilitate the assembly of large structures, we introduce a live digital twin simulation tool for controlling and coordinating assembly robots that enables both global planning for a target structure and live user design, interaction, or intervention. To improve assembly throughput, we introduce a new modular assembly robot, designed for hierarchical voxel handling. We validate this system by demonstrating the voxelization, hierarchical blocking, path planning, and robotic fabrication of a set of meter-scale objects. SCF ’25, Cambridge, MA, USA

Are Crypto Ecosystems (De)centralizing? A Framework for Longitudinal Analysis

2025-12-13T03:10:12Z

Are Crypto Ecosystems (De)centralizing? A Framework for Longitudinal Analysis Ju, Harang; Valavi, Eshan; Kumar, Madhav; Aral, Sinan Blockchain technology relies on decentralization to resist faults and attacks while operating without trusted intermediaries. Although industry experts have touted decentralization as central to their promise and disruptive potential, it is still unclear whether the crypto ecosystems built around blockchains are becoming more or less decentralized over time. As crypto plays an increasing role in facilitating economic transactions and peer-to-peer interactions, measuring their decentralization becomes even more essential.We thus propose a systematic framework for measuring the decentralization of crypto ecosystems over time and compare commonly used decentralization metrics. We applied this framework to seven prominent blockchains, across five distinct subsystems and across their lifetime for over 15 years. Our analysis revealed that while crypto has largely become more decentralized over time, recent trends show a shift toward centralization in the consensus layer, NFT marketplaces, and developers. Our framework and results inform researchers, policymakers, and practitioners about the design, regulation, and implementation of crypto ecosystems and provide a systematic, replicable foundation for future studies.

MIT hosts 11th Undergraduate Research Technology Conference

2025-12-13T03:12:15Z

MIT hosts 11th Undergraduate Research Technology Conference Beyah, Malakhi; Placides, Jojo From Oct. 10 to Oct. 12, the Stata Center was abuzz with bright minds and fresh faces as the Institute geared up for its 11th annual Undergraduate Research Technology Conference (URTC), where high school and undergraduate students from across the country came to present their latest research to experts and industry leaders.

Multimodal AI for Human Sensing and Interaction

2025-12-13T03:09:38Z

Multimodal AI for Human Sensing and Interaction Liang, Paul Pu; Ahuja, Karan; Luo, Yiyue A significant body of HCI research today focuses on applying AI to sense, learn, and interact with humans through a wide range of wearable and ubiquitous sensors. These methods typically involve learning features from multimodal sensory data using AI methods. To aid HCI researchers who want to apply AI to their sensing problems, this course will cover the fundamental challenges and approaches in multimodal AI for human sensing and interaction. It is planned for 3 parts, one given by each organizer. The first covers the foundations of multimodal AI, studying how AI systems can represent, combine, and learn information from many interconnected sensory inputs. The second part discusses the practice of multimodal AI for human sensing, covering the latest methods for cross-modal learning across diverse sensors, human-centered application domains, and real-world concerns around their usage. The final part covers the hardware, fabrication, and data collection challenges that must be tackled to deploy these multimodal AI systems in the real world. By the end of this course, attendees should understand the fundamental principles and challenges of multimodal AI, identify the right AI approaches for their problems, prototype basic hardware systems for efficient and robust sensing, be aware of real-world concerns around ethics, interpretability, and privacy, and appreciate the range of human-centered applications enabled by multimodal AI and sensing. CHI EA ’25, Yokohama, Japan

Deep-Time Architecture: Building as Material-Event

2025-12-13T03:10:47Z

Deep-Time Architecture: Building as Material-Event Alonso, Cristina Parreño Despite our tendency to conceive, perceive, and represent buildings as static objects, buildings are, in their abundant reality, matter and energy in flux. As Heraclitus famously remarked in his panta rhei (πάντα ῥεῖ,): “everything flows.”1 Buildings are no different, and they need to be better thought through as entities in motion. In architectural literature, many voices have challenged the prevailing notion of the building as a static object. Bruno Latour, for instance, claims that a building is rather “a moving project, and that even once it has been built, it ages, it is transformed by its users, modified by all of what happens inside and outside, and that it will pass or be renovated, adulterated and transformed beyond recognition.”2 Another attempt to express architecture’s fluidity is Bernard Tschumi’s triad, “space, event and movement,” with which he aimed to expand what constitutes building beyond a static object and form: “There is no space without event, no architecture without movement.”3 And here we must add that there is no movement without time—and further, that given enough time, even a solid-like material (think of a building here) flows.

Involuntary vs. voluntary flexible work: insights for scholars and stakeholders

2025-12-13T03:10:26Z

Involuntary vs. voluntary flexible work: insights for scholars and stakeholders Kaduk, Anne; Genadek, Katie; Kelly, Erin L; Moen, Phyllis Building on insights from the early stages of our research partnership with a U.S. Fortune 500 organization, we came to differentiate between voluntary and involuntary schedule variability and remote work. This differentiation underscores the complexity behind flexible schedules and remote work, especially among white-collar, salaried professionals. We collected survey data among the partner firm's information technology (IT) workforce to evaluate whether these forms of flexibility had different implications for workers, as part of the larger Work, Family, and Health Network Study. We find that a significant minority of these employees report working variable schedules and working at home involuntarily. Involuntary variable schedules are associated with greater work-to-family conflict, stress, burnout, turnover intentions, and lower job satisfaction in models that adjust for personal characteristics, job, work hours, family demands, and other factors. Voluntary remote work, in contrast, is protective and more common in this professional sample. Employees working at least 20% of their hours at home and reporting moderate or high choice over where they work have lower stress and intentions to leave the firm. These findings point to the importance of both stakeholders and scholars distinguishing between voluntary and involuntary forms of flexibility, even in a relatively advantaged workforce.

Machine learning demand forecasting and supply chain performance

2025-12-13T03:10:33Z

Machine learning demand forecasting and supply chain performance Feizabadi, Javad In many supply chains, firms staged in upstream of the chain suffer from variance amplification emanating from demand information distortion in a multi-stage supply chain and, consequently, their operation inefficiency. Prior research suggest that employing advanced demand forecasting, such as machine learning, could mitigate the effect and improve the performance; however, it is less known what is the extent and magnitude of savings as tangible supply chain performance outcomes. In this research, hybrid demand forecasting methods grounded on machine learning i.e. ARIMAX and Neural Network is developed. Both time series and explanatory factors are feed into the developed method. The method was applied and evaluated in the context of functional product and a steel manufacturer. The statistically significant supply chain performance improvement differences were found across traditional and ML-based demand forecasting methods. The implications for the theory and practice are also presented.

The future of sperm: a biovariability framework for understanding global sperm count trends

2025-12-13T03:10:46Z

The future of sperm: a biovariability framework for understanding global sperm count trends Boulicault, Marion; Perret, Meg; Galka, Jonathan; Borsa, Alex; Gompers, Annika; Reiches, Meredith; Richardson, Sarah The past 50 years have seen heated debate in the reproductive sciences about global trends in human sperm count. In 2017, Levine and colleagues published the largest and most methodologically rigorous meta-regression analysis to date and reported that average total sperm concentration among men from ‘Western’ countries has decreased by 59.3% since 1973, with no sign of halting. These results reverberated in the scientific community and in public discussions about men and masculinity in the modern world, in part because of scientists’ public-facing claims about the societal implications of the decline of male fertility. We find that existing research follows a set of implicit and explicit assumptions about how to measure and interpret sperm counts, which collectively form what we term the Sperm Count Decline hypothesis (SCD). Using the study by Levine and colleagues, we identify weaknesses and inconsistencies in the SCD, and propose an alternative framework to guide research on sperm count trends: the Sperm Count Biovariability hypothesis (SCB). SCB asserts that sperm count varies within a wide range, much of which can be considered non-pathological and species-typical. Knowledge about the relationship between individual and population sperm count and life-historical and ecological factors is critical to interpreting trends in average sperm counts and their relationships to health and fertility.

Estimating Pedestrian Flows on Street Networks

2025-12-13T03:10:41Z

Estimating Pedestrian Flows on Street Networks Sevtsuk, Andres City governments and planners alike commonly seek to increase pedestrian activity on city streets as part of broader sustainability, community building, and economic development strategies. Though walkability has received ample attention in planning literature, most planners still lack practical methods for predicting how development proposals could affect pedestrian activity on specific streets or public spaces at different times of the day. Cities typically require traffic impact assessments (TIAs) but not pedestrian impact assessments. In this study I present a methodology for estimating pedestrian trip generation and distribution between detailed origins and destinations in both existing and proposed built environments. Using the betweenness index from network analysis, I introduce a number of methodological improvements that allow the index to model pedestrian trips with parameters and constraints to account for pedestrian behavior in different settings. I demonstrate its application in the Kendall Square area of Cambridge (MA), where estimated foot traffic is compared during lunch and evening peak periods with observed pedestrian counts. The proposed approach can be particularly useful for TIAs, neighborhood plans, and large-scale development projects, where pedestrian flow estimates can be used to guide pedestrian infrastructure and safety improvements and public space investments or for locating pedestrian priority streets during the COVID-19 pandemic.

Understanding individuals with spinal cord injury’s self-care practices: a technology probe study to promote pressure relief adherence

2025-12-13T03:10:23Z

Understanding individuals with spinal cord injury’s self-care practices: a technology probe study to promote pressure relief adherence Oh, Hannah Hye Yeon; Pontis, Sheila Pressure reliefs (PRs) are self-care practices essential for individuals with spinal cord injury (SCI) to prevent life-threatening pressure injuries (PIs). Despite the benefits, individuals often do not do these exercises at home, leading to increased patient morbidity and mortality. To examine how digital technology could improve this population's adherence to PR exercises, we conducted a technology probe study with five individuals with SCI over ten consecutive business days. A chat-based intervention was created to send user-scheduled PR reminders, which were personalized with visual elements and progress trackers. Participants were interviewed before and after interacting with the probe to better understand their experiences with PIs and PR practices. Results shed light on specific factors that may impact individuals with SCI's behaviours towards PRs and four considerations to design a customisable reminder intervention: (1) easy to use and friendly technology, (2) design-your-own- schedule feature, (3) communication style feature, and (4) dialogue support features. Personalisation supported with gamified visual progress tracking and motivational messages emerged as a strong strategy to increase PR adherence. Both sets of findings expand upon the human-computer interaction (HCI) literature for mobile health tools that encourage self-care practices; in particular, to the specific needs of individuals with SCI and the use of visual elements to increase engagement.

Health and toxicity in content moderation: the discursive work of justification

2025-12-13T03:10:25Z

Health and toxicity in content moderation: the discursive work of justification Gibson, Anna D.; Docherty, Niall; Gillespie, Tarleton Within academia, industry, and government, the terms ‘health’ and ‘toxicity’ are widely used to describe and justify decisions around online content and its removal. However, the meanings of these terms are assumed to be self-evident and therefore are rarely examined. This article turns a critical eye to the health and toxicity metaphor to unpack its hidden political work. We trace the metaphor through three different discourses: the historical political economy of the term, the usage by cultural elites in the last two decades, and finally through its contemporary instrumental usage by volunteer content moderators on Facebook. By linking these discourses together, we argue that the metaphor of health and toxicity serves as a means for justification and legitimacy under contemporary neoliberalized orders that typically chafe at modes of public intervention and the language of democratic statecraft. Rather than elucidating the challenges of online content, we find that the metaphor often serves to obfuscate or sidestep the hardest problems in democratic governance. This analysis therefore has practical significance for researchers, policymakers, journalists, and other speakers that publicly traffic in this discourse at large.

Balancing Covariates in Randomized Experiments with the Gram–Schmidt Walk Design

2025-12-13T03:10:28Z

Balancing Covariates in Randomized Experiments with the Gram–Schmidt Walk Design Harshaw, Christopher; Sävje, Fredrik; Spielman, Daniel A; Zhang, Peng The design of experiments involves a compromise between covariate balance and robustness. This article provides a formalization of this tradeoff and describes an experimental design that allows experimenters to navigate it. The design is specified by a robustness parameter that bounds the worst-case mean squared error of an estimator of the average treatment effect. Subject to the experimenter’s desired level of robustness, the design aims to simultaneously balance all linear functions of potentially many covariates. Less robustness allows for more balance. We show that the mean squared error of the estimator is bounded in finite samples by the minimum of the loss function of an implicit ridge regression of the potential outcomes on the covariates. Asymptotically, the design perfectly balances all linear functions of a growing number of covariates with a diminishing reduction in robustness, effectively allowing experimenters to escape the compromise between balance and robustness in large samples. Finally, we describe conditions that ensure asymptotic normality and provide a conservative variance estimator, which facilitate the construction of asymptotically valid confidence intervals. Supplementary materials for this article are available online.

From natural language to simulations: applying AI to automate simulation modelling of logistics systems

2025-12-13T03:10:49Z

From natural language to simulations: applying AI to automate simulation modelling of logistics systems Jackson, Ilya; Jesus Saenz, Maria; Ivanov, Dmitry Our research strives to examine how simulation models of logistics systems can be produced automatically from verbal descriptions in natural language and how human experts and artificial intelligence (AI)-based systems can collaborate in the domain of simulation modelling. We demonstrate that a framework constructed upon the refined GPT-3 Codex is capable of generating functionally valid simulations for queuing and inventory management systems when provided with a verbal explanation. As a result, the language model could produce simulation models for inventory and process control. These results, along with the rapid improvement of language models, enable a significant simplification of simulation model development. Our study offers guidelines and a design of a natural language processing-based framework on how to build simulation models of logistics systems automatically, given the verbal description. In generalised terms, our work offers a technological underpinning of human-AI collaboration for the development of simulation models.

HiTop 2.0: combining topology optimisation with multiple feature size controls and human preferences

2025-12-13T03:10:43Z

HiTop 2.0: combining topology optimisation with multiple feature size controls and human preferences Schiffer, Gillian; Ha, Dat Quoc; Carstensen, Josephine V Topology optimisation is a computational design approach that generates high-performing, efficient structures uniquely suited to a design engineer’s goal. However, there exist two major obstacles to the accessibility, or ease of use, of topology optimisation: expensive computational costs and users’ binary decision between personal intuition and the algorithm’s result. Human-informed topology optimisation, or HiTop, presents an alternative approach to topology optimisation when a user lacks access to a high-performance computer or knowledge of code parameters. HiTop 2.0 prompts users to interactively identify a region of interest in the preliminary design and modify the size of the solid and/or void features. The novel contribution of this paper implements multi-phase minimum and maximum solid feature size controls in HiTop 2.0, and demonstrates 2D and 3D benchmark examples, including test cases that show how the user can interactively enhance issues related to eigenvalues, stress, and energy absorption, while solving the minimum compliance problem.

Integrated urban heat sinks for low-carbon neighbourhoods: dissipating heat to the ground and sky through building structures

2025-12-13T03:10:39Z

Integrated urban heat sinks for low-carbon neighbourhoods: dissipating heat to the ground and sky through building structures Gascón Alvarez, Eduardo; Feickert, Kiley; Ismail, Mohamed A; Mueller, Caitlin T; Norford, Leslie K In a global context of simultaneous urbanization and rising ambient temperatures, it is imperative to design heat-resilient and material-efficient neighbourhoods that respond to the pressing demand for housing with minimal environmental impact. With this goal in mind, the work presented here focuses on the integration of heat dissipation systems within structural building components, introducing a novel framework for their systems-level simulation and design. Two well-studied, low-cost systems (shallow geothermal and night-sky cooling) are modelled within a parametric design workflow that combines bottom-up structural embodied carbon calculations with annual building energy simulations that account for heat sink availability. The proposed method results in a fast and reliable early-stage design tool that allows urban planners, policymakers, and designers to evaluate the suitability of available heat dissipation technologies across climates and urban morphologies. This paper analyzes specifically the multi-domain performance of a hypothetical urban geometry within three different cooling-dominated locations (Algiers, Cairo, and Bangkok).

China’s Potential Lessons from Ukraine for Conflict over Taiwan

2025-12-13T03:10:34Z

China’s Potential Lessons from Ukraine for Conflict over Taiwan Taylor Fravel, M What lessons for a conflict over Taiwan might China be learning from Russia’s invasion of Ukraine and the global responses to the war? And what are the strategic implications of these lessons? To answer these questions, I examine how the war in Ukraine may be shaping China’s assessments of the political, military and economic costs of military action against Taiwan, and how these assessments may influence China’s decision to use force against Taiwan.

BoundarEase: Fostering Constructive Community Engagement to Inform More Equitable Student Assignment Policies

2025-12-12T05:25:02Z

BoundarEase: Fostering Constructive Community Engagement to Inform More Equitable Student Assignment Policies Overney, Cassandra; Moe, Cassandra; Chang, Alvin; Gillani, Nabeel Public school districts across the United States (US) play a pivotal role in shaping access to quality education through their student assignment policies—most prominently, school attendance boundaries. Community engagement processes for changing such policies, however, are often opaque, cumbersome, and highly polarizing—hampering equitable access to quality schools in ways that can perpetuate disparities in achievement and future life outcomes. In this paper, we describe a collaboration with a large US public school district serving nearly 150,000 students to design and evaluate a new sociotechnical system, “BoundarEase”, for fostering more constructive community engagement around changing school attendance boundaries. Through a formative study with 16 community members, we first identify several frictions in existing community engagement processes during boundary planning, like individualistic over collective thinking; a failure to understand and empathize with different community members when considering policy impacts; and challenges in accessing and understanding the impacts of boundary changes. We then use these frictions to inspire the design and development of BoundarEase, a web platform that allows community members to explore and offer feedback on potential boundaries based on their preferences. A user study with 12 community members reveals that BoundarEase prompts reflection among community members on how policies might impact families beyond their own, and increases transparency around the details of policy proposals. Our paper offers education researchers insights into the challenges and opportunities involved in community engagement for designing student assignment policies; human-computer interaction researchers a case study of how new sociotechnical systems might help mitigate polarization in local policymaking; and school districts a practical tool they might use to facilitate community engagement to foster more equitable student assignment policies. Cassandra Overney, Cassandra Moe, Alvin Chang, and Nabeel Gillani. 2025. BoundarEase: Fostering Constructive Community Engagement to Inform More Equitable Student Assignment Policies. Proc. ACM Hum.-Comput. Interact. 9, 2, Article CSCW040 (May 2025), 37 pages.

Multi-objective Evolutionary Learning for Near Pareto-Optimal Optimization of Solar Deployment

2025-12-12T05:25:52Z

Multi-objective Evolutionary Learning for Near Pareto-Optimal Optimization of Solar Deployment Sigrist, Cooper; Li, Archimedes; Zhang, Alice; Lechowicz, Adam; Bashir, Noman; Lertsaroj, Pichsinee; Bahlous-Boldi, Ryan; Hajiesmaili, Mohammad Existing residential rooftop photovoltaic (PV) installations in the United States are inequitable, as they are concentrated in high-income neighborhoods, and carbon-inefficient because they are often not located in electric grids dominated by fossil-fuel generators. Prior work, however, shows that prioritizing socioeconomic equity can also significantly increase the carbon efficiency of new installations. In this paper, we formalize the problem of site selection for rooftop PV installations as a multi-objective optimization problem, with metrics including energy generation, carbon offsetting, and demographic equity. We introduce a novel method called Evolutionary Value Assignment (EVA) that uses a neural network trained via evolutionary learning to select ideal sites for deployment. We evaluate our proposed approach in a case study using a dataset of U.S. solar generation and demographic information. Compared to projections of current installation trends, our method improves Carbon Efficiency by 43%, Income Equity by 41%, and Racial Equity by 24%, while increasing Energy Generation Potential by up to 10%. Therefore, our optimized placement can achieve the estimated carbon offset needed for net-zero emissions from electricity generation earlier than current deployment trends. BUILDSYS ’25, Golden, CO, USA

Robust and expert-agnostic digital twin calibration via ensemble learning and Bayesian optimization

2025-12-12T05:25:38Z

Robust and expert-agnostic digital twin calibration via ensemble learning and Bayesian optimization Zhan, Sicheng; Cui, Bosen Digital twins have emerged as a critical tool in tackling climate change. Considering the data scarcity of complex systems, a promising approach to developing digital twins involves combining physics-based models with data assimilation. However, model calibration remains challenging due to uncertainties in both the physical models and observational data, and the reliance on domain knowledge. In this study, we develop an ensemble learning-based approach that aggregates sub-models with diversified calibration configurations. The proposed method streamlines calibration without expert-driven parameter screening and improves the digital twin's extrapolation capability, enabling more robust predictive applications. We demonstrate the effectiveness of our approach by calibrating the energy model of an office building, significantly reducing the extrapolation error and the associated risks. To the best of our knowledge, this is the first study to facilitate the calibration of physics-based models using ensemble learning, especially in the parameter space. BUILDSYS ’25, Golden, CO, USA

Talk to the Hand: an LLM-powered Chatbot with Visual Pointer as Proactive Companion for On-Screen Tasks

2025-12-12T05:21:10Z

Talk to the Hand: an LLM-powered Chatbot with Visual Pointer as Proactive Companion for On-Screen Tasks Prasongpongchai, Thanawit; Pataranutaporn, Pat; Lertsutthiwong, Monchai; Maes, Pattie This paper presents Pointer Assistant, a novel human-AI interaction technique for on-screen tasks. The design features a chatbot displayed as an extra mouse pointer, alongside the user’s, which proactively gives feedback on user actions while directing them to relevant areas on the screen and responding to the user’s direct chat messages. The effectiveness of the design’s key characteristics, pointer form and proactivity, was investigated in a study involving 220 participants in a financial budget planning task. Results demonstrated that the pointer design and interaction reduced task load while improving satisfaction with the experience, and increased the number of budget categories ideated during the task compared to the traditional passive chat log design. Participants viewed Pointer Assistant as a fun, innovative, and helpful visual guide while noting that its assertiveness can be improved. Future developments could offer even further enhancements to the user experience of human-AI collaboration and task outcomes. CHI ’25, Yokohama, Japan

WireBend-kit: A Computational Design and Fabrication Toolkit for Wirebending Custom 3D Wireframe Structures

2025-12-12T05:25:40Z

WireBend-kit: A Computational Design and Fabrication Toolkit for Wirebending Custom 3D Wireframe Structures Faruqi, Faraz; Paonaskar, Josha; Schuler, Riley; Prevey, Aiden; Taylor, Carson; Tak, Anika; Guinto, Anthony; Shilamkar, Eeshani; Cheenaruenthong, Natarith; Nisser, Martin This paper introduces WireBend-kit, a desktop wirebending machine and computational design tool for creating 3D wireframe structures. Combined, they allow users to rapidly and inexpensively create custom 3D wireframe structures from aluminum wire. Our design tool is implemented in freely available software and allows users to generate virtual wireframe designs and assess their fabricability. A path-planning procedure automatically converts the wireframe design into fabrication instructions for our machine while accounting for material elasticity and kinematic error sources. The custom machine costs $293 in parts and can form aluminum wire into 3D wireframe structures through an ordered sequence of feed, bend, and rotate instructions. Our technical evaluation reveals our system’s ability to overcome odometrically accumulating errors inherent to wirebending in order to produce accurate 3D structures from inexpensive hardware. Finally, we provide application examples demonstrating the design space enabled by Wirebend-kit. SCF ’25, Cambridge, MA, USA

Bias Delayed is Bias Denied? Assessing the Effect of Reporting Delays on Disparity Assessments

2025-12-12T05:25:05Z

Bias Delayed is Bias Denied? Assessing the Effect of Reporting Delays on Disparity Assessments Gosciak, Jennah; Balagopalan, Aparna; Ouyang, Derek; Koenecke, Allison; Ghassemi, Marzyeh; Ho, Daniel Prior work has documented widespread racial and ethnic inequities across sectors, such as healthcare, finance, and technology. Across all of these domains, conducting disparity assessments at regular time intervals is critical for surfacing potential biases in decision-making and improving outcomes across demographic groups. Because disparity assessments fundamentally depend on the availability of demographic information, their efficacy is limited by the availability and consistency of available demographic identifiers. While prior work has considered the impact of missing data on fairness, little attention has been paid to the role of delayed demographic data. Delayed data, while eventually observed, might be missing at the critical point of monitoring and action – and delays may be unequally distributed across groups in ways that distort disparity assessments. We characterize such impacts in healthcare, using electronic health records of over 5M patients across primary care practices in all 50 states. Our contributions are threefold. First, we document the high rate of race and ethnicity reporting delays in a healthcare setting and demonstrate widespread variation in rates at which demographics are reported across different groups. Second, through a set of retrospective analyses using real data, we find that such delays impact disparity assessments and hence conclusions made across a range of consequential healthcare outcomes, particularly at more granular levels of state-level and practice-level assessments. Third, we find limited ability of conventional methods that impute missing race in mitigating the effects of reporting delays on the accuracy of timely disparity assessments. Our insights and methods generalize to many domains of algorithmic fairness where delays in the availability of sensitive information may confound audits, thus deserving closer attention within a pipeline-aware machine learning framework. FAccT ’25, Athens, Greece

Securing Cryptographic Software via Typed Assembly Language

2025-12-12T05:25:37Z

Securing Cryptographic Software via Typed Assembly Language Song, Shixin; Dong, Tingzhen; Nwabueze, Kosi; Zanders, Julian; Erbsen, Andres; Chlipala, Adam; Yan, Mengjia Authors of cryptographic software are well aware that their code should not leak secrets through its timing behavior, and, until 2018, they believed that following industry-standard constant-time coding guidelines was sufficient. However, the revelation of the Spectre family of speculative execution attacks injected new complexities. To block speculative attacks, prior work has proposed annotating the program's source code to mark secret data, with hardware using this information to decide when to speculate (i.e., when only public values are involved) or not (when secrets are in play). While these solutions are able to track secret information stored on the heap, they suffer from limitations that prevent them from correctly tracking secrets on the stack, at a cost in performance. This paper introduces SecSep, a transformation framework that rewrites assembly programs so that they partition secret and public data on the stack. By moving from the source-code level to assembly rewriting, SecSep is able to address limitations of prior work. The key challenge in performing this assembly rewriting stems from the loss of semantic information through the lengthy compilation process. The key innovation of our methodology is a new variant of typed assembly language (TAL), Octal, which allows us to address this challenge. Assembly rewriting is driven by compile-time inference within Octal. We apply our technique to cryptographic programs and demonstrate that it enables secure speculation efficiently, incurring a low average overhead of 1.2%. CCS ’25, Taipei

Study on LLMs for Promptagator-Style Dense Retriever Training

2025-12-12T05:25:42Z

Study on LLMs for Promptagator-Style Dense Retriever Training Gwon, Daniel; Jedidi, Nour; Lin, Jimmy Promptagator demonstrated that Large Language Models (LLMs) with few-shot prompts can be used as task-specific query generators for fine-tuning domain-specialized dense retrieval models. However, the original Promptagator approach relied on proprietary and large-scale LLMs which users may not have access to or may be prohibited from using with sensitive data. In this work, we study the impact of open-source LLMs at accessible scales (≤14B parameters) as an alternative. Our results demonstrate that open-source LLMs as small as 3B parameters can serve as effective Promptagator-style query generators. We hope our work will inform practitioners with reliable alternatives for synthetic data generation and give insights to maximize fine-tuning results for domain-specific applications. Our code is available at https://www.github.com/mitll/promptodile CIKM ’25, Seoul, Republic of Korea

One-Sided Bounded Noise: Theory, Optimization Algorithms and Applications

2025-12-12T05:25:34Z

One-Sided Bounded Noise: Theory, Optimization Algorithms and Applications Xiao, Hanshen; Wan, Jun; Shi, Elaine; Devadas, Srinivas We investigate the optimal trade-off between utility and privacy using one-sided perturbation. Unlike conventional privacy-preserving statistical releases, randomization for obfuscating side-channel information is often constrained by infrastructure limitations. In practical scenarios, these constraints may only allow positive and bounded perturbations. For example, extending processing time or sending and storing dummy messages/data is typically feasible. However, implementing modifications in the opposite direction is challenging due to restrictions imposed by hardware capacity, communication protocols, and data management systems. In this paper, we establish the foundation of the positive noise mechanism within three semantic privacy frameworks: Differential Privacy (DP), Maximal Leakage (MaxL), and Probably Approximately Correct (PAC) Privacy. We then present a series of results that characterize or approximate the optimal one-sided noise distribution, subject to a second-moment budget and a bounded maximal magnitude. Building on this theoretical foundation, we develop efficient tools to solve the underlying optimization problems. Through experiments conducted in various scenarios, we demonstrate that existing techniques, such as Truncated Biased Laplace noise, are often suboptimal and result in excessive performance degradation. For instance, in an anonymous communication system with a 250K message budget, our optimized DP noise mechanism achieves a 21× reduction in dummy messages and an 18× reduction in dummy message latency overhead compared to traditional methods. CCS ’25, Taipei, Taiwan

TH-Wood: Developing Thermo-Hygro-Coordinating Driven Wood Actuators to Enhance Human-Nature Interaction

2025-12-12T05:24:58Z

TH-Wood: Developing Thermo-Hygro-Coordinating Driven Wood Actuators to Enhance Human-Nature Interaction Wang, Guanyun; Chen, Chuang; Jin, Xiao; Chen, Yulu; Zheng, Yangweizhe; Zhen, Qianzi; Zhang, Yang; Li, Jiaji; Yang, Yue; Tao, Ye; Luo, Shijian; Sun, Lingyun Wood has become increasingly applied in shape-changing interfaces for its eco-friendly and smart responsive properties, while its applications face challenges as it remains primarily driven by humidity. We propose TH-Wood, a biodegradable actuator system composed of wood veneer and microbial polymers, driven by both temperature and humidity, and capable of functioning in complex outdoor environments. This dual-factor-driven approach enhances the sensing and response channels, allowing for more sophisticated coordinating control methods. To assist in designing and utilizing the system more effectively, we developed a structure library inspired by dynamic plant forms, conducted extensive technical evaluations, created an educational platform accessible to users, and provided a design tool for deformation adjustments and behavior previews. Finally, several ecological applications demonstrate the potential of TH-Wood to significantly enhance human interaction with natural environments and expand the boundaries of human-nature relationships. CHI ’25, Yokohama, Japan

From blades to tracks: a case study in structural reuse of curved surfaces for circular design

2025-12-12T05:25:49Z

From blades to tracks: a case study in structural reuse of curved surfaces for circular design Pupping, Jesse; Riso, Marzia; Popescu, Mariana; Bousseau, Adrien; Joustra, Jelle We explore the fabrication of curved surfaces by reusing panels extracted from decommissioned wind turbine blades, using cycling pumptracks as a case study. We first present real-world prototypes of pumptrack modules that we manufactured to evaluate the practicality of this reuse scenario and to define the boundary conditions for harvesting blade panels and assembling a track. We then propose an algorithm to optimize the segmentation of a wind turbine blade into quadrilateral panels whose sides fall within a small set of compatible boundaries. These panels form a library of modules that designers can connect side by side to create pumptracks of various lengths and curvatures. Together, these contributions provide a proof-of-concept of how computer-aided design and manufacturing can support circular design through the reuse of curved surfaces.

Fintech Innovation in China

2025-12-12T05:25:15Z

Fintech Innovation in China Cusumano, Michael This column discusses innovation in payment platforms in China and what Western central banks and governments might learn. Private Chinese companies led in the introduction of the mobile payment systems Alipay and WeChat Pay, using QR codes, and most transactions in the country are now digital. China also has banned private crypto currencies and stablecoins and introduced a public digital currency and payment system using crypto technology. However, it has been very difficult to get users to switch to the new central bank digital currency, despite aggressive promotions, subsidies, and mandates. China's experience suggests that other central banks around the world will have difficulty introducing their own digital currencies and competing with private stablecoins and cryptocurrencies as well as other private digital payment platforms.

The Stable Marriage Problem and Sudoku

2025-12-11T03:12:42Z

The Stable Marriage Problem and Sudoku Borodin, Matvey; Chen, Eric; Duncan, Aidan; Khovanova, Tanya; Litchev, Boyan; Liu, Jiahe; Moroz, Veronika; Qian, Matthew; Raghavan, Rohith; Rastogi, Garima; Voigt, Michael Are you having trouble getting married? These days, there are lots of products on the market for dating, from apps to websites and matchmakers, but we know a simpler way! That’s right—your path to coupled life isn’t through Tinder; it’s through Sudoku! Read our fabulous paper, where we explore the Stable Marriage Problem to help you find happiness and stability in marriage through math. As a bonus, you get two Sudoku puzzles with a new flavor.

How Medical Technologies Materialize Oppression

2025-12-11T03:12:52Z

How Medical Technologies Materialize Oppression Boulicault, Marion Biomedical practice can encode and perpetuate oppressive ideologies. This encoding and perpetuation, scholars like Liao and Carbonell (Citation2023) convincingly argue, can occur not only via social practices, but also through medical technologies themselves. In other words, medical technologies can “materialize oppression”: they can be biased in a way that systematically “reflects and perpetuates unjust power relations” (Liao and Carbonell Citation2023, 9). In this paper, I examine how medical technologies materialize oppression, offering a preliminary, non-exhaustive taxonomy of the mechanisms of this materialization. While scholars like Liao and Carbonell focus primarily on physical medical instruments, I offer new examples that illustrate these mechanisms at work, focusing on medical data classification technologies and infrastructures. A clearer view of how these mechanisms operate suggests possibilities for building technologies that liberate rather than oppress.

Wild Wood Gridshells: Mixed-Reality Construction of Nonstandard Wood

2025-12-11T03:12:47Z

Wild Wood Gridshells: Mixed-Reality Construction of Nonstandard Wood Cousin, Tim; Alkhayat, Latifa; Pearl, Natalie; Dewart, Christopher B; Mueller, Caitlin Irregular wood is often downcycled despite having significant embedded strength. Reintegrating this wood into structural assemblies can improve material efficiency in the built environment. This work implemented material logic in a design-to-fabrication workflow for building structures using bifurcated tree branches to leverage this potential (Figure 1). This process is demonstrated through the design and construction of a prototype. A user-oriented computational interface is proposed that manages irregular geometries, matching and optimization algorithms, and structural simulation for design iteration. The demonstrated workflow, which concludes with augmented reality (AR) assisted fabrication, facilitates designing with varying materials, enabling upcycling a wide range of nonstandard building elements. At scale, this methodology can significantly reduce the environmental impact of construction.

MiNav: Autonomous Drone Navigation Indoors using Millimeter-Waves

2025-12-11T03:12:23Z

MiNav: Autonomous Drone Navigation Indoors using Millimeter-Waves Lam, Maisy; Herrera, Joshua; Afzal, Sayed Saad; Zhou, Kaichen; Adib, Fadel We present the design, implementation, and evaluation of MiNav, a system capable of accurate, efficient and fully autonomous drone navigation in challenging indoor environments, including those where vision-based systems fail. MiNav builds on recent literature in millimeter-wave (mmWave) backscatter localization and makes the leap to full end-to-end autonomous mmWave-based navigation. MiNav leverages a mmWave radar mounted on a drone and one or more mmWave backscatter tags deployed in the environment. To enable autonomous navigation, our design introduces key innovations. First, MiNav derives a novel Joint DOP-SNR formulation to probabilistically model uncertainty in localization, and uses this uncertainty to generate an RF-Navigation Map that maximizes the accuracy and reliability of mmWave backscatter localization throughout an environment. It then applies a RF-aware Autonomous Path Planning technique that jointly optimizes for navigation efficiency and localization performance. We built an end-to-end real-time implementation of MiNav consisting of a custom built drone and mmWave backscatter tags. We tested it in practical indoor environments. We run over 165 successful autonomous missions across different tag deployments and demonstrate a median 3D navigation error of 9.1 cm. Our results also show that in comparison to baseline implementations that rely on more classical uncertainty metrics, MiNav achieves a 20% increase in navigation reliability and nearly 3x improvement in self-tracking in millimeter-wave backscatter localization. Finally, we demonstrate first of its kind capabilities, such as fully autonomous, end-to-end mmWave-based drone navigation and path planning in featureless and dark environments. Demo video: http://y2u.be/EpnWibRcxBI

A Civics Lesson for Corporations Seeking to Join a University Community of Innovation

2025-12-11T03:12:51Z

A Civics Lesson for Corporations Seeking to Join a University Community of Innovation Wright, Randall S. Civics, according to Merriam-Webster(2023), is “a social science dealing withthe rights and duties of citizens.”We’ve reached an inflection point.The headline of the July 2023 edi-tion of University-Industry EngagementAdvisor (Lewis 2023) reads “Beforesigning off on strategic partnerships,experts stress value of solid due dili-gence process.”

Artificial intelligence for telemedicine diabetic retinopathy screening: a review

2025-12-11T03:12:49Z

Artificial intelligence for telemedicine diabetic retinopathy screening: a review Nakayama, Luis Filipe; Zago Ribeiro, Lucas; Novaes, Frederico; Miyawaki, Isabele Ayumi; Miyawaki, Andresa Emy; de Oliveira, Juliana Angélica Estevão; Oliveira, Talita; Malerbi, Fernando Korn; Regatieri, Caio Vinicius Saito; Celi, Leo Anthony; Silva, Paolo S PURPOSE: This study aims to compare artificial intelligence (AI) systems applied in diabetic retinopathy (DR) teleophthalmology screening, currently deployed systems, fairness initiatives and the challenges for implementation. METHODS: The review included articles retrieved from PubMed/Medline/EMBASE literature search strategy regarding telemedicine, DR and AI. The screening criteria included human articles in English, Portuguese or Spanish and related to telemedicine and AI for DR screening. The author's affiliations and the study's population income group were classified according to the World Bank Country and Lending Groups. RESULTS: The literature search yielded a total of 132 articles, and nine were included after full-text assessment. The selected articles were published between 2004 and 2020 and were grouped as telemedicine systems, algorithms, economic analysis and image quality assessment. Four telemedicine systems that perform a quality assessment, image preprocessing and pathological screening were reviewed. A data and post-deployment bias assessment are not performed in any of the algorithms, and none of the studies evaluate the social impact implementations. There is a lack of representativeness in the reviewed articles, with most authors and target populations from high-income countries and no low-income country representation. CONCLUSIONS: Telemedicine and AI hold great promise for augmenting decision-making in medical care, expanding patient access and enhancing cost-effectiveness. Economic studies and social science analysis are crucial to support the implementation of AI in teleophthalmology screening programs. Promoting fairness and generalizability in automated systems combined with telemedicine screening programs is not straightforward. Improving data representativeness, reducing biases and promoting equity in deployment and post-deployment studies are all critical steps in model development.

The Removal Chain & Sentient Life Cycles

2025-12-11T03:12:41Z

The Removal Chain & Sentient Life Cycles Schrage, Leonard; Duarte, Fábio; Ratti, Carlo As our cities are growing, managing waste is becoming increasingly challenging. Global plastic waste is set to almost triple by 2060 (OECD Citation2020) while recycling rates are staying below expectations. At the same time, landfills are being relocated away from cities, reaching their maximum capacities, or forced to shut down due to contamination with hazardous materials. As waste management infrastructure is increasingly removed from urban areas, we are becoming further disconnected from its ubiquitous, indispensable, yet invisible life of its own. In recent years, supply chain issues have been an omnipresent reflection of our consumerist reality. For example, when the Ever Given—one of the largest container ships in the world—got stuck in the Suez Canal in 2021 (Chellel et al. Citation2021), we were reminded that our globalized goods travel a long way around the world before they arrive at our doorstep. Still, we tend to forget that there is a life after the supply. On a planet with finite resources and growing piles of (hazardous) trash, we need to look further than the obvious. We urgently need to embrace a circular economy to combat the climate crisis. And to do so, we need to mind both the supply and removal chains.

A two-level machine learning approach for predicting thermal striping in T-junctions with upstream elbow

2025-12-11T03:12:54Z

A two-level machine learning approach for predicting thermal striping in T-junctions with upstream elbow Wang, Yu-Jou; Baglietto, Emilio; Shirvan, Koroush Thermal striping is a phenomenon characterized by oscillatory mixing of non-isothermal streams, which is commonly seen in industrial processes such as nuclear coolant piping, petrochemical plants and liquefied natural gas transportation. The oscillatory mixing of hot and cold fluid can produce thermal field fluctuations and pose a potential risk of high-cycle thermal fatigue failures. Predicting and evaluating spatiotemporal fluctuations in thermal striping often requires high resolution and massive computational power. Although there have been extensive studies using machine learning algorithms on surrogate modeling, research focused on spatiotemporal fluctuation predictions is very limited. Due to the high dimensionality, it often requires complex algorithms with a large amount of high-fidelity training data, which limits the adoption of such methods for industrial applications. In this research, a two-level machine learning framework based on turbulence coherent structures is proposed and its application to a practical problem is demonstrated. The two-level design leverages vortex identification and local bias correction techniques, efficiently reducing the number of full-order simulations required for training. In the first level, well-organized coherent structures are extracted by performing Proper Orthogonal Decomposition on local parameters and then a tree-based machine-learning model is used to down-select the reference structures for the field reconstruction. In the second level, a parameterized convolution neural network is trained to predict the bias introduced by reference structures approximation. The demonstration of the methodology shows that the method can accurately capture the fluctuation frequencies and amplitudes of the spatiotemporal fields in a highly variational setting. Based on the vortex identification method, the methodology is expected to be applicable to general phenomenon driven by large coherent structures.

IP Networks Over Heterogeneous Embedded Serial Links

2025-12-11T03:08:19Z

IP Networks Over Heterogeneous Embedded Serial Links Perry, Nathan The Internet Protocol (IP) provides a number of key benefits to networked devices: it serves as a "narrow waist" enabling functional modularity by decoupling lower-layer devices from application behavior, it provides a notion of transitive connectivity and a number of standardized methods to achieve it, and most importantly, it is ubiquitous, enabling almost all networked applications to mutually communicate. Many embedded microcontrollers cannot take advantage of the benefits of IP because they lack the dedicated networking hardware which is as a practical matter required to interact with nontrivial networks. I observe that multihop point-to-point IP networks can in principle be constructed over the communication media that microcontrollers commonly do have, such as UARTs, I2C, SPI, and CAN bus, but software support is lacking to make this networking approach accessible. Therefore, this thesis develops and evaluates interstice, a platform-independent, open-source software library designed to enable the flexible implementation of modular packet forwarders in userspace. It can be used to interconnect devices and their IP stacks across a variety of conventional and unconventional links. Interstice exposes a reprogrammable, dynamically-updatable packet-forwarding strategy, enabling forwarder nodes in principle to act as hubs, bridges, full routers, or implement firewalls or NAT, as application requirements and platform constraints permit. This approach enables benefits for modular, networked systems of microcontrollers which need to talk to the outside world: using IP enables internal microcontrollers to communicate with external devices such as PCs and smartphones without the need for application gateways. Further, to the extent that such networks are runtime-reconfigurable, features of IP such as address assignment, dynamic routing, and link-agnosticity can be incredibly beneficial. Interstice is evaluated here primarily against networks of various types of serial links (UART, I2c, CAN) speaking PPP, selected to demonstrate utility of the approach to connect embedded devices lacking dedicated networking peripherals, and further that link drivers can be specialized to take advantage of the specific characteristics of each link. The approach is showcased in application scenarios including a networked milling machine, and is analyzed for a number of performance metrics.

BioLIG: Designing Biologically Derived Electronics and Their Speculative Lives

2025-12-11T03:08:22Z

BioLIG: Designing Biologically Derived Electronics and Their Speculative Lives Li, Yuqing Lucy Imagination is the origin of reality. Cultivating new infrastructural and ecological imaginaries is crucial to addressing the climate crisis. Where is the space to prototype new social and technological relations? Transient electronics is an emerging field in advanced materials focused on making electronics that don’t last. Devices are designed to be transient for biomedical, environmental monitoring, or energy storage applications. It is a fascinating and unconventional direction that advances the area of biocompatibility, redefining waste and time-programmable decay {Making electronics that, 2022}. However, in a manufacturing system that fundamentally favors the inert and invariant, transient properties can be precisely the qualities that make adaptation most challenging, often failing at the very stage of imagination. Taking inspiration from transient electronics, this thesis consists of a set of novel biomaterials, a workflow, and three fictional stories to enrich our imagination and instill agency amidst entangled humanitarian, ecological, and technological crises. BioLIG is a material for prototyping accessible and compostable electronics. It uses laser-induced graphene as an organic, bio-derived conductor and affordable biomaterials as the substrate. Three sheets and two inks make up a toolkit to create biocomposites with different properties, colors, and textures specifically designed for prototyping sensors and circuits with transient behaviours. Through a series of characterisations, BioLIG is evaluated and demonstrates that with one material, its electrical performance is on par with synthetic substrates. However, the goal is not to create a replacement material but to prototype new social and technological relations to transient materials. Through a questionnaire, I collected stories, ideas, and questions from makers, designers, and artists for BioLIG and used those as the basis for imagination. In a speculative house, on three floors, three stories unfold of a hoarder, a city forester, and a family living in a time with a leap in our relationship to fabrication, to electronics, and to decay.

Being. Creative. Together. Designing Technologies That Center Human Connection, Co-creativity, and Calm in the Era of AI

2025-12-11T03:06:17Z

Being. Creative. Together. Designing Technologies That Center Human Connection, Co-creativity, and Calm in the Era of AI Dhariwal, Manuj As Artificial Intelligence (AI) becomes increasingly interwoven into our creative, social, and learning experiences, we must ask: Will these technologies deepen our connection to the timeless human experiences of Being, Being Together, and Being Creative Together—or will they pull us apart, leaving us more anxious and isolated? In an era where AI systems are increasingly framed as our “co-creators” and “companions,” enabling hyper-personalized yet hyper-isolated interactions, this dissertation reclaims the prefix ‘co-’ as fundamentally interhuman—introducing a set of new paradigms that center human connection, co-creativity, and calm in the design of technologies. Central to this work, we’ve developed CoCo (coco.build), a general-purpose, real-time co-creative learning platform that empowers young people to engage in a wide variety of safe, shared creative experiences with their peers—spanning creative computing, AI education, digital art, writing, and more. Through the platform, we showcase how digital environments can move beyond isolated modes of learning and creating to support multiple ways of being creative together with others—introducing a new paradigm for real-time digital collaboration. We further illuminate how CoCo has been envisioned as a “self-less” social platform that de-emphasizes comparison-based, self-centric metrics (profiles, likes, followers) prevalent in most online systems for young people. We weave these interconnected ideas into the unifying theme of “Being. Creative. Together.”— values we believe are both timeless and especially timely in the AI era. We supplement the broader design, technical, practical, and pedagogical contributions of this work by sharing insights and feedback from pilots with over 2,000 young people and educators across diverse settings. Ultimately, we see this dissertation as both a contribution and a call—to preserve the human essence of co-, to distinguish it from the useful, powerful, but instrumental AI interactions, and to shape digital environments that nurture our capacity to co-imagine, co-create, co-learn, co-exist, and co-evolve—with and through one another. Note: This work has been co-developed with Shruti Dhariwal. See https://coco.build/thesis for suggested citation and updates on this work.

Toward the computational transformation of legal theory and practice

2025-12-11T03:05:54Z

Toward the computational transformation of legal theory and practice Mahari, Robert This doctoral thesis seeks to advance the formalization of computational law as a distinct research discipline. It explores three interwoven key themes: the empirical understanding of legal systems through advanced computational methods; the development of computational tools to augment the capabilities of legal practitioners, thereby expanding access to justice; and the identification of novel, computationally-enabled regulatory interventions. This research directly confronts the global access to justice crisis and the shortcomings of conventional legal services that frequently leave businesses and individuals without adequate support. Furthermore, the thesis investigates innovative regulatory strategies for emerging technologies, aiming to synchronize legal frameworks with contemporary technological progress by exploring adaptive and forward-looking governance approaches.

Modular Development Platforms and Creative Ecosystems: Design & Deployment for Wide Impact Across Fields

2025-12-11T03:06:21Z

Modular Development Platforms and Creative Ecosystems: Design & Deployment for Wide Impact Across Fields Shtarbanov, Ali Physical, digital, and conceptual tools and building blocks are fundamental enablers and accelerators of humanity’s progress in technology, science, medicine, art, and even in abstract fields like mathematics, philosophy, and social sciences. Hardware development platforms present a special class of tools and building blocks, facilitating and accelerating innovation, prototyping, and research. They drastically reduce prototyping time and complexity, improve efficiency for experts, democratize access to innovation, and even inspire entirely new ideas. This research investigates how to design, develop, and deploy development platforms in ways that maximize their real-world impact potential. It focuses not only on the technical and engineering aspects, but also on the complete ecosystem a platform needs in order to have impact, including community building, engagement with users and volunteers, content strategy, online presence, publicity, deployment, feedback loops modularity, financial viability, and symbiotic relationships. A comprehensive Design & Deployment Framework is introduced as a conceptual tool for creating high-impact platforms and creative ecosystems, recognizing and fostering the positive feedback loops that sustain them and that shape their evolution and growth. This framework is applied in the development and deployment of multiple novel platform and ecosystem projects, including FlowIO, SleeveIO, and ModiStrap, as well as the ecosystem SoftRobotics.IO. Those works have benefited thousands of people around the world, providing researchers, designers, and engineers with powerful, reconfigurable, modular enabling artifacts that streamline prototyping, accelerate research, and lower barriers in fields like soft robotics, haptics, assistive technology, shape-changing interfaces, interactive arts, and more. A multitude of research, art, and engineering projects made possible by FlowIO and SoftRobotics.IO are presented, as well as over a dozen case studies showcasing how other users across disciplines have adopted, utilized, and extended these systems to advance their own creative, educational, and technical endeavors. Additionally, this thesis also investigates various deployment models for hardware and introduces a new hardware deployment model for equitable access to expensive hardware that may otherwise be financially out of reach for many users, as well as an “earned open-source” model, which preserves the essence of the traditional open-source model, while eliminating many of its pitfalls.

Advancing Biosecurity in the Age of AI: Integrating Novel Detection, Suppression, and Evaluation Approaches

2025-12-11T03:08:21Z

Advancing Biosecurity in the Age of AI: Integrating Novel Detection, Suppression, and Evaluation Approaches Justen, Lennart J. Civilization confronts a growing challenge: advancing transformative biological science while safeguarding against catastrophic misuse, a tension amplified by the rapid convergence between biology and artificial intelligence. The COVID-19 pandemic starkly revealed our vulnerabilities to self-replicating, exponential biological phenomena, yet current defenses remain dangerously inadequate—often blind to novel pathogens until too late and lacking barriers against rapid airborne transmission. This thesis argues that robust biosecurity enables, rather than hinders, progress, and advances three key defensive capabilities. First, it evaluates blood metagenomics for pathogen-agnostic surveillance, reanalyzing public datasets to quantify viral signatures and guide the implementation of much-needed early-warning systems sensitive to novel pathogens. Second, it advances far-UVC, a type of ultraviolet between 200-235 nm, for continuous indoor air disinfection, critically assessing its safety profile through an international expert review and establishing research priorities essential for deploying this vital physical defense against airborne threats. Third, it develops rigorous methodologies for evaluating AI's rapidly evolving biological capabilities, benchmarking frontier models across diverse tasks to track progress, reveal limitations in current assessments, and guide responsible innovation in this powerful dual-use technology. Collectively, these contributions help accelerate technologies to mitigate biological risks, thereby helping secure the conditions for continued, beneficial advancement of biology in the age of AI.

From Dialogue to Decision: An LLM-Powered Framework for Analyzing Collective Idea Evolution and Voting Dynamics in Deliberative Assemblies

2025-12-11T03:08:17Z

From Dialogue to Decision: An LLM-Powered Framework for Analyzing Collective Idea Evolution and Voting Dynamics in Deliberative Assemblies Poole-Dayan, Elinor Deliberative assemblies—representative samples of citizens engaged in collective decision-making through facilitated learning and deliberation—are increasingly recognized as powerful tools for revitalizing democratic governance. Yet, core aspects of how deliberation shapes which ideas advance, how perspectives evolve, and why certain recommendations succeed remain opaque and underexamined. This thesis addresses these gaps by investigating: (1) How might we trace the evolution and distillation of ideas into concrete recommendations within deliberative assemblies? and (2) How does the deliberative process shape delegate perspectives and influence voting dynamics over the course of the assembly? To answer these questions, I develop LLM-based methodologies for empirically analyzing transcripts from a tech-enhanced student deliberative assembly. The first framework identifies and visualizes the space of expressed suggestions, revealing that seemingly large gaps between ideas and final recommendations often reflect productive deliberative filtering—while also surfacing overlooked viable ideas. A second analysis integrates post-assembly survey data with transcript-grounded voting patterns to uncover the primary drivers of vote change: edits to recommendations, evolving opinions, and strategic shifts in response to updated priorities. Building on this, I introduce a framework for reconstructing each delegate’s evolving stance across the assembly, linking shifts in perspective to specific deliberative moments and justifications. Together, these methods contribute novel empirical insight into deliberative processes and demonstrate how LLMs can surface high-resolution dynamics otherwise invisible in traditional assembly outputs. The findings lay groundwork for new tools that support facilitators and delegates during live assemblies, improve transparency for decision-makers, and elevate ideas that may otherwise be missed. Looking ahead, this work opens pathways for comparative research across assemblies and highlights the potential for human-centered AI to meaningfully enhance deliberative democratic practice. As societies seek new modes of participatory governance amid growing polarization and institutional mistrust, tools that strengthen deliberation without compromising its core human character are urgently needed.

Private, Verifiable, and Auditable AI Systems

2025-12-11T03:06:13Z

Private, Verifiable, and Auditable AI Systems South, Tobin The growing societal reliance on artificial intelligence necessitates robust frameworks for ensuring its security, accountability, and trustworthiness. This thesis addresses the complex interplay between privacy, verifiability, and auditability in modern AI, particularly in foundation models. It argues that technical solutions that integrate these elements are critical for responsible AI innovation. Drawing from international policy contributions and technical research to identify key risks in the AI pipeline, this work introduces novel technical solutions for critical privacy and verifiability challenges. Specifically, the research introduces techniques for enabling verifiable and auditable claims about AI systems using zero-knowledge cryptography; utilizing secure multi-party computation and trusted execution environments for auditable, confidential deployment of large language models and information retrieval; and implementing enhanced delegation mechanisms, credentialing systems, and access controls to secure interactions with autonomous and multi-agent AI systems. Synthesizing these technical advancements, this dissertation presents a cohesive perspective on balancing privacy, verifiability, and auditability in foundation model-based AI systems, offering practical blueprints for system designers and informing policy discussions on AI safety and governance.

Language Models as Mirrors and Bridges for Intergroup Communication

2025-12-11T03:06:08Z

Language Models as Mirrors and Bridges for Intergroup Communication Jiang, Hang This dissertation explores how large language models (LLMs) can serve dual roles in intergroup communication: as mirrors that reflect intergroup differences, and as bridges that facilitate communication across group boundaries. Intergroup communication refers to interactions between individuals from different social groups, such as political, cultural, or professional communities, where divergent perspectives often lead to misunderstandings, unequal access to information, and social fragmentation. The first part of the dissertation presents LLMs as mirrors that reveal intergroup differences. We first introduce CommunityLM, a novel framework for probing public opinion by fine-tuning LLMs on social media posts from specific communities. Our case study comparing Republican and Democratic groups reveals that model predictions align well with human survey responses, substantially outperforming established baselines. Building on this foundation, we develop PersonaLLM to investigate whether prompt-based LLM agents can generate content aligned with assigned personas, which has emerged as a popular approach for modeling the behaviors of social groups. Through automated and human evaluations, we demonstrate that these agents can complete personality tests and write stories that reflect the distinctive behavioral patterns of specific personality profiles. Together, these complementary projects illustrate how LLMs can effectively capture and simulate the unique perspectives and behaviors that characterize diverse social groups. The second part of the dissertation presents LLMs as bridges that facilitate communication across group boundaries. First, we introduce Bridging Dictionary, an interactive tool that uses retrieval-augmented generation (RAG) techniques with LLMs to identify polarized language and suggest more inclusive alternatives. In collaboration with PBS Frontline, we demonstrate the potential of LLMs to reduce misunderstanding in journalism and political communication. Second, we present Legal Storytelling, a human-LLM collaboration framework that generates accessible narratives to explain complex legal concepts to non-experts. Through randomized controlled trials (RCTs), we find that LLM-generated narratives can improve legal literacy and help bridge communication gaps between experts and laypeople, particularly among non-native English speakers. Third, we develop FaciliTrain, a voice-based, LLM-powered system that enables facilitators to learn and practice intergroup dialogue skills with multiple LLM agents representing diverse social backgrounds and personas in a small-group setting. User studies with campus participants show encouraging early results, suggesting that LLMs can effectively support the development of communication skills essential for constructive intergroup dialogue. Together, these projects illustrate how LLMs can actively foster mutual understanding across social divides by promoting more inclusive, accessible, and constructive communication.

Delibrary: From Discussion to Outcomes and Back(casted) Again, a Visualization Tool for Deliberative Assemblies

2025-12-11T03:08:14Z

Delibrary: From Discussion to Outcomes and Back(casted) Again, a Visualization Tool for Deliberative Assemblies Wong, Wing Cheung Michael With trust in traditional democratic institutions waning, it is increasingly important to examine how potential new institutions could be created and bolstered, with particular emphasis on restoring trust and empowering the public. One potential solution, the citizen's or deliberative assembly, can serve to bridge the governance and legitimacy gap between real-world policy decision-making processes and citizen-driven impact by leveraging random sortition and a well-designed deliberation process. In this thesis, I explore how AI-driven sensemaking via GPT4o-mini--a Large-Language Model (LLM)--synthesized with custom-built visualization tools, can potentially reveal the dynamics within citizen deliberative assemblies where representative, randomly selected citizens navigate public interest issues through facilitated deliberation--and how such tools can serve to amplify transparency within both the assembly process itself and the issues they explore. Through building three different prototype visualization frameworks and the development of an AI-powered topic identification process called backcasting, I analyze novel datasets from two tech-enhanced assemblies; fully recorded discussions from both an on-the-ground citizens' assembly in Deschutes County, Oregon, as well as an MIT student assembly on sustainability. In backcasting, assembly outcomes are linked to transcriptions of assembly discussions via LLM tagging, uncovering what, when, who, and where participants deliberate about topics that eventually become proposals/recommendations/outcomes. Furthermore, I analyze the sentiment with which an assembly delegate presented their view on a certain recommendation (agreement, disagreement, etc.) in addition to the supporting reasoning patterns this delegate used to express their view, if any (e.g. whether they draw from personal experience, reference outside expertise, etc.). To evaluate the final prototype tool, I interview subject matter and assembly experts, assembly organizers/facilitators, as well as assembly delegate members to assess the potential and drawbacks of this visualization tool and AI sensemaking backbone. Positive feedback obtained from these user studies include the clear potential for research, narrative building, and facilitation improvement, in addition to greater perceived transparency into the workings of an assembly process. Further work is still needed, however, to address significant lingering issues, such as adjusting presentation to better serve specific use cases and to reduce complexity and confusion, the most referenced drawback of Delibrary. Overall, my thesis aims to \textbf{build transparent insights into the human-led structures of assemblies, enabling relevant stakeholders--from delegates, policy makers, to the general public--to achieve a better understanding of the assembly process and engender legitimacy perception by illustrating that delegates drawn from all walks of life do have meaningful voice in an impactful process}. By helping to promote this understanding and perception of legitimacy of an effective and respectful deliberation process, I strive to ultimately help scaffold healthier democratic decision-making.

Facilitating Creative Learning: Engaging in a Practice of Care

2025-12-11T03:06:15Z

Facilitating Creative Learning: Engaging in a Practice of Care Presicce, Carmelo Creative learning is shaped not only by tools and activities, but by relationships. This dissertation explores facilitation in creative learning environments as a relational practice centered on care—not as a set of techniques, but as a deeply human way of being with others, a commitment to creating spaces where people feel supported enough to explore, connected enough to share, and valued enough to express themselves. Grounded in constructionist, socioconstructivist, and humanistic pedagogies, the research draws from my multi-year engagement with Learning Creative Learning (LCL)—an online course and global community for educators—and WeScratch, a series of hands-on, collaborative online workshops introducing educators to creative coding. Through qualitative analysis of small-group facilitation during WeScratch workshops, I explore how volunteer facilitators experience and reflect on their practice. Drawing from three case studies, I examine how care takes shape in the situated, relational work of creative learning facilitation. In particular, I identify three interrelated forms of care: epistemic care, which focuses on what and how people learn; affirming care, which supports what learners value and who they are; and convivial care, which attends to how learners feel and relate to one another in a group. After introducing these three forms of care through the work of individual facilitators, I show how epistemic, affirming, and convivial care are deeply interwoven in practice—at times reinforcing one another, at times pulling in different directions. Facilitators must navigate these tensions in the moment, making situated judgments about when to step in, when to hold back, and how to respond to the evolving needs of individuals and groups. By centering care, this research highlights facilitation as deeply human, relational work that sustains the conditions for creative learning, contributing to the broader and evolving discourse on constructionism. It also makes the case for seeing facilitation as an ethical and political practice. In a time when educational discourse is increasingly shaped by ideals of efficiency and optimization—and the world faces rising authoritarianism and dehumanization—choosing to care is not only pedagogically meaningful, but also politically urgent.

Novel Earth Abundant Catalytic Materials for Abatement of Atmospheric Methane Sources, and Evaluation of Agricultural Deployment Environments

2025-12-11T03:06:10Z

Novel Earth Abundant Catalytic Materials for Abatement of Atmospheric Methane Sources, and Evaluation of Agricultural Deployment Environments Brenneis, Rebecca J. Annual global average temperatures in the past year have already exceeded the international target limit of 1.5°C, and the window to prevent that rise from extending is rapidly closing. The high global warming potential (GWP) and short atmospheric residence time (half-life of around 12 years) of methane make it a critical target for action to slow the pace of climate change in this decade. Yet technological solutions for methane abatement are challenged by methane’s inertness, dilute atmospheric concentrations, and diffuse, variable emissions sources. In this thesis, I propose the use of a bio-inspired, earth-abundant, heterogeneous catalysts as a novel tool for atmospheric and emissions-based methane abatement. Copper zeolites were characterized for their ability to convert low levels of methane, continuously, at low temperatures, for moderate durations, and in the presence of a variety of gaseous mixture influents, designed to mimic atmospheric air at standard temperatures and pressures. Catalytic performance was tested under conditions designed to mimic those found at two of the primary sources of low-level, anthropogenic emissions: ventilation air methane (VAM) and industrial dairy. Laboratory synthesized catalysts were shown to completely oxidize methane at concentrations ranging from atmospheric to 1%, covering the range of subflarable levels. Conversion was demonstrated at temperature as low a 270°C, with complete conversion achievable as low as 350°C, in the presence of 20% oxygen. While the presence of water vapor, nitric oxide, and hydrogen sulfide were shown to partially reduce catalytic efficiency, conversion efficiency was restored with increased temperature. The presence of carbon dioxide, alkanes, ammonia and hydrogen, at industrially relevant concentrations, had no effect on catalytic performance. Finally, atmospheric samples were collected at six industrial scale dairy barns across the Midwest and compared with the simulated laboratory conditions. Dairy samples fell within the ranges tested at the bench scale showing no evidence of any impediment to copper zeolite as a potential abatement tool. Methane concentrations at dairies were shown to be on the order of atmospheric to low 10s of ppmv making copper zeolites the only currently identified abatement strategy to address methane emissions at these locations. While it remains to be shown that these zeolites can provide net greenhouse gas benefit in the conditions required, copper zeolites are a strong option on a short list of technologies to address methane at any subflarable concentration, sources of which comprise 80% of global emissions sources, showing great promise as a climate technology breakthrough.

To Co- Is Human: Designing Technologies That Center Human Connection, Co-creativity, and Calm in the Era of AI

2025-12-11T03:06:06Z

To Co- Is Human: Designing Technologies That Center Human Connection, Co-creativity, and Calm in the Era of AI Dhariwal, Shruti In an era where Artificial Intelligence (AI) systems are increasingly framed as our “companions” and “co-creators,” this dissertation reclaims “co-” as a fundamental marker of shared human experience—using it as a foundation to reimagine and build technologies that consciously center interhuman connection and co-creativity. Central to this work, we’ve developed CoCo (coco.build)—a general-purpose, real-time co-creative learning platform that empowers young people to engage in a wide variety of safe, shared creative experiences with their peers, spanning creative computing, AI education, digital art, writing, and more. Through the platform, we showcase how digital environments can move beyond isolated modes of learning and creating to support multiple ways of being creative together with others—introducing a new paradigm for real-time digital collaboration. We further illuminate how CoCo has been envisioned as a “self-less” social platform that de-emphasizes comparison-based, self-centric metrics (profiles, likes, followers) prevalent in most online systems for youth. We anchor these interconnected ideas in a unifying theme of “Being. Creative. Together.”—reflecting timeless values that have become especially timely in an era when AI tools can further accentuate individualized digital experiences for young people. We supplement the broader design, technical, practical, and pedagogical contributions of this work by sharing insights and feedback from pilots with over 2,000 young people and educators across diverse settings. Ultimately, we see this dissertation as both a contribution and a call—to preserve the human essence of co-, to distinguish it from the useful, powerful, but instrumental AI interactions, and to shape digital environments that nurture young people’s capacity to co-imagine, co-create, co-learn, co-exist, and co-evolve—with and through one another. Note: This work has been co-developed with Manuj Dhariwal. See https://coco.build/thesis for suggested citation and updates on this work.

A Dual-Branch Coupled Fourier Neural Operator for High-Resolution Multi-Phase Flow Modeling in Porous Media

2025-12-11T03:12:56Z

A Dual-Branch Coupled Fourier Neural Operator for High-Resolution Multi-Phase Flow Modeling in Porous Media Al Hashim, Hassan; Elyas, Odai; Williams, John This paper investigates a physics-informed surrogate modeling framework for multi-phase flow in porous media based on the Fourier Neural Operator. Traditional numerical simulators, though accurate, suffer from severe computational bottlenecks due to fine-grid discretizations and the iterative solution of highly nonlinear partial differential equations. By parameterizing the kernel integral directly in Fourier space, the operator provides a discretization-invariant mapping between function spaces, enabling efficient spectral convolutions. We introduce a Dual-Branch Adaptive Fourier Neural Operator with a shared Fourier encoder and two decoders: a saturation branch that uses an inverse Fourier transform followed by a multilayer perceptron and a pressure branch that uses a convolutional decoder. Temporal information is injected via Time2Vec embeddings and a causal temporal transformer, conditioning each forward pass on step index and time step to maintain consistent dynamics across horizons. Physics-informed losses couple data fidelity with residuals from mass conservation and Darcy pressure, enforcing the governing constraints in Fourier space; truncated spectral kernels promote generalization across meshes without retraining. On SPE10-style heterogeneities, the model shifts the infinity-norm error mass into the 10−2 to 10−1 band during early transients and sustains lower errors during pseudo-steady state. In zero-shot three-dimensional coarse-to-fine upscaling from 30 ×110 ×5 to 60 ×220 ×5, it attains 𝑅2 =0.90, RMSE = 4.4 ×10−2, and MAE = 3.2 ×10−2, with more than 90% of voxels below five percent absolute error across five unseen layers, while the end-to-end pipeline runs about three times faster than a full-order fine-grid solve and preserves water-flood fronts and channel connectivity. Benchmarking against established baselines indicates a scalable, high-fidelity alternative for high-resolution multi-phase flow simulation in porous media.

Unveiling IPv6 Scanning Dynamics: A Longitudinal Study Using Large Scale Proactive and Passive IPv6 Telescopes

2025-12-10T06:57:25Z

Unveiling IPv6 Scanning Dynamics: A Longitudinal Study Using Large Scale Proactive and Passive IPv6 Telescopes Tanveer, Hammas Bin; Chan, Echo; Mok, Ricky K. P.; Kappes, Sebastian; Richter, Philipp; Gasser, Oliver; Ronan, John; Berger, Arthur; Claffy, kc We introduce new tools and vantage points to develop and integrate proactive techniques to attract IPv6 scan traffic, thus enabling its analysis. By deploying the largest-ever IPv6 proactive telescope in a production ISP network, we collected over 600M packets of unsolicited traffic from 1.9k Autonomous Systems in 10 months. We characterized the sources of unsolicited traffic, evaluated the effectiveness of five major features across the network stack, and inferred scanners' sources of target addresses and their strategies.

HARMONY: A Scalable Distributed Vector Database for High-Throughput Approximate Nearest Neighbor Search

2025-12-10T06:57:53Z

HARMONY: A Scalable Distributed Vector Database for High-Throughput Approximate Nearest Neighbor Search Xu, Qian; Zhang, Feng; Li, Chengxi; Cao, Lei; Chen, Zheng; Zhai, Jidong; Du, Xiaoyong Approximate Nearest Neighbor Search (ANNS) is essential for various data-intensive applications, including recommendation systems, image retrieval, and machine learning. Scaling ANNS to handle billions of high-dimensional vectors on a single machine presents significant challenges in memory capacity and processing efficiency. To address these challenges, distributed vector databases leverage multiple nodes for the parallel storage and processing of vectors. However, existing solutions often suffer from load imbalance and high communication overhead, primarily due to traditional partition strategies that fail to effectively distribute the workload. In this paper, we introduce Harmony, a distributed ANNS system that employs a novel multi-granularity partition strategy, combining dimension-based and vector-based partition. This strategy ensures a balanced distribution of computational load across all nodes while effectively minimizing communication costs. Furthermore, Harmony incorporates an early-stop pruning mechanism that leverages the monotonicity of distance computations in dimensionbased partition, resulting in significant reductions in both computational and communication overhead. We conducted extensive experiments on diverse real-world datasets, demonstrating that Harmony outperforms leading distributed vector databases, achieving 4.63× throughput on average in four nodes and 58% performance improvement over traditional distribution for skewed workloads.

Kernel Extension DSLs Should Be Verifier-Safe!

2025-12-10T06:57:52Z

Kernel Extension DSLs Should Be Verifier-Safe! Solleza, Franco; Adam, Justus; Crotty, Andrew; Narayan, Akshay; Schwarzkopf, Malte; Tatbul, Nesime eBPF allows developers to write safe operating system extensions, but writing these extensions remains challenging because it requires detailed knowledge of both the extension's domain and eBPF's programming interface. Most importantly, the extension must pass the eBPF verifier. This paper argues that DSLs for extensions should guarantee verifier-safety: valid DSL programs should result in eBPF code that always passes the verifier. This avoids complex debugging and the need for extension developers to be eBPF experts. We show that three existing DSLs for different domains are compatible with verifier-safety. Beyond verifier-safety, practical extension DSLs must also achieve good performance. Inspired by database query optimization, we sketch an approach to creating DSL-specific optimizers capable of maintaining verifier-safety. A preliminary evaluation shows that optimizing verifier-safe extension performance is feasible. eBPF ’25, September 8–11, 2025, Coimbra, Portugal

Experiencing EmbedNet: Embedding self-sensing to 3D casting objects

2025-12-10T06:57:19Z

Experiencing EmbedNet: Embedding self-sensing to 3D casting objects Liu, Fangzheng; Dementyev, Artem; Wicaksono, Irmandy; Paradiso, Joseph This paper introduces EmbedNet, a method for integrating dense sensor networks into casting objects. With EmbedNet, sensor nodes are seamlessly incorporated into casting objects during fabrication. The process involves extruding base materials like silicone rubber or liquid plastic and a custom-designed sensor strip using a hand-held extruder into a mold tailored to specific applications. The base material mixes with the sensor strip in the mold, and upon curing, the result is an object with a defined shape housing a sensor network. EmbedNet employs a small Host node to access sensor data from all nodes on the strip. Each sensor node is self-contained and provides status indications through an onboard RGB LED. The Host connects with all sensor nodes using just three wires: power, ground, and data. This one-wire communication is facilitated through a custom-designed software serial port for each sensor node. The paper showcases various applications of EmbedNet, including wearables, home sensing, and entertainment devices. UIST Adjunct ’25, Busan, Republic of Korea

Demonstrating NeuroFlux: A Non-Invasive Peripheral Magnetic Stimulation Device for Multimodal Haptic Feedback

2025-12-10T06:57:50Z

Demonstrating NeuroFlux: A Non-Invasive Peripheral Magnetic Stimulation Device for Multimodal Haptic Feedback Huang, Bingjian; Chin, Sam; Wigdor, Daniel; Paradiso, Joseph We demonstrate NeuroFlux, a wearable armband that delivers multimodal haptic feedback through non-invasive peripheral magnetic stimulation. Unlike conventional haptic devices limited to either tactile or kinesthetic modalities, NeuroFlux stimulates peripheral nerves to independently evoke both muscle movements and localized skin sensations. Our system features a custom-designed control circuit and a multi-coil armband, enabling precise, real-time control of stimulation location and intensity. This hardware innovation significantly expands the design space of haptic feedback by bridging kinesthetic and tactile modalities through a single, compact device. In our demonstration, participants will experience a wide range of magnetically induced haptic sensations, including independent stimulation of muscular and cutaneous nerves in the forearm. The setup includes interactive tasks that showcase NeuroFlux’s ability to generate diverse haptic effects such as finger flexion, wrist movement, as well as immersive virtual reality object interactions. By offering hands-on exposure to peripheral magnetic stimulation, we aim to spark new research directions in multimodal haptic feedback and make neural stimulation more accessible to the HCI community. UIST Adjunct ’25, Busan, Republic of Korea

Supernotes: Driving Consensus in Crowd-Sourced Fact-Checking

2025-12-10T06:57:49Z

Supernotes: Driving Consensus in Crowd-Sourced Fact-Checking De, Soham; Bakker, Michiel; Baxter, Jay; Saveski, Martin X's Community Notes, a crowd-sourced fact-checking system, allows users to annotate potentially misleading posts. Notes rated as helpful by a diverse set of users are prominently displayed below the original post. While demonstrably effective at reducing misinformation's impact when notes are displayed, there is an opportunity for notes to appear on many more posts: for 91% of posts where at least one note is proposed, no notes ultimately achieve sufficient support from diverse users to be shown on the platform. This motivates the development of Supernotes: AI-generated notes that synthesize information from several existing community notes and are written to foster consensus among a diverse set of users. Our framework uses an LLM to generate many diverse Supernote candidates from existing proposed notes. These candidates are then evaluated by a novel scoring model, trained on millions of historical Community Notes ratings, selecting candidates that are most likely to be rated helpful by a diverse set of users. To test our framework, we ran a human subjects experiment in which we asked participants to compare the Supernotes generated by our framework to the best existing community notes for 100 sample posts. We found that participants rated the Supernotes as significantly more helpful, and when asked to choose between the two, preferred the Supernotes 75.2% of the time. Participants also rated the Supernotes more favorably than the best existing notes on quality, clarity, coverage, context, and argumentativeness. Finally, in a follow-up experiment, we asked participants to compare the Supernotes against LLM-generated summaries and found that the participants rated the Supernotes significantly more helpful, demonstrating that both the LLM-based candidate generation and the consensus-driven scoring play crucial roles in creating notes that effectively build consensus among diverse users. WWW ’25, Sydney, NSW, Australia

I Feel Your Pain: a Haptic Interface for Improving Pain Literacy

2025-12-10T06:57:16Z

I Feel Your Pain: a Haptic Interface for Improving Pain Literacy Yin, Peggy; Chen, Sofia; Chang, Ethan There is no sensation more universal and misunderstood than pain. While pain presents itself in nearly every eukaryotic organism, it remains one of the most elusive disease states to express, let alone treat. Here, we introduce Pain by Numbers, a haptic, immersive storytelling interface that facilitates user recognition and communication of low-to-medium-intensity pain, in order to improve pain literacy for patients, physicians, and society-at-large. UIST Adjunct ’25, Busan, Republic of Korea

Understanding and Scaling Collaborative Filtering Optimization from the Perspective of Matrix Rank

2025-12-10T06:57:47Z

Understanding and Scaling Collaborative Filtering Optimization from the Perspective of Matrix Rank Loveland, Donald; Wu, Xinyi; Zhao, Tong; Koutra, Danai; Shah, Neil; Ju, Mingxuan Collaborative Filtering (CF) methods dominate real-world recommender systems given their ability to learn high-quality, sparse ID-embedding tables that effectively capture user preferences. These tables scale linearly with the number of users and items, and are trained to ensure high similarity between embeddings of interacted user-item pairs, while maintaining low similarity for non-interacted pairs. Despite their high performance, encouraging dispersion for non-interacted pairs necessitates expensive regularization (e.g., negative sampling), hurting runtime and scalability. Existing research tends to address these challenges by simplifying the learning process, either by reducing model complexity or sampling data, trading performance for runtime. In this work, we move beyond model-level modifications and study the properties of the embedding tables under different learning strategies. Through theoretical analysis, we find that the singular values of the embedding tables are intrinsically linked to different CF loss functions. These findings are empirically validated on real-world datasets, demonstrating the practical benefits of higher stable rank -- a continuous version of matrix rank which encodes the distribution of singular values. Based on these insights, we propose an efficient warm-start strategy that regularizes the stable rank of the user and item embeddings. We show that stable rank regularization during early training phases can promote higher-quality embeddings, resulting in training speed improvements of up to 65.9%. Additionally, stable rank regularization can act as a proxy for negative sampling, allowing for performance gains of up to 21.2% over loss functions with small negative sampling ratios. Overall, our analysis unifies current CF methods under a new perspective -- their optimization of stable rank -- motivating a flexible regularization method that is easy to implement, yet effective at enhancing CF systems. WWW ’25, April 28-May 2, 2025, Sydney, NSW, Australia

21W.749 / CMS.935 Documentary Photography and Photojournalism: Still Images of a World in Motion, Spring 2016

2025-12-09T19:31:47Z

21W.749 / CMS.935 Documentary Photography and Photojournalism: Still Images of a World in Motion, Spring 2016 Colen, B. D. In this course, you will be exposed to the work of many great documentary photographers and photojournalists, as well as to writing about the documentary tradition. Further, throughout the term, you will hone your photographic skills and 'eye,' and you will work on a photo documentary project of your own, attempting to reduce a tiny area of the moving world to a set of still images that convey what the viewer needs to know about what you saw—without hearing the sounds, smelling the odors, experiencing what was happening outside the viewfinder, and without seeing the motion.

Unified and Generalizable Reinforcement Learning for Facility Location Problems on Graphs

2025-12-10T06:56:51Z

Unified and Generalizable Reinforcement Learning for Facility Location Problems on Graphs Guo, Wenxuan; Wang, Runzhong; Xu, Yanyan; Jin, Yaohui Facility location problems on graphs are ubiquitous in the real world and hold significant importance, yet their resolution is often impeded by NP-hardness. MIP solvers can find the optimal solutions but fail to handle large instances, while algorithm efficiency has a higher priority in cases of emergency. Recently, machine learning methods have been proposed to tackle such classical problems with fast inference, but they are limited to the myopic constructive pattern and only consider simple cases in Euclidean space. This paper introduces a unified and generalizable approach to tackle facility location problems on weighted graphs with deep reinforcement learning, demonstrating a keen awareness of complex graph structures. Striking a harmonious balance between solution quality and running time, our method stands out with superior efficiency and steady performance. Our model trained on small graphs is highly scalable and consistently generates high-quality solutions, achieving a speedup of more than 2000 times to Gurobi on instances with 1000 nodes. The experiments on Shanghai road networks further demonstrate its practical value in solving real-world problems. The source codes are available at https://github.com/AryaGuo/PPO-swap. WWW ’25, Sydney, NSW, Australia

Paratrouper: Exploratory Creation of Character Cast Visuals Using Generative AI

2025-12-10T06:57:09Z

Paratrouper: Exploratory Creation of Character Cast Visuals Using Generative AI Leong, Joanne; Ledo, David; Driscoll, Thomas; Grossman, Tovi; Fitzmaurice, George; Anderson, Fraser Great characters are critical to the success of many forms of media, such as comics, games, and films. Designing visually compelling casts of characters requires significant skill and consideration, and there is a lack of specialized tools to support this endeavor. We investigate how AI-driven image-generation techniques can empower creatives to explore a variety of visual design possibilities for individual and groups of characters. Informed by interviews with character designers, Paratrouper is a multi-modal system that enables creating and experimenting with multiple permutations for character casts and visualizing them in various contexts as part of a holistic approach to design. We demonstrate how Paratrouper supports different aspects of the character design process, and share insights from its use by eight creators. Our work highlights the interplay between creative agency and serendipity, as well as the visual interrelationships among character aesthetics. Joanne Leong, David Ledo, Thomas Driscoll, Tovi Grossman, George Fitzmaurice, and Fraser Anderson. 2025. Paratrouper: Exploratory Creation of Character Cast Visuals Using Generative AI. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25). Association for Computing Machinery, New York, NY, USA, Article 189, 1–20.

FiberCircuits: A Miniaturization Framework To Manufacture Fibers That Embed Integrated Circuits

2025-12-10T06:57:23Z

FiberCircuits: A Miniaturization Framework To Manufacture Fibers That Embed Integrated Circuits Honnet, Cedric; Babatain, Wedyan; Luo, Yiyue; Kilic Afsar, Ozgun; Bensahel, Chloe; Nicita, Sarah; Zhu, Yunyi; Danielescu, Andreea; Gershenfeld, Neil; Paradiso, Joseph UIST ’25, Busan, Republic of Korea

FIP: Endowing Robust Motion Capture on Daily Garment by Fusing Flex and Inertial Sensors

2025-12-10T06:57:04Z

FIP: Endowing Robust Motion Capture on Daily Garment by Fusing Flex and Inertial Sensors Zheng, Ruonan; Fang, Jiawei; Yao, Yuan; Gao, Xiaoxia; Zuo, Chengxu; Guo, Shihui; Luo, Yiyue What if our clothes could capture our body motion accurately? This paper introduces Flexible Inertial Poser (FIP), a novel motion-capturing system using daily garments with two elbow-attached flex sensors and four Inertial Measurement Units (IMUs). To address the inevitable sensor displacements in loose wearables which degrade joint tracking accuracy significantly, we identify the distinct characteristics of the flex and inertial sensor displacements and develop a Displacement Latent Diffusion Model and a Physics-informed Calibrator to compensate for sensor displacements based on such observations, resulting in a substantial improvement in motion capture accuracy. We also introduce a Pose Fusion Predictor to enhance multimodal sensor fusion. Extensive experiments demonstrate that our method achieves robust performance across varying body shapes and motions, significantly outperforming SOTA IMU approaches with a 19.5% improvement in angular error, a 26.4% improvement in elbow angular error, and a 30.1% improvement in positional error. FIP opens up opportunities for ubiquitous human-computer interactions and diverse interactive applications such as Metaverse, rehabilitation, and fitness analysis. Our project page can be seen at Flexible Inertial Poser. CHI ’25, Yokohama, Japan

GPU-accelerated dynamic nonlinear optimization with ExaModels and MadNLP

2025-12-10T06:57:44Z

GPU-accelerated dynamic nonlinear optimization with ExaModels and MadNLP Pacaud, François; Shin, Sungho We investigate the potential of Graphics Processing Units (GPUs) to solve large-scale nonlinear programs with a dynamic structure. Using ExaModels, a GPU-accelerated automatic differentiation tool, and the interior-point solver MadNLP, we significantly reduce the time to solve dynamic nonlinear optimization problems. The sparse linear systems formulated in the interior-point method is solved on the GPU using a hybrid solver combining an iterative method with a sparse Cholesky factorization, which harness the newly released NVIDIA cuDSS solver. Our results on the classical distillation column instance show that despite a significant pre-processing time, the hybrid solver allows to reduce the time per iteration by a factor of $\mathbf{2 5}$ for the largest instance. 2024 IEEE 63rd Conference on Decision and Control (CDC), Milan, Italy, 2024

Scalable Primal Decomposition Schemes for Large-Scale Infrastructure Networks

2026-03-08T03:32:20Z

Scalable Primal Decomposition Schemes for Large-Scale Infrastructure Networks Engelmann, Alexander; Shin, Sungho; Pacaud, François; Zavala, Victor M The operation of large-scale infrastructure networks requires scalable optimization schemes. To guarantee safe system operation, a high degree of feasibility in a small number of iterations is important. Decomposition schemes can help to achieve scalability. In terms of feasibility, however, classical approaches, such as the alternating direction method of multipliers (ADMMs), often converge slowly. In this work, we present primal decomposition schemes for hierarchically structured strongly convex quadratic programs. These schemes offer high degrees of feasibility in a small number of iterations in combination with global convergence guarantees. We benchmark their performance against the centralized off-the-shelf interior-point solver Ipopt and ADMM on problems with up to 300 000 decision variables and constraints. We find that the proposed approaches solve problems as fast as Ipopt, but with reduced communication and without requiring a full model exchange. Moreover, the proposed schemes achieve a higher accuracy than ADMM.

Coplanarity of rooted spanning-tree vectors

2025-12-09T03:11:01Z

Coplanarity of rooted spanning-tree vectors Polettini, Matteo; Harunari, Pedro E.; Cengio, Sara D.; Lecomte, Vivien Employing a recent technology of tree surgery, we prove a “deletion–constriction” formula for products of rooted spanning-trees on weighted directed graphs that generalizes deletion–contraction on undirected graphs. The formula implies that, letting τ x ∅ , τ x + , and τ x - be the rooted spanning-tree polynomials obtained, respectively, by removing both directed edges between two vertices, or by forcing the tree to pass through either edge, the vectors ( τ x ∅ , τ x + , τ x - ) are coplanar for all roots x . We deploy the result to give an alternative derivation of a recently found mutual linearity of stationary currents of Markov chains. We generalize deletion–constriction and current linearity for two pairs of edges and conjecture that similar results may hold for arbitrary subsets of edges.

Search for the decay B0 → ϕϕ

2026-03-08T03:32:10Z

Search for the decay B0 → ϕϕ Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J.; Aleksiejunas, R. A search for the decay B0 → ϕϕ is made using pp collision data collected with the LHCb detector at centre-of-mass energies of 7, 8 and 13 TeV, corresponding to an integrated luminosity of 9 fb−1. No significant signal is observed, and an upper limit on the branching fraction of 1.3 (1.4) × 10−8 at 90 (95)% confidence level is set. This result supersedes the previous LHCb study and improves the upper limit by a factor of two.

Regulating Sommerfeld resonances for multi-state systems and higher partial waves

2026-03-08T03:32:09Z

Regulating Sommerfeld resonances for multi-state systems and higher partial waves Parikh, Aditya; Sato, Ryosuke; Slatyer, Tracy R. Long-range attractive interactions between dark matter particles can significantly enhance their annihilation, particularly at low velocities. This “Sommerfeld enhancement” is typically computed by evaluating the deformation of the two-particle wavefunction due to the long-range potential, while ignoring the physics associated with the annihilation, and then scaling the appropriate annihilation matrix elements by factors that depend on the wavefunction in the limit where the particles approach zero relative separation. It has long been recognized that this approach is a valid approximation only in the limit where the annihilation rate is small, and breaks down in the regime where the enhanced annihilation rate approaches the unitarity bound, in which case ignoring the impact of the annihilation physics on the two-particle wavefunction cannot be justified and leads to apparent violations of unitarity. In the case where the physics relevant to annihilation occurs at a parametrically shorter distance scale (higher energy scale) compared with the long-range potential, we provide a simple prescription for correcting the Sommerfeld enhancement for the effects of the short-range physics, valid for all partial waves and for systems where multiple states are coupled by the long-range potential.

A field guide to event-shape observables using optimal transport

2026-03-08T03:32:07Z

A field guide to event-shape observables using optimal transport Cesarotti, Cari; LeBlanc, Matt We lay out the phenomenological behavior of event-shape observables evaluated by solving optimal transport problems between collider events and reference geometries — which we name manifold distances — to provide guidance regarding their use in future studies. This discussion considers several choices related to the metric used to quantify these distances. We explore the differences between the various options, for the first time using a combination of analytical studies and simulated minimum-bias and multi-jet events. Making judicious choices when defining the metric and reference geometry can improve sensitivity to interesting signal features and reduce sensitivity to non-perturbative effects in QCD. The goal of this article is to provide a ‘field guide’ that can inform how choices made when defining a manifold distance can be tailored for the analysis at-hand.

Search for the lepton-flavour-violating decays B0 → K*0τ±e∓

2026-03-08T03:32:10Z

Search for the lepton-flavour-violating decays B0 → K*0τ±e∓ Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J. A first search at LHCb for the lepton-flavour-violating decays B0 → K*0τ±e∓ is presented. The analysis is performed using a sample of proton-proton collision data, collected with the LHCb detector at a centre-of-mass energy of 13 TeV between 2016 and 2018, corresponding to an integrated luminosity of 5.4 fb−1. No significant signal is observed, and upper limits on the branching fractions are determined to be B B 0 → K ∗ 0 τ − e + < 5.9 7.1 × 10 − 6 and B B 0 → K ∗ 0 τ + e − < 4.9 5.9 × 10 − 6 at the 90% (95%) confidence level. These results correspond to the current most stringent upper limits for b → sτl transitions.

Guidelines for environmental life cycle assessment of cultivated meat

2026-03-08T03:32:18Z

Guidelines for environmental life cycle assessment of cultivated meat Blackstone, Nicole T.; Pavlova, Anisiya; Trinidad, Kirsten R.; Nikkhah, Amin; Sinke, Pelle; Heller, Martin; Duncan-Duggal, Joe; Ridoutt, Brad; Smetana, Sergiy; Makov, Tamar; Shabtai, Shira Purpose Cultivated meat is produced by growing animal cells in vitro without using, or reducing the use of, animals for meat, poultry, or seafood production. Responsibly and consistently investigating the environmental impacts of cultivated meat is essential to provide reliable performance benchmarks and realistic comparisons with animal-based production systems. In this contribution, we provide technical, actionable guidelines for conducting life cycle assessments (LCAs) of cultivated meat and highlight further research needs for the field. Methods We assembled a global team of recognized and active scientists in cultivated meat LCA, livestock systems LCA, and ISO LCA standards to develop this set of guidelines using a workshop (in person and online) and online meetings, as well as asynchronous feedback, to reach consensus. Results and discussion These guidelines provide specifications throughout the four phases of LCA, from goal definition to the interpretation of LCA results. Data gaps, including the availability and quality of feed or food-grade culture media component inventories, are among the areas highlighted for further exploration. Conclusion We invite LCA practitioners to apply these guidelines when investigating cultivated meat systems to increase the consistency and reliability of environmental impact evaluations for these emerging products.

Sharp Bound for the Erdős–Straus Non-averaging Set Problem

2026-03-08T03:32:21Z

Sharp Bound for the Erdős–Straus Non-averaging Set Problem Pham, Huy T.; Zakharov, Dmitrii A set of integers A is non-averaging if there is no element a in A which can be written as an average of a subset of A not containing a . We show that the largest non-averaging subset of { 1 , … , n } has size n 1 / 4 + o ( 1 ) , thus solving the Erdős–Straus problem. We also determine the largest size of a non-averaging set in a d -dimensional box for any fixed d . Our main tool includes the structure theorem for the set of subset sums due to Conlon, Fox and the first author, together with a result about the structure of a point set in nearly convex position.

Seasonal variations of the atmospheric muon neutrino spectrum measured with IceCube

2026-03-08T03:32:19Z

Seasonal variations of the atmospheric muon neutrino spectrum measured with IceCube IceCube Collaboration This study presents an analysis of seasonal variations in the atmospheric muon neutrino flux, using 11.3 years of data from the IceCube Neutrino Observatory. By leveraging a novel spectral unfolding method, we explore the energy range from 125 GeV to 10 TeV for zenith angles from 90 ∘ to 110 ∘ , corresponding to the Antarctic atmosphere. Our findings reveal that the differential measurement of the amplitudes of the seasonal variation is consistent with an energy-dependent decrease reaching ( - 4.5 ± 1.2)% during Austral winter and increase to (+ 3.9 ± 1.3)% during Austral summer relative to the annual average at 10 TeV. While the unfolded flux exceeds the model predictions by up to 30%, the differential measurement of the seasonal to annual average flux remains unaffected. The measured seasonal variations of the muon neutrino spectrum are consistent with theoretical predictions using the MCEq code and the NRLMSISE-00 atmospheric model.

Capacity lower bound for the Ising perceptron

2026-03-08T03:32:21Z

Capacity lower bound for the Ising perceptron Ding, Jian; Sun, Nike We consider the Ising perceptron with gaussian disorder, which is equivalent to the discrete cube { - 1 , + 1 } N intersected by M random half-spaces. The perceptron’s capacity is the largest integer M N for which the intersection is nonempty. It is conjectured by Krauth and Mézard (1989) that the (random) ratio M N / N converges in probability to an explicit constant α ⋆ ≐ 0.83 . Kim and Roche (1998) proved the existence of a positive constant γ such that γ ⩽ M N / N ⩽ 1 - γ with high probability; see also Talagrand (1999). In this paper we show that the Krauth–Mézard conjecture α ⋆ is a lower bound with positive probability, under the condition that an explicit univariate function S ⋆ ( λ ) is maximized at λ = 0 . Our proof is an application of the second moment method to a certain slice of perceptron configurations, as selected by the so-called TAP (Thouless, Anderson, and Palmer, 1977) or AMP (approximate message passing) iteration, whose scaling limit has been characterized by Bayati and Montanari (2011) and Bolthausen (2012).

Tackling the UK’s regional economic inequality: binding constraints and avenues for policy intervention

2026-03-08T03:32:17Z

Tackling the UK’s regional economic inequality: binding constraints and avenues for policy intervention Stansbury, Anna; Turner, Dan; Balls, Ed We analyse binding constraints to productivity growth in the UK’s regions outside London and the greater South East. These analyses challenge a number of common arguments about the UK’s regional economic inequality problem. We find little evidence consistent with the hypotheses (i) that low shares of university graduates remain the primary constraint on growth for the UK’s regions; (ii) that there is a generalised issue with access to finance for firms outside the South East; or (iii) that low or falling regional migration rates are to blame for the persistence of the UK’s regional economic inequalities. Instead, we find evidence consistent with (i) a specific relative shortage of STEM degrees; (ii) binding transport infrastructure constraints within major non-London conurbations; (iii) a failure of public innovation policy to support clusters beyond the South East, in particular through the regional distribution of public support for Research and Development (R&D); and (iv) missed opportunities for higher internal mobility due to London’s overheating housing market. We also find some suggestive evidence consistent with constraints on access to early-stage equity financing for high-growth-potential SMEs in certain regions.

Chile’s Inclusion Law: the arduous drive to regulate an unequal education system, 2006–19

2026-03-08T03:32:12Z

Chile’s Inclusion Law: the arduous drive to regulate an unequal education system, 2006–19 Cummings, Peter MM; Mizala, Alejandra; Schneider, Ben Ross Chile’s Inclusion Law, passed in 2015, significantly increased government regulation of one of the most privatised education systems in the world and provided major redistributive benefits. How did Chile’s government succeed in passing and implementing this legislation in the face of a powerful and cohesive opposition? Our study finds that student protesters served as the initial impetus, shaping the education debate and increasing the political salience and urgency of education reform. In line with power resource theory, other left movement organisations and voters used their power to support redistributive education reform, and Bachelet’s centre-left coalition followed through on its mandate by proposing the Inclusion Law. Also, a well-connected policy network helped articulate problems with the status quo and shaped the specifics of the education bill. To develop this argument, the paper draws on historical information on the student movement in Chile, quantitative data on education stakeholder appearances in the press, public opinion surveys, and detailed analysis of the 13-month legislative proceedings – to explain the law’s passage in congress. To underscore the significance of the Inclusion Law and to contextualise the Chilean case, the paper also compares Chile to other countries with nation-wide school choice systems.

Bayesian Network–Based Fault Diagnostic System for Nuclear Power Plant Assets

2026-03-08T03:32:11Z

Bayesian Network–Based Fault Diagnostic System for Nuclear Power Plant Assets Zhao, Xingang; Wang, Xinyan; Golay, Michael W Future advances in nuclear power technologies call for enhanced operator advice and autonomous control capabilities that can leverage simpler designs and increased safety features to reduce reliance on human labor. One of the first tasks in the development of such capabilities is the formulation of symptom-based conditional failure probabilities for the plant structures, systems, and components (SSCs) of interest. The primary goal is to aid plant personnel in (1) deducing the probabilistic performance status of the monitored SSCs and (2) detecting impending faults/failures. The task of estimating conditional failure probability is a bidirectional inference problem, and a logical approach is to use the Bayesian network (BN) method. As a knowledge-based explainable artificial intelligence tool and a probabilistic graphical model, BN offers reasoning capability under uncertainty, graphical representation emulating physical behavior of the target SSC, and interpretability of the model structure and results. This paper provides a systematic overview of the BN technique and the software tools for implementing BN models, along with the associated knowledge representation and reasoning paradigm. Both operational data and expert judgment can be readily incorporated into the knowledge base of a BN model. The challenges with data availability are highlighted, and the general approach to target SSC identification is presented. The focus is on failure-prone and risk-important balance of plant assets, especially for cases with strong operator involvement. Two example case studies on the failure of (1) a centrifugal pump and (2) an electric motor are conducted to demonstrate the usefulness and technical feasibility of the proposed BN-based fault diagnostic system using an expert system shell.

Your body tells more than words – predicting perceived meeting productivity through body signals

2026-03-08T03:32:19Z

Your body tells more than words – predicting perceived meeting productivity through body signals Zeyda, Maximilian; Stracke, Selina; Knipfer, Kristin; Gloor, Peter A The productivity of work meetings is mostly assessed through post-hoc questionnaires. These questionnaires are impractical as they require additional time after the meeting has ended. Thus, measuring meeting productivity in a non-intrusive manner is of practical and theoretical importance. Extending research on physiological arousal and the healthy physiological variability thesis to the context of work meetings, we take a novel approach and investigate whether physiological arousal and the variability in implicit body signals of meeting participants (heart rate, arm movements, and speech intensity) can be accurate predictors of perceived meeting productivity. In a preliminary field study, we used smartwatches and tracked the body signals of 16 team members in 26 team meetings. The perceived meeting productivity was assessed at the end of the meetings. Partly supporting our assumptions, multilevel analysis showed that the variance in arm acceleration was a significant predictor of perceived meeting productivity. Further, using a random forest classifier, we accurately predicted perceived meeting productivity in roughly 60% of the cases with body signals. This study adds to previous work on meeting effectiveness by tapping into the potential of wearables to provide valid information about perceived meeting productivity. Cultivating our findings, we discuss lessons learned for future research.

Phonon Sampling Method for Inelastic Thermal Neutron Scattering Events

2026-03-08T03:32:17Z

Phonon Sampling Method for Inelastic Thermal Neutron Scattering Events Trainer, Amelia; Forget, Benoit Accurate representation of thermal neutron scattering in Monte Carlo transport simulations requires that the molecular vibrations of the target material be accounted for. Historically, this has been achieved by precomputing large multidimensional tables that are a function of temperature and the cosine of the scattering angle, as well as incoming and outgoing neutron energy. Most commonly used sampling techniques for thermal neutron scattering rely on large multidimensional tables, where higher resolution results in an increase in required memory and attempts to reduce memory can result in grid coarseness errors. An alternative sampling method is introduced here that is a significant departure from precomputed tables and instead relies on a more physical model of the scattering behavior. The phonon sampling method classifies neutron scattering events by the number of phonons excited/de-excited during the scattering collision. In doing so, energy exchange may be obtained via rejection sampling, and an analytical representation of the momentum exchange is obtained. This sampling method has been tested on graphite, yttrium hydride, and uranium nitride, and preliminary implementation of the phonon sampling method shows accurate results for angular and energy distributions, though resulting in up to a 40% slowdown in overall calculation time. This notable slowdown is countered, however, by a large reduction in storage (over 99% reduction compared to standard multidimensional tables).

Assessment of Engineering Turbulence Models in Buoyant Diabatic Turbulent Flow

2026-03-08T03:32:14Z

Assessment of Engineering Turbulence Models in Buoyant Diabatic Turbulent Flow Wiser, Ralph; Baglietto, Emilio Turbulent heat transfer in buoyancy-dominated flows is a challenging problem for computational fluid dynamics (CFD). Many authors attribute model error in these conditions to the Reynolds analogy. We leverage a brand-new direct numerical simulation database to evaluate the performance of several popular turbulence models in buoyant diabatic channel flow. We find that heat transfer results are relatively accurate, with a Nusselt number error less than 20%. However, the turbulent flow solution is very inaccurate, with wall shear overpredicted by up to 100%. This indicates significant turbulence model error in such flows. We determined that the dominant sources of model error are missing physics in the algebraic Reynolds stress framework and the simple buoyancy production term used in industrial CFD. We suggest that future modeling efforts focus on these two sources of model error. We demonstrate that the Reynolds analogy is not the dominant source of model error.

The insurgent smart city: How a social movement created an alternative imaginary of the smart city

2026-03-08T03:32:16Z

The insurgent smart city: How a social movement created an alternative imaginary of the smart city Stokols, Andrew Urban scholars have critiqued smart cities for their association with neoliberal governance, narrow focus on quantifiable aspects of urban systems, and failure to incorporate citizens’ needs or aspirations. The “smart city” remains a contested concept and as such is subject to reappropriation. Here, I analyze the case of an urban social movement, the 2019–2020 Hong Kong Anti-ELAB protests, as an alternative, “insurgent smart city.” Following from an earlier network analysis of Telegram channels used during the protests, I show how the communications system underpinning much of the protest action simultaneously enabled coordination while also remaining open to grassroots decision-making and innovations of new protest formats as the movement responded to countertactics of the state and police. Telegram channels linked neighborhood-based organizing to the citywide movement. These actions not only emulated but also inverted top-down visions of a total urban information system underpinning many smart city projects. Framing the Hong Kong Anti-ELAB protests as an insurgent smart city offers an alternative sociotechnical imaginary of what smart cities could be, and raises possibilities for an “insurgent digital citizenship” as an alternative to both state and platform-mediated forms of digital citizenship.

The evolution of global cybersecurity norms in the digital age: A longitudinal study of the cybersecurity norm development process

2026-03-08T03:32:13Z

The evolution of global cybersecurity norms in the digital age: A longitudinal study of the cybersecurity norm development process Madnick, Benjamin; Huang, Keman; Madnick, Stuart Developing cybersecurity norms and global normative cybersecurity behaviors play an increasingly critical role in global cybersecurity governance. This paper takes a longitudinal approach to analyze cybersecurity norms development activities during the period 1997–2020. A total of 206 individual cases were collected, and 233 individual cybersecurity norms were identified and compiled into 25 subject categories. Categorizing the norm subjects alongside the frequency of cases and norms identified each year allowed for a longitudinal view of cyber norm activities and the evolution in developments over these years. This examination enables us to categorize cybersecurity norms, including their dynamic focus and evolution patterns. By studying those viewed as “successful,” we gain guidance regarding the construction of global cybersecurity governance in the digital age.

Anthropology Has One Job (On Genocide in the United States)

2026-03-08T03:32:15Z

Anthropology Has One Job (On Genocide in the United States) Lowry, David Shane In an introductory anthropology course, the instructor might provide a definition of anthropology similar to this: “Anthropology is the most scientific of the humanities, and it is the most humanistic of the sciences.” If something like that is said, it stems from a statement in Anthropology, a 1964 book by famed anthropologist Eric Wolf in which he attempted to define the discipline. Wolf’s approach came at a time when many anthropologists were attempting to intervene in the historical telling of the world.Footnote1 In particular, Wolf argued that non-Europeans were also participants in global, colonial processes. The value of Wolf’s voice—indeed, the value of most anthropology at the time—was that it offered a wide-scale view of human events for which the anthropologist was merely an observer, hence not responsible.

Online Bidding under RoS Constraints without Knowing the Value

2025-12-06T03:09:02Z

Online Bidding under RoS Constraints without Knowing the Value Vijayan, Sushant; Feng, Zhe; Padmanabhan, Swati; Shanmugam, Karthikeyan; Suggala, Arun; Wang, Di We consider the problem of bidding in online advertising, where an advertiser aims to maximize value while adhering to budget and Return-on-Spend (RoS) constraints. Unlike prior work that assumes knowledge of the value generated by winning each impression (e.g., conversions), we address the more realistic setting where the advertiser must simultaneously learn the optimal bidding strategy and the value of each impression opportunity. This introduces a challenging exploration-exploitation dilemma: the advertiser must balance exploring different bids to estimate impression values with exploiting current knowledge to bid effectively. To address this, we propose a novel Upper Confidence Bound (UCB)-style algorithm that carefully manages this trade-off. Via a rigorous theoretical analysis, we prove that our algorithm achieves Õ(₲T log(|B|T) ) regret and constraint violation, where T is the number of bidding rounds and B is the domain of possible bids. This establishes the first optimal regret and constraint violation bounds for bidding in the online setting with unknown impression values. Moreover, our algorithm is computationally efficient and simple to implement. We validate our theoretical findings through experiments on synthetic data, demonstrating that our algorithm exhibits strong empirical performance compared to existing approaches. WWW ’25, April 28–May 2, 2025, Sydney, NSW, Australia.

Tactile Vega-Lite: Rapidly Prototyping Tactile Charts with Smart Defaults

2025-12-06T03:09:11Z

Tactile Vega-Lite: Rapidly Prototyping Tactile Charts with Smart Defaults Chen, Mengzhu (Katie); Pedraza Pineros, Isabella; Satyanarayan, Arvind; Zong, Jonathan Tactile charts are essential for conveying data to blind and low vision (BLV) readers but are difficult for designers to construct. Non-expert designers face barriers to entry due to complex guidelines, while experts struggle with fragmented and time-consuming workflows that involve extensive customization. Inspired by formative interviews with expert tactile graphics designers, we created Tactile Vega-Lite (TVL): an extension of Vega-Lite that offers tactile-specific abstractions and synthesizes existing guidelines into a series of smart defaults. Predefined stylistic choices enable non-experts to produce guideline-compliant tactile charts quickly. Expert users can override defaults to tailor customizations for their intended audience. In a user study with 12 tactile graphics creators, we show that Tactile Vega-Lite enhances flexibility and consistency by automating tasks like adjusting spacing and translating braille while accelerating iterations through pre-defined textures and line styles. Through expert critique, we also learn more about tactile chart design best practices and design decisions. CHI ’25, Yokohama, Japan

Toward Everyday Perceptual and Physiological Augmentation

2025-12-06T03:09:17Z

Toward Everyday Perceptual and Physiological Augmentation Tao, Yujie; Gemicioglu, Tan; Chin, Sam; Huang, Bingjian; Brooks, Jas; Follmer, Sean; Lopes, Pedro; Nanayakkara, Suranga Human senses are fundamental to how we interpret and interact with the world. Computing devices are increasingly coupled with the human sensory system through interfaces such as smart glasses, earbuds, and wristbands. This opens up opportunities to dynamically mediate, modify, and augment perceptual experiences and physiological processes through multisensory stimulation. These devices go beyond assistive technologies designed for individuals with sensory impairments (e.g., hearing aids) and are now available for everyday use. Applications range from enriching immersive entertainment experiences to supporting well-being through multisensory interventions. The UIST community has been a key venue for introducing many proof-of-concept prototypes in multisensory stimulation. However, gaps remain in systematically understanding how such technologies can be designed, studied, and contextualized in long-term, everyday use. This workshop will examine barriers to transitioning prototypes from proof-of-concepts into systems for real-world use. The session will feature keynote talks, demo sessions, and an interactive device-swap activity where participants exchange and wear different devices during the afternoon session, and conclude with an open discussion to develop implementation frameworks. UIST Adjunct ’25, Busan, Republic of Korea

GyFoam: Fabricating Lattice Foam with Customizable Stiffness through Uniform Expansion

2025-12-06T03:09:34Z

GyFoam: Fabricating Lattice Foam with Customizable Stiffness through Uniform Expansion Wang, Guanyun; Chen, Haotian; Wang, Yufeng; Li, Songyun; Tao, Yue; Qi, Fanke; Cao, Lizhuo; Jin, Xiao; Tao, Ye; Li, Jiaji We present GyFoam, a fabrication method integrating foam material with lattice structure to enable controlled and uniform expansion, which supports high-quality forming in appearance and customizable stiffness in function, using standard 3D printers, filaments, commercially available Thermo-Expandable Microspheres and silicone. To achieve customizable stiffness, we propose two methods: modifying material concentration and adjusting lattice structural parameters. Additionally, we propose three shape control strategies for creating complex shapes: bending, wavy edges, and internal doming. Furthermore, a user-friendly design tool is established for users to construct lattice structures, preview basic deformation, and generate mold models for printing. Finally, through a series of applications, we validate GyFoam’s practical usage of fabricating large objects, wearable products, enabling flexible interactions and creating aesthetic designs. UIST ’25, Busan, Republic of Korea

EmbroChet: A Hybrid Textile Fabrication Approach for 3D Personalized Handicraft via Heat-Shrinking

2025-12-06T03:09:28Z

EmbroChet: A Hybrid Textile Fabrication Approach for 3D Personalized Handicraft via Heat-Shrinking Wang, Guanyun; Wang, Zhiqi; Li, Fanyu; Liu, Qinyang; Dong, Tianshu; Hong, Zixiang; Li, Xinyi; Zhu, Kuangqi; Li, Jiaji; Zhao, Xiaoliang; Tao, Ye We propose EmbroChet, a hybrid approach that bridges digital fabrication and textile craftsmanship, empowering individuals unfamiliar with intricate craft techniques to design and fabricate 3D textile handicrafts intuitively. EmbroChet allows the creation of handicrafts by embroidering chain stitches (a fundamental embroidery technique) onto a heat-shrinkable film, which subsequently self-transforms from a 2D composite to a 3D textile through a freely controllable heating triggering process. Through a single stitch type, the method enables custom designs and intricate geometries to be achieved without complex manual skills that often requires expertise between different stitch knowledge. To better demonstrate EmbroChet, we propose a design tool that includes shape-changing libraries to assist users in customizing 3D shapes. The evaluation demonstrates its unique strength in balancing geometric complexity and textile softness. Furthermore, our workshop verifies the feasibility of EmbroChet, exploring its potential for personalized textile fabrication, and synergizing the precision of digital fabrication with the tactile artistry of textile craftsmanship. UIST ’25, Busan, Republic of Korea

Meta-antenna: Mechanically Frequency Reconfigurable Metamaterial Antennas

2025-12-06T03:09:24Z

Meta-antenna: Mechanically Frequency Reconfigurable Metamaterial Antennas AlAlawi, Marwa; Zheng, Regina; Ahn, Sooyeon; Yan, Katherine; Sethapakdi, Ticha; Zhu, Junyi; Mueller, Stefanie We introduce Meta-antenna, a design and fabrication pipeline for creating frequency reconfigurable antennas while making use of a single type of mechanical metamaterial structure. Unlike traditional static antenna systems with fixed radiation patterns and frequency responses per geometry, Meta-antenna leverages mechanical reconfiguration to alter the radiation and geometry characteristics of the antenna, making it more versatile for sensing and communication. Meta-antenna provides a design space of resonance frequency from 500 MHz to 6.3 GHz (≥ 10dB) upon the structure’s compression, bending, or rotation. Additionally, we provide an Ansys-based editor that allows users to generate metamaterial antenna geometries and simulate their resonance frequency. We also provide a code template for Meta-antenna based sensing interactions. Our technical evaluation demonstrates that our fabricated Meta-antenna structures remain functional even after 10,000 compression cycles. Finally, we contribute three example applications showcasing Meta-antenna’s potential in adaptive personal devices, smart home systems, and tangible user interfaces. UIST ’25, Busan, Republic of Korea

Computational Tailor-Making for Personalized, Shape-changing, and Sustainable Fabrics

2025-12-06T03:09:20Z

Computational Tailor-Making for Personalized, Shape-changing, and Sustainable Fabrics Narumi, Koya; Hirose, Yuichi; Lee, Hsuanling; Larsson, Maria; He, Liang; Leake, Mackenzie; Forman, Jack; Farahi, Behnaz; Yao, Lining; Igarashi, Takeo Fabrics are fundamental elements of our daily lives, which are woven, knitted, or embroidered into diverse products like clothing and furniture. Recent advances in materials science and digital fabrication have enabled us to fabricate personalized and responsive fabric products computationally and interactively, which we call “computational tailor-making.” In this workshop, we will build an interdisciplinary network of researchers on computational tailor-making and discuss (1) computational fabric design, (2) novel fabric fabrication tools, (3) shape-changing fabrics, and (4) sustainable fabric production, from the viewpoint of HCI. The workshop session will help attendees build a shared vision, recognize potential challenges, find unexpected solutions and ideas, collaborate beyond disciplines, and explore the possible connection to industries. UIST Adjunct ’25, Busan, Republic of Korea

Ori-TENG: 3D Printed Origami Tessellations as Triboelectric Nanogenerators for Self-powered Sensing and Energy Harvesting

2025-12-06T03:09:19Z

Ori-TENG: 3D Printed Origami Tessellations as Triboelectric Nanogenerators for Self-powered Sensing and Energy Harvesting AlAlawi, Marwa; Wang, Kexin; Zheng, Regina; Chan, Adelene; Feick, Martin; Mueller, Stefanie We introduce Ori-TENG, a design and fabrication framework for 3D printed origami tessellations that function as triboelectric sensors and energy harvesters. Ori-TENG structures are 3D printed flat in a single step, then folded, with internal electrical routing optimized for both folding mechanics and triboelectric performance. UIST Adjunct ’25, Busan, Republic of Korea

GreenMix: Energy-Efficient Serverless Computing via Randomized Sketching on Asymmetric Multi-Cores

2025-12-06T03:09:49Z

GreenMix: Energy-Efficient Serverless Computing via Randomized Sketching on Asymmetric Multi-Cores Basu Roy, Rohan; Patel, Tirthak; Li, Baolin; Samsi, Siddharth; Gadepally, Vijay; Tiwari, Devesh GreenMix is motivated by the renewed interest in asymmetric multi-core processors and the emergence of the serverless computing model. Asymmetric multi-cores offer better energy and performance trade-offs by placing different core types on the same die. However, existing serverless scheduling techniques do not leverage these benefits. GreenMix is the first serverless work to reduce energy and serverless keep-alive costs while meeting QoS targets by leveraging asymmetric multi-cores. GreenMix employs randomized sketching, tailored for serverless execution and keep-alive, to perform within 10% of the optimal solution in terms of energy efficiency and keep-alive cost reduction. GreenMix’s effectiveness is demonstrated through evaluations on clusters of ARM big.LITTLE and Intel Alder Lake asymmetric processors. It outperforms competing state-of-the-art schedulers, offering a novel approach for energy-efficient serverless computing. SC ’25, St Louis, MO, USA

Scalable and Low Power Localization for Underwater Robots

2025-12-06T03:09:46Z

Scalable and Low Power Localization for Underwater Robots Afzal, Sayed Saad; Rademacher, Jack; Chen, Weitung; Wang, Purui; Adib, Fadel Localization is a critical task for underwater robots, yet today’s underwater localization systems are limited by their accuracy, scalability, and/or energy consumption (i.e., longevity). We present the design, implementation, and evaluation of EchoBLUE– an accurate, scalable, and low-power localization system for underwater robots. In EchoBLUE, an underwater robot transmits SONARstyle (FMCW) signals, and leverages ultra-low power underwater backscatter nodes as location anchors. EchoBLUE’s design introduces two key innovations. The first is a novel doppler compensation mechanism that enables it to accurately self-localize under mobility: the technique employs a cross-chirp mechanism that exploits the quad-band nature of the resulting backscatter response to overcome the rangedoppler ambiguity. Second, it introduces the first semi-active retrodirective underwater backscatter design and uses it for location anchors; this design achieves wide bandwidth to backscatter the full FMCW signal, enabling fine-grained localization. We implemented a proof of concept prototype of EchoBLUE by building a base station mounted on a BlueROV2 underwater robot and custom-designed low-power retrodirective location anchors deployed in a pool. Our evaluation across 700 real-world trials demonstrates that EchoBLUE achieves a median 3D localization accuracy of 28 cm and 90th percentile of 48 cm. Moreover, these anchors consume only 740 𝜇𝑊 for semi-active backscatter, paving the way for truly low-power and scalable underwater localization.

A polyurethane-urea elastomer at low to extreme strain rates

2025-12-06T03:09:59Z

A polyurethane-urea elastomer at low to extreme strain rates Lee, Jaehee; Veysset, David; Hsieh, Alex J; Rutledge, Gregory C; Cho, Hansohl A finite strain nonlinear constitutive model is presented to study the extreme mechanical behavior of a polyurethane-urea (PUU) well suited for many engineering applications. The micromechanically- and thermodynamically based constitutive model captures salient features in resilience and dissipation in the material from low to extreme strain rates. The extreme deformation features are further elucidated by laser-induced micro-particle impact tests for the material, where an ultrafast strain rate ( > 1 0 6 s−1) incurs. Numerical simulations for the strongly inhomogeneous deformation events are in good agreement with the experimental data, supporting the predictive capabilities of the constitutive model for the extreme deformation features of the PUU material over at least 9 orders of magnitude in strain rates ( 1 0 − 3 to 1 0 6 s−1).

Molecular simulation of flow-enhanced nucleation of polyethylene crystallites in biaxial flows

2025-12-06T03:09:53Z

Molecular simulation of flow-enhanced nucleation of polyethylene crystallites in biaxial flows Gangal, Chinmay S; Rutledge, Gregory C Flow-enhanced nucleation (FEN) of n-pentacontahectane (C150) under biaxial extensional flows of varying strain rate ratios is studied using nonequilibrium molecular dynamics simulation. The nucleation rates thus calculated are used to test previously published FEN models based on invariants of the conformation tensor of Kuhn segments and the extra stress tensor. Models based on the conformation tensor provide a more accurate description of FEN observed in biaxial flow simulations than those based on the extra stress tensor. In addition, the formation of nematic domains previously reported to be stabilized by shear or extensional flow is absent in equibiaxial flows. However, such domains do form in non-equibiaxial flows, and nucleation occurs in these domains preferentially. The shape and orientation of nuclei formed under biaxial flows of various strengths and strain rate ratios are also reported.

Cholesterol Nanofiber Patches with Sustainable Oil Delivery Eliminate Inflammation in Atopic Skin

2025-12-06T03:10:01Z

Cholesterol Nanofiber Patches with Sustainable Oil Delivery Eliminate Inflammation in Atopic Skin Sroczyk, Ewa A; Tarasiuk, Aleksandra; Talar, Marcin; Rutledge, Gregory C; Makaro, Adam; Misztal, Zofia; Wołyniak, Maria; Berniak, Krzysztof; Sałaga, Maciej; Fichna, Jakub; Stachewicz, Urszula Atopic skin is dry and itchy and lacks integrity. Impaired skin barrier results from altered lipid composition of the skin. A crucial skin lipid, cholesterol, provides flexibility and homeostasis of the cell membranes' lipid bilayer. Cholesterol-based creams and natural oils, especially blackcurrant seed oil, are beneficial for skin care as they hydrate the skin and improve its integrity. The major atopic symptom, skin dryness, can be overcome by the application of porous patches enhanced with cholesterol and natural oil. The base of the patches is constructed of polyimide (PI) nanofibers with cholesterol coatings and externally added blackcurrant seed oil. The presence of cholesterol in PI mats hinders the passage of oil through the patches to the skin, resulting in sustained and prolonged skin hydration. The theoretical and numerical investigations of oil dynamics in porous mats confirmed the experimental results, showing a prolonged skin hydration effect up to 6 h. Additionally, as demonstrated by in vivo tests on atopic mice, cholesterol patches lower serum immunoglobulin E levels and expression of proinflammatory cytokines in the skin, thereby accelerating skin healing. Our results hold great promise for the long-term application of the patches in atopic dermatitis treatment.

The Polls and the U.S. Presidential Election in 2020 …. and 2024

2025-12-06T03:09:55Z

The Polls and the U.S. Presidential Election in 2020 …. and 2024 Barnett, Arnold; Sarfati, Arnaud Arguably, the single greatest determinant of U.S. public policy is the identity of the president. And if trusted, polls not only provide forecasts about presidential-election outcomes but can act to shape those outcomes. Looking ahead to the 2024 U.S. presidential election and recognizing that polls before the 2020 presidential election were sharply criticized, we consider whether such harsh assessments are warranted. Initially, we explore whether such polls as processed by the sophisticated aggregator FiveThirtyEight successfully forecast actual 2020 state-by-state outcomes. We evaluate FiveThirtyEight’s forecasts using customized statistical methods not used previously, methods that take account of likely correlations among election outcomes in similar states. We find that, taken together, the pollsters and FiveThirtyEight did an excellent job in predicting who would win in individual states, even those “tipping point” states where forecasting is more difficult. However, we also find that FiveThirtyEight underestimated Donald Trump’s vote shares by state to a modest but statistically significant extent. We further consider how the polls performed when the more primitive aggregator Real Clear Politics combined their results, and then how well single statewide polls performed without aggregation. It emerges that both Real Clear Politics and the individual polls fared surprisingly well.

The Profession’s Vanguards: Arab Architects and Regional Architectural Exchange, 1900–50

2025-12-06T03:09:58Z

The Profession’s Vanguards: Arab Architects and Regional Architectural Exchange, 1900–50 Abusaada, Nadi Writings on architecture in the Middle East during the first half of the twentieth century have often focused on the legacies of colonial architects and planners in shaping Middle Eastern cities and built environments. Contrarily, this article focuses on the overlooked history of the first milieu of trained Arab architects in Middle East, focusing on Palestine, Syria, Lebanon and Egypt. Examining unstudied historical materials and archives, it maps out the trajectories of individual architects as well as the architectural profession more generally in this period of rapid change. It is divided into three main sections that highlight this: first, architecture’s transition from the Ottoman guild system to its professionalisation by the turn of the century; second, the mobility of architectural knowledge and expertise in the Arab region following the First World War; finally, the development of a new institutionalised architectural culture that sought to cultivate bonds between Arab architects not only in their individual countries, but also regionally throughout the Arab world towards the mid-twentieth century.

Fossil fuel divestment and public climate change policy preferences: an experimental test in three countries

2025-12-06T03:10:03Z

Fossil fuel divestment and public climate change policy preferences: an experimental test in three countries Schwartz, Joshua A.; Lendway, Paul; Nuri, Abolfazl Divestment is a prominent strategy championed by activists to induce positive social change. For example, the current fossil fuel divestment movement includes over 1,500 institutions that control $40 trillion in assets. A primary pathway through which divestment is theorized to be effective is by influencing public beliefs and policy preferences, thus pressuring policymakers to take action. However, prior research only tests this argument via qualitative case studies. We assess the impact of exposure to information about fossil fuel divestment on public opinion through the use of national survey experiments in three major greenhouse gas emitters: the U.S., India, and South Africa. We find surprisingly little evidence that exposure to information about the fossil fuel divestment movement can increase public support for policies that address climate change. Our findings suggest that divestment movements may be less effective at changing beliefs and policy preferences than previously realized.

“Sculptress Interprets Land’s Spirit”: Elizabeth Wyn Wood, the Group of Seven, and analogy as equivalence

2025-12-06T03:09:56Z

“Sculptress Interprets Land’s Spirit”: Elizabeth Wyn Wood, the Group of Seven, and analogy as equivalence Nikčević, Hana Canadian sculptor Elizabeth Wyn Wood (1903–66), best known for her modernist landscape sculptures, has since the inception of her artistic career been compared, through analogy, with the Group of Seven (fl. 1920–33), Canada’s enduringly famous and overtly nationalistic collective of modernist landscape painters. Critics claimed that Wood “achieved for sculpture what the Group of Seven achieved for painting” and, occasionally, invoked specific Group artists, dubbing Wood the “Lawren Harris of sculpture.” Analogizing across disciplines, the Wood/Group likening appears to posit a formal comparison in gendered language: the Group’s bold, decorative portrayals of the northern Ontario “wilderness” find clear visual comparands in Wood’s abstracted compositions of the same region. In this article, however, I demonstrate that the apparently visual basis for the comparison is inextricable from the textual discourse fundamental to Canadian art in the early twentieth century and beyond; it is only through analyzing this discourse that an understanding of the Wood/Group analogy can be reached. The Group ostensibly pioneered the first genuine Canadian landscape aesthetic; through immersing himself in the land, the mythology went, the Canadian artist learned to paint Canada on its own terms. This landscape artist-as-woodsman myth was a form of settler indigenization by which Canada laid cultural claim to colonized land. Analogy frames Wood as not an epigone but an equal of the Group: in producing organically, anew, a genuine Canadian landscape aesthetic for sculpture, Wood “achieved for sculpture what the Group of Seven achieved for painting”—its deployment as a medium in the service of Canada’s land claim.

A Landscape “Difficult to Describe”: The Model Village and the Capital City

2025-12-06T03:09:51Z

A Landscape “Difficult to Describe”: The Model Village and the Capital City Springstubb, Phoebe In mid-twentieth-century Punjab, grassroots development projects sought to modernize the countryside by decentralizing power to villages. The capital city Chandigarh, built in the same period, seems to represent the opposite: a national symbol of a newly independent India’s centralized power. Yet, this article argues, rural and urban were reciprocal and volatile counterparts. Through the work of M.S. Randhawa, it reorients analysis of Chandigarh to reveal how the materiality of landscape itself was a medium for territorial planning, indelibly linking—and managing the distinctions between—city and countryside. A botanist and civil servant, Randhawa used landscape to realize modernizing agendas and to constrain social change in projects from model villages and a “bioaesthetic” plan for the city to new land-grant universities that ushered in the Green Revolution’s industrialized agriculture. His work offers a revisionist history of development’s practitioners and periodization. It shows how an uneven fabric of late-colonial rural uplift shaped the contours of postcolonial, state-directed agrarian transformation. Following the civil servant in the landscape, this article calls for the grounding of abstract theories like development and state formation in histories of their local inflections.

Policy Search through Genetic Programming and LLM-assisted Curriculum Learning

2025-12-05T04:15:40Z

Policy Search through Genetic Programming and LLM-assisted Curriculum Learning Jorgensen, Steven; Nadizar, Giorgia; Pietropolli, Gloria; Manzoni, Luca; Medvet, Eric; O'Reilly, Una-May; Hemberg, Erik Curriculum learning (CL) consists in using a diverse set of user-provided test cases, with varying levels of difficulty and organized in a suitable progression, for learning a policy. The quality of test cases is important to allow optimization techniques as genetic programming (GP) to solve policy search problems. In this work, we evaluate large language models (LLMs) as providers of test cases for GP-based policy search. We consider two policy search tasks, a single-player and a multi-player game, and four LLMs differing in complexity and specialization, which we prompt in order to generate suitable test cases for the two games. We experimentally assess the intrinsic quality of LLM-generated test cases and their utility when inserted in a curriculum consumed by a GP optimization. We evaluate the robustness of the approach with respect to the way cases are scheduled in curricula and with respect to the policy representation, for which we use both graphs and linear programs evolved by GP. We observe that the effectiveness of LLM-assisted CL depends on both the choice of LLM and the design of the prompting and scheduling strategies. These findings highlight important considerations for leveraging LLMs in automated curriculum design for GP-based optimization.

Robust Biharmonic Skinning Using Geometric Fields

2025-12-05T04:15:49Z

Robust Biharmonic Skinning Using Geometric Fields Dodik, Ana; Sitzmann, Vincent; Solomon, Justin; Stein, Oded Bounded bihramonic weights are a popular tool used to rig and deform characters for animation, to compute reduced-order simulations, and to define feature descriptors for geometry processing. They necessitate tetrahedralizing the volume bounded by the surface, introducing the possibility of meshing artifacts or tetrahedralization failure. We introduce a mesh-free and robust automatic skinning technique that generates weights comparable to the current state of the art, but works reliably even on open surfaces, triangle soups, and point clouds where current methods fail. We achieve this through the use of a specialized Lagrangian representation enabled by the advent of hardware ray-tracing, which circumvents the need for finite elements while optimizing the biharmonic energy and enforcing boundary conditions. The flexibility of our formulation allows us to integrate artistic control through weight painting during the optimization. We offer a thorough qualitative and quantitative evaluation of our method.

SquareLoop: Explore Optimal Authentication Block Strategy for ML

2025-12-05T04:15:46Z

SquareLoop: Explore Optimal Authentication Block Strategy for ML Strzeszynski, Jan; Tong, Jianming; Lee, Kyungmi; Xiong, Nathan; Parashar, Angshuman; Emer, Joel; Krishna, Tushar; Yan, Mengjia Off-chip memory in ML accelerators is vulnerable to both hardware and software attack, which needs encryption and authentication. Precise performance modeling of it requires (1) representation of authentication blocks (AuthBlock) to cover the full design space of shapes and orientations, and (2) precise memory behavior modeling, as encryption and authentication mainly increase memory traffic. This paper introduces 𝑆 2Loop, a framework that resolves these challenges by introducing (1) flexible, all-level partitioning based AuthBlocks for ensuring full coverage of the entire design space, (2) a realistic layout-based memory model, and (3) an Mapping-LayoutAuthentication co-search algorithm to explore the drastic combinatorial design space to figure out optimal mapping, layout, and AuthBlock shape choice for multi-layer workloads. SquareLoop’s detailed memory model helps find better mapping to achieve 1.32× speedup on ResNet18 compared to the SotA SecureLoop, and our latency predictions are validated to within 7.3% of an RTL implementation. 𝑆 2𝐿𝑜𝑜𝑝 also achieve up-to 1.08×/1.82× overall speedup for authenticated ResNet18/MobileNet-V3 on various accelerators with AuthBlock and Mapping co-searching. We open-source 𝑆 2Loop to provide a powerful and validated tool for designing efficient, secure accelerators at https://github.com/maeri-project/squareloop. HASP 2025, Seoul, Republic of Korea

A Close Look at RMP Entry Caching and Its Security Implications in SEV-SNP

2025-12-05T04:15:44Z

A Close Look at RMP Entry Caching and Its Security Implications in SEV-SNP Bagia, Alexis; Ulitzsch, Vincent; Trujillo, Dani?l; Li, Mengyuan; Yan, Mengjia; Seifert, Jean-Pierre AMD’s Secure Encrypted Virtualization (SEV) technology is a pivotal component in AMD server processors that boosts cloud computing security. It achieves this by offering transparent memory encryption and managing keys for protecting virtual machines (VMs), independently of the hypervisor’s trustworthiness. The latest iteration, SEV-Secure Nested Paging (SEV-SNP), introduces memory integrity protection through a data structure called the Reverse Map Table (RMP), which maps system physical addresses to guest physical addresses and tracks ownership of physical pages. The RMP is maintained in a dedicated region in DRAM. As every memory write triggers a check against an RMP entry, caching RMP entries is crucial to alleviating the RMP’s performance impact. However, caching may create new security challenges, as it can introduce new microarchitectural side-channels. In addition, maintaining cache coherence is crucial for the RMP’s security guarantees. However, so far, neither the details of the RMP’s caching behavior nor its security implications have been explored. This paper aims to fill this gap by conducting a systematic study of the RMP’s caching behavior. Through reverse engineering, we identify that the RMP is not only cached in the TLB, but also in the L1D and L2 data cache. Interestingly, this caching depends on the access type on Zen 5. We also uncover the mechanisms by which cache coherence across the TLB is enforced. We find that each update to the RMP table triggers a global TLB flush across all cores. Finally, we present several potential security implications and demonstrate that an attacker can exploit RMP’s caching to leak physical address information. A user process can leak 6 bits of the Physical Frame Number (PFN) of its pages via the L1D cache within 2.5 µs per page, with success rates of 97 % (Zen 4) and 99 % (Zen 3 and Zen 5).

Guarding LLM-aided Software Transformation Tasks via Component Exoskeletons

2025-12-05T04:15:47Z

Guarding LLM-aided Software Transformation Tasks via Component Exoskeletons Lamprou, Evangelos; Kalhauge, Christian; Rinard, Martin; Vasilakis, Nikos Large language models (LLMs) are achieving state-of-the-art results across a wide variety of software transformation tasks---including translating across languages and lifting opaque software components to high-level languages. Unfortunately, their results are often subtly incorrect, insecure, or underperformant---affecting the widespread deployment of these LLM-driven techniques in settings that go beyond the narrow scope of academic papers. This paper posits that such widespread deployment crucially depends on developing appropriate model guardrails for safeguarding the results of the transformation process. Such guardrails can be supported by component exoskeletons, tunable partial specifications extracted mostly automatically from the original, pre-transformed component. Exoskeletons serve as component projections that supplement, and often go through, the entire transformation process, confirming that the new, transformed component meets the original specifications. They show promise on several real-world scenarios and unearth exciting research directions. PACMI ’25, October 13-16, 2025, Seoul, Republic of Korea

The Continuous Tensor Abstraction: Where Indices Are Real

2025-12-05T04:15:43Z

The Continuous Tensor Abstraction: Where Indices Are Real Won, Jaeyeon; Ahrens, Willow; Collin, Teodoro Fields; Emer, Joel S.; Amarasinghe, Saman This paper introduces the continuous tensor abstraction, allowing indices to take real-number values (e.g., A[3.14]). It also presents continuous tensor algebra expressions, such as Cx,y = Ax,y ∗ Bx,y, where indices are defined over a continuous domain. This work expands the traditional tensor model to include continuous tensors. Our implementation supports piecewise-constant tensors, on which infinite domains can be processed in finite time. We also introduce a new tensor format for efficient storage and a code generation technique for automatic kernel generation. For the first time, our abstraction expresses domains like computational geometry and computer graphics in the language of tensor programming. Our approach demonstrates competitive or better performance to hand-optimized kernels in leading libraries across diverse applications. Compared to hand-implemented libraries on a CPU, our compiler-based implementation achieves an average speedup of 9.20× on 2D radius search with ∼60× fewer lines of code (LoC), 1.22× on genomic interval overlapping queries (with ∼18× LoC saving), and 1.69× on trilinear interpolation in Neural Radiance Field (with ∼6× LoC saving).

A Domain-Specific Probabilistic Programming Language for Reasoning about Reasoning (Or: A Memo on memo)

2025-12-05T04:15:21Z

A Domain-Specific Probabilistic Programming Language for Reasoning about Reasoning (Or: A Memo on memo) Chandra, Kartik; Chen, Tony; Tenenbaum, Joshua B.; Ragan-Kelley, Jonathan The human ability to think about thinking ("theory of mind") is a fundamental object of study in many disciplines. In recent decades, researchers across these disciplines have converged on a rich computational paradigm for modeling theory of mind, grounded in recursive probabilistic reasoning. However, practitioners often find programming in this paradigm challenging: first, because thinking-about-thinking is confusing for programmers, and second, because models are slow to run. This paper presents memo, a new domain-specific probabilistic programming language that overcomes these challenges: first, by providing specialized syntax and semantics for theory of mind, and second, by taking a unique approach to inference that scales well on modern hardware via array programming. memo enables practitioners to write dramatically faster models with much less code, and has already been adopted by several research groups.

Pyrosome: Verified Compilation for Modular Metatheory

2025-12-05T04:15:31Z

Pyrosome: Verified Compilation for Modular Metatheory Jamner, Dustin; Kammer, Gabriel; Nag, Ritam; Chlipala, Adam We present Pyrosome, a generic framework for modular language metatheory that embodies a novel approach to extensible semantics and compilation, implemented in Coq. Common techniques for semantic reasoning are often tied to the specific structures of the languages and compilers that they support. Contextual equivalence is difficult to work with directly, and both logical relations and transition system-based approaches typically fix a specific notion of effect globally. While modular transition systems have been effective in imperative settings, they are suboptimal for functional code. These limitations restrict the extension and composition of semantics in these systems. In Pyrosome, verified compilers are fully extensible, meaning that to extend a language simply requires defining and verifying the compilation of the new feature, reusing the old correctness theorem for all other cases. The novel enabling idea is an inductive formulation of equivalence preservation that supports the addition of new rules to the source language, target language, and compiler. Pyrosome defines a formal, deeply embedded notion of programming languages with semantics given by dependently sorted equational theories, so all compiler-correctness proofs boil down to type-checking and equational reasoning. We support vertical composition of any compilers expressed in our framework in addition to feature extension. Since our design requires compilers to support open programs, our correctness guarantees support linking with any target code of the appropriate type. As a case study, we present a multipass compiler from System F with simple references, through CPS translation and closure conversion. Specifically, we demonstrate how we can build such a compiler incrementally by starting with a compiler for simply typed lambda-calculus and adding natural numbers, the unit type, recursive functions, and a global heap, then extending judgments with a type environment and adding type abstraction, all while reusing the original theorems. We also present a linear version of the simply typed CPS pass and compile a small imperative language to the simply typed target to show how Pyrosome handles substructural typing and imperative features.

What You See Is What It Does: A Structural Pattern for Legible Software

2025-12-05T04:15:32Z

What You See Is What It Does: A Structural Pattern for Legible Software Meng, Eagon; Jackson, Daniel The opportunities offered by LLM coders (and their current limitations) demand a reevaluation of how software is structured. Software today is often “illegible”—lacking a direct correspondence between code and observed behavior—and insufficiently modular, leading to a failure of three key requirements of robust coding: incrementality (the ability to deliver small increments by making localized changes), integrity (avoiding breaking prior increments) and transparency (making clear what has changed at build time, and what actions have happened at runtime). A new structural pattern offers improved legibility and modularity. Its elements are concepts and synchronizations: fully independent services and event-based rules that mediate between them. A domain-specific language for synchronizations allows behavioral features to be expressed in a granular and declarative way (and thus readily generated by an LLM). A case study of the RealWorld benchmark is used to illustrate and evaluate the approach. Onward! ’25, Singapore, Singapore

Gauguin, Descartes, Bayes: A Diurnal Golem’s Brain

2025-12-05T04:15:29Z

Gauguin, Descartes, Bayes: A Diurnal Golem’s Brain Chandra, Kartik; Liu, Amanda; Ragan-Kelley, Jonathan; Tenenbaum, Joshua B. A "quine" is a deterministic program that prints itself. In this essay, I will show you a "gauguine": a probabilistic program that infers itself. A gauguine is repeatedly asked to guess its own source code. Initially, its chances of guessing correctly are of course minuscule. But as the gauguine observes more and more of its own previous guesses, it detects patterns of behavior and gains information about its inner workings. This information allows it to bootstrap self-knowledge, and ultimately discover its own source code. We will discuss how—and why—we might write a gauguine, and what we stand to learn by constructing one. Onward! ’25, Singapore, Singapore

Low-Fidelity vs. High-Fidelity Spatial Design in Virtual Reality for Non-professionals

2025-12-05T04:15:37Z

Low-Fidelity vs. High-Fidelity Spatial Design in Virtual Reality for Non-professionals Wei, Lan; Dai, Chenyue; Peng, Xuening; Tong, Xin; Liu, Can In spatial design, non-professionals lack effective hands-on opportunities to participate in the design process. Although VR platforms can support spatial design with immersive interaction, existing tools simply provide high-fidelity 3D objects for users to choose and place around. Low-fidelity design approach is rarely supported, nor investigated in this context. In this work, we present a user study comparing low-fidelity and high-fidelity spatial design in VR. Eighteen participants were recruited to use both versions of a prototype with varied geometric fidelity to complete home designs. Their design outcome and intent was evaluated by professional designers. Our findings show, the low-fidelity version allowed participants to think more openly and creatively, leading to a more holistic expression of their design intent and needs, while the high-fidelity version promoted users’ thinking of realistic scenarios. We discuss the design implications and how they can be combined in co-design activities. CHCHI 2024, Shenzhen, China

Trans Data: A Research and Design Agenda from Trans Activists' Transformative Data Science

2025-12-05T04:15:38Z

Trans Data: A Research and Design Agenda from Trans Activists' Transformative Data Science Stevens, Nikko; D'Ignazio, Catherine; Doğan, Amelia Trans activists play a deeply important role in caring for and advocating for the transgender community using data. Through an interview study with 16 trans activists working in trans-led and trans-serving organizations in the United States, we document how they use restorative/transformative data science processes of resolving, researching, recording, and refusing and using data. We incorporate their data practices with trans technology and trans competent interaction design approaches to propose a research agenda for trans data: materially improve trans lives, cross data boundaries, and constantly engage in power analysis. We expound on how a trans data research agenda can benefit data advocacy and CSCW research and design.

Techno-economic analysis and life cycle assessment for catalytic fast pyrolysis of mixed plastic waste

2025-12-05T04:15:58Z

Techno-economic analysis and life cycle assessment for catalytic fast pyrolysis of mixed plastic waste Yadav, Geetanjali; Singh, Avantika; Dutta, Abhijit; Uekert, Taylor; DesVeaux, Jason S; Nicholson, Scott R; Tan, Eric CD; Mukarakate, Calvin; Schaidle, Joshua A; Wrasman, Cody J; Carpenter, Alberta C; Baldwin, Robert M; Román-Leshkov, Yuriy; Beckham, Gregg T yrolysis of waste plastics has gained interest as a candidate chemical recycling technology. To examine the potential of this approach, we conducted a techno-economic analysis (TEA) and life cycle assessment (LCA) of a conceptual catalytic fast pyrolysis (CFP) facility that converts 240 metric tons/day of mixed plastic waste. The modeled base case predicts the minimum selling price (MSP) of a benzene, toluene, and xylenes (BTX) mixture at $1.07 per kg when co-products are sold at their average market prices. We predict that the aromatic product stream can be cost-competitive with virgin BTX mixtures ($0.68/kg) if the mixed waste plastics are available for less than $0.10/kg or if crude oil prices exceed $60/barrel. Moreover, we estimate that CFP-based conversion of waste plastics can reduce the total supply chain energy use by 24% but with a 2.4-fold increase in greenhouse gas (GHG) emissions per kilogram of BTX, relative to incumbent manufacturing process. Sensitivity analysis highlights that feedstock cost, co-product selling prices, capital cost for product separations, and operating costs are key cost drivers. Further, we examine three additional CFP processes that differ in product composition, namely naphtha, and a case where the products are rich in either C2–C4 olefins or BTX aromatic hydrocarbons. Whereas the MSP of naphtha ($2.18/kg) is ∼4-fold higher than virgin naphtha, both the olefin-rich and aromatics-rich product cases exhibit a potential reduction in MSP up to 40%, with a 21%–45% reduction in total supply chain energy and 2.2–3.8-fold increase in GHG emissions relative to incumbent manufacturing processes. LCA predicts that the CFP process exhibits lower fossil fuel depletion than virgin manufacturing across all cases as well as lower acidification, ozone depletion, and smog formation for select cases, but high utility and feedstock preparation requirements result in poorer performance across other metrics. Overall, this study highlights important process parameters for improving CFP of mixed waste plastics from economic and environmental perspectives.

Togedule: Scheduling Meetings with Large Language Models and Adaptive Representations of Group Availability

2025-12-05T04:15:34Z

Togedule: Scheduling Meetings with Large Language Models and Adaptive Representations of Group Availability Song, Jaeyoon; Ashktorab, Zahra; Malone, Thomas Scheduling is a perennial-and often challenging-problem for many groups. Existing tools are mostly static, showing an identical set of choices to everyone, regardless of the current status of attendees' inputs and preferences. In this paper, we propose Togedule, an adaptive scheduling tool that uses large language models to dynamically adjust the pool of choices and their presentation format. With the initial prototype, we conducted a formative study (N=10) and identified the potential benefits and risks of such an adaptive scheduling tool. Then, after enhancing the system, we conducted two controlled experiments, one each for attendees and organizers (total N=66). For each experiment, we compared scheduling with verbal messages, shared calendars, or Togedule. Results show that Togedule significantly reduces the cognitive load of attendees indicating their availability and improves the speed and quality of the decisions made by organizers.

Core‐passivation: A concept for stable core‐shell nanoparticles in aqueous electrocatalysis

2025-12-04T03:14:19Z

Core‐passivation: A concept for stable core‐shell nanoparticles in aqueous electrocatalysis Göhl, Daniel; Paciok, Paul; Wang, Zhenshu; Kang, Jin Soo; Heggen, Marc; Mayrhofer, Karl JJ; Román‐Leshkov, Yuriy; Ledendecker, Marc The stability of nanoparticles is a major challenge in thermal and electrocatalysis. This is especially true for core‐shell nanoparticles where only a few monolayers of noble metal protect the usually non‐noble core material. In this work, we utilize the practical nobility concept to engineer stable core‐shell nanoparticles with a self‐passivating core material. Specifically, tantalum carbide as core material in combination with a 1–3 monolayer thick platinum shell exhibits exceptional stability in aqueous media. The core‐shell catalyst shows no sign of structural changes after 10,000 degradation cycles up to 1.0 VRHE. Due to the efficient passivation of tantalum carbide at the solid/liquid interface, the dissolution reduces by a factor of eight compared to bare Pt. Our findings confirm that passivating core materials are highly beneficial for the stabilization of core‐shell nanomaterials in aqueous media. They open up new ways for the rational design of cost‐efficient but stable non‐noble core – platinum shell nanoparticles where harsh, oxidizing conditions are employed.

Interdependence of Solvent and Catalyst Selection on Low Pressure Hydrogen-Free Reductive Catalytic Fractionation

2025-12-04T03:14:28Z

Interdependence of Solvent and Catalyst Selection on Low Pressure Hydrogen-Free Reductive Catalytic Fractionation Facas, Gregory G; Brandner, David G; Bussard, Jeremy R; Román-Leshkov, Yuriy; Beckham, Gregg T Hydrogen-free reductive catalytic fractionation (RCF) is a promising method to produce aromatic compounds directly from native biomass without the use of external hydrogen gas. In this work, we show that by using high boiling point diols as a solvent in hydrogen-free RCF, reaction pressures can be reduced by an order of magnitude compared to conventional RCF with methanol and hydrogen gas, while still producing appreciable aromatic monomer yields. Importantly, the use of diols with secondary alcohol functional groups increases hydrogenation activity on Ru/C, Pt/C, and Ni/C, measured by the yield of aromatic compounds with saturated propyl side chains, compared to processing in ethylene glycol, indicating that the choice of solvent and catalyst together can be tuned to control product selectivity of aromatic monomers in RCF.

Propylene Metathesis over Molybdenum Silicate Microspheres with Dispersed Active Sites

2025-12-04T03:14:25Z

Propylene Metathesis over Molybdenum Silicate Microspheres with Dispersed Active Sites Skoda, David; Zhu, Ran; Hanulikova, Barbora; Styskalik, Ales; Vykoukal, Vit; Machac, Petr; Simonikova, Lucie; Kuritka, Ivo; Poleunis, Claude; Debecker, Damien P; Román-Leshkov, Yuriy In this work, we demonstrate that amorphous and porous molybdenum silicate microspheres are highly active catalysts for heterogeneous propylene metathesis. Homogeneous molybdenum silicate microspheres and aluminum-doped molybdenum silicate microspheres were synthesized via a nonaqueous condensation of a hybrid molybdenum biphenyldicarboxylate-based precursor solution with (3-aminopropyl)triethoxysilane. The as-prepared hybrid metallosilicate products were calcined at 500 °C to obtain amorphous and porous molybdenum silicate and aluminum-doped molybdenum silicate microspheres with highly dispersed molybdate species inserted into the silicate matrix. These catalysts contain mainly highly dispersed MoOx species, which possess high catalytic activity in heterogeneous propylene metathesis to ethylene and butene. Compared to conventional silica-supported MoOx catalysts prepared via incipient wetness impregnation (MoIWI), the microspheres with low Mo content (1.5–3.6 wt %) exhibited nearly 2 orders of magnitude higher steady-state propylene metathesis rates at 200 °C, approaching site time yields of 0.11 s–1.

Accessing monomers from lignin through carbon–carbon bond cleavage

2025-12-04T03:14:20Z

Accessing monomers from lignin through carbon–carbon bond cleavage Palumbo, Chad T; Ouellette, Erik T; Zhu, Jie; Román-Leshkov, Yuriy; Stahl, Shannon S; Beckham, Gregg T Lignin, the heterogeneous aromatic macromolecule found in the cell walls of vascular plants, is an abundant feedstock for the production of biochemicals and biofuels. Many valorization schemes rely on lignin depolymerization, with decades of research focused on accessing monomers through C–O bond cleavage, given the abundance of β–O–4 bonds in lignin and the large number of available C–O bond cleavage strategies. Monomer yields are, however, invariably lower than desired, owing to the presence of recalcitrant C–C bonds whose selective cleavage remains a major challenge in catalysis. In this Review, we highlight lignin C–C cleavage reactions, including those of linkages arising from biosynthesis (β–1, β–5, β–β and 5–5) and industrial processing (5–CH2–5 and α–5). We examine multiple approaches to C–C cleavage, including homogeneous and heterogeneous catalysis, photocatalysis and biocatalysis, to identify promising strategies for further research and provide guidelines for definitive measurements of lignin C–C bond cleavage.

Direct propylene epoxidation via water activation over Pd-Pt electrocatalysts

2025-12-04T03:14:16Z

Direct propylene epoxidation via water activation over Pd-Pt electrocatalysts Chung, Minju; Maalouf, Joseph H; Adams, Jason S; Jiang, Chenyu; Román-Leshkov, Yuriy; Manthiram, Karthish Direct electrochemical propylene epoxidation by means of water-oxidation intermediates presents a sustainable alternative to existing routes that involve hazardous chlorine or peroxide reagents. We report an oxidized palladium-platinum alloy catalyst (PdPtOx/C), which reaches a Faradaic efficiency of 66 ± 5% toward propylene epoxidation at 50 milliamperes per square centimeter at ambient temperature and pressure. Embedding platinum into the palladium oxide crystal structure stabilized oxidized platinum species, resulting in improved catalyst performance. The reaction kinetics suggest that epoxidation on PdPtOx/C proceeds through electrophilic attack by metal-bound peroxo intermediates. This work demonstrates an effective strategy for selective electrochemical oxygen-atom transfer from water, without mediators, for diverse oxygenation reactions.

Interaction Configurations and Prompt Guidance in Conversational AI for Question Answering in Human-AI Teams

2025-12-04T03:14:02Z

Interaction Configurations and Prompt Guidance in Conversational AI for Question Answering in Human-AI Teams Song, Jaeyoon; Ashktorab, Zahra; Pan, Qian; Dugan, Casey; Geyer, Werner; Malone, Thomas Understanding the dynamics of human-AI interaction in question answering is crucial for enhancing collaborative efficiency. Extending from our initial formative study, which revealed challenges in human utilization of conversational AI support, we designed two configurations for prompt guidance: a Nudging approach, where the AI suggests potential responses for human agents, and a Highlight strategy, emphasizing crucial parts of reference documents to aid human responses. Through two controlled experiments, the first involving 31 participants and the second involving 106 participants, we compared these configurations against traditional human-only approaches, both with and without AI assistance. Our findings suggest that effective human-AI collaboration can enhance response quality, though merely combining human and AI efforts does not ensure improved outcomes. In particular, the Nudging configuration was shown to help improve the quality of the output when compared to AI alone. This paper delves into the development of these prompt guidance paradigms, offering insights for refining human-AI collaborations in conversational question-answering contexts and contributing to a broader understanding of human perceptions and expectations in AI partnerships.

Pushing on an Open Door: Japan’s Evolutionary Security Posture

2025-12-04T03:14:27Z

Pushing on an Open Door: Japan’s Evolutionary Security Posture Heginbotham, Eric; Leiter, Samuel; Samuels, Richard J At the 2022 Shangri-La Dialogue, Japan’s Prime Minister Fumio Kishida warned defense ministers from across the Indo-Pacific region that “Ukraine today may be East Asia tomorrow.” Russia’s war of aggression and China’s tacit support for the invasion have amplified the urgency of the threat posed by China’s economic and military rise and have informed material changes to Japanese defense policy.

Dissecting User Experience of Social Virtual Reality: A Tale of Five Platforms

2025-12-04T03:14:05Z

Dissecting User Experience of Social Virtual Reality: A Tale of Five Platforms Cheng, Ruizhi; Li, Jie; Chen, Songqing; Han, Bo Social virtual reality (VR) has the potential to replace conventional online social media by offering quasi-realworld social experiences. As such, it has been extensively examined by the research community. However, existing studies fall short of providing a comprehensive understanding of how different aspects of social VR platforms interact to affect user experience. Motivated by this limitation, we conduct a user study with Oculus Quest 2 headsets and dissect the user experience on five social VR platforms. We evenly and randomly divide 42 participants into short-term (spending 10–30 minutes/platform) and long-term (spending at least 120 minutes/platform) groups. Besides employing surveys and interviews, we measure the frame rate and resolution of these platforms and explore how various factors interplay to influence the user experience of social VR. Our findings reveal that the frame rate, resolution, and interactive events of social VR platforms have a more significant impact on the experience of long-term users compared to short-term users. The scalability limitations of these platforms, as evidenced by decreased frame rates with the increasing number of concurrent users, result in an increased prevalence of motion sickness among long-term users, negatively impacting their overall experience. Moreover, the absence of highly interactive events also deteriorates their overall experience, and the low resolution combined with the lack of interactive events further decreases their sense of social presence. Additionally, our study demonstrates several common limitations negatively affecting the experience of both long-term and short-term users. For example, the harassment prevention mechanisms on all five platforms are inadequate, and being harassed has a detrimental effect on users’ overall experience and sense of social presence. The avatar embodiment of investigated platforms has limited contribution to users’ sense of social presence, mainly due to the lack of realism and full-body tracking. Our findings call for more research in scalability support, motion sickness relief, interactive event design, harassment prevention, and avatar development for improving social VR platforms in the future.

Investigation of critical heat flux enhancement on nanoengineered surfaces in pressurized subcooled flow boiling using infrared thermometry

2025-12-04T03:14:24Z

Investigation of critical heat flux enhancement on nanoengineered surfaces in pressurized subcooled flow boiling using infrared thermometry Wang, Chi; Su, Guanyu; Akinsulire, Olorunsola; Zhang, Limiao; Rahman, Md Mahamudur; Bucci, Matteo Enhancing the flow boiling critical heat flux (CHF) is beneficial to the economics and safety margins of many industrial applications cooled by boiling heat transfer. While many studies have shown that surfaces with hydrophilic nanoscale and micro-scale features can enhance CHF in pool boiling, it is still not clear how these engineered surfaces affect the CHF in subcooled flow boiling at ambient pressure, let alone high-pressure conditions. Here, two nano-engineered surfaces, i.e., a surface coated with a porous layer of hydrophilic silica nanoparticles and a surface coated with zinc oxide nanowires, were tested. Flow boiling tests with a 10 K subcooling and a mass flux of 1000 kg/(m2·s) were conducted at 1 bar and 4 bars using infrared thermometry diagnostics. At 1 bar, the CHF enhancement is around 15% for both coatings. At 4 bars, the CHF enhancement is around 17% for the nanowire surface, and around 25% for the nano-porous surface. Infrared thermometry measurements reveal that the CHF enhancement comes from an increase of both two-phase heat transfer and single-phase heat transfer mechanisms, which is due to a change of bubble dynamics on the nanoengineered surfaces. It is also shown that the boiling crisis can be predicted using a percolation model based on Monte Carlo (MC) simulations.

Connecting Digitalization and Sustainability: Proptech in the Real Estate Operations and Management

2025-12-04T03:14:31Z

Connecting Digitalization and Sustainability: Proptech in the Real Estate Operations and Management Tan, Zhengzhen; Miller, Norm G. Digitalization of building operations and maintenance enable real-time monitoring, optimization, and automation for environment sustainability. Proptech startups are important change agents in accelerating building digitalization. While many researchers analyze economic and environmental savings from deployment of digital technology, far less attention has been devoted to challenges for proptech startups to transform efficiency gains into viable businesses. We analyze the Unissu global proptech startup database to reveal the scope and competitive landscape of proptech solutions. We conduct interviews with building owners/operators to understand what impedes the adoption of proptech solutions. Despite rapid growth, ongoing challenges remain for sustainability-focused proptech firms with three adoption barriers: (1) integration of the technology stacks; (2) integration of technology stacks with business processes; and (3) integration of owner/operators and the occupants’ solutions. Proptech with applications that work with existing infrastructure or provide more complete holistic solutions with extensive capital reserves, are more likely to survive. Other pathways include having data standardization and security protocols in place; technology partnership with technology incumbents; and effective communication with owners/operators to fill the knowledge gap. Findings can provide insights to emerging digital proptech startups as they spearhead market adoption in the real estate sector and monetize the sustainability value creation.

Driver response and recovery following automation initiated disengagement in real-world hands-free driving

2025-12-04T03:14:22Z

Driver response and recovery following automation initiated disengagement in real-world hands-free driving Gershon, Pnina; Mehler, Bruce; Reimer, Bryan Objective Advanced driver assistance systems are increasingly available in consumer vehicles, making the study of drivers’ behavioral adaptation and the impact of automation beneficial for driving safety. Concerns over driver’s being out-of-the-loop, coupled with known limitations of automation, has led research to focus on time-critical, system-initiated disengagements. This study used real-world data to assess drivers’ response to, and recovery from, automation-initiated disengagements by quantifying changes in visual attention, vehicle control, and time to steady-state behaviors. Methods Fourteen drivers drove for one month each a Cadillac CT6 equipped with Super Cruise (SC), a partial automation system that, when engaged, enables hands-free driving. The vehicles were instrumented with data acquisition systems recording driving kinematics, automation use, GPS, and video. The dataset included 265 SC-initiated disengagements identified across 5,514 miles driven with SC. Results Linear quantile mixed-effects models of glance behavior indicated that following SC-initiated disengagement, the proportions of glances to the Road decreased (Q50Before=0.91, Q50After=0.69; Q85Before=1.0, Q85After=0.79), the proportions of glances to the Instrument Cluster increased (Q50Before=0.14, Q50After=0.25; Q85Before=0.34, Q85After=0.45), and mean glance duration to the Road decreased by 4.86 sec in Q85. Multinomial logistic regression mixed-models of glance distributions indicated that the number of transitions between glance locations following disengagement increased by 43% and that glances were distributed across fewer locations. When driving hands-free, take over time was significantly longer (2.4 sec) compared to when driving with at least one hand on the steering wheel (1.8 sec). Analysis of moment-to-moment distributional properties of visual attention and steering wheel control following disengagement indicated that on average it took drivers 6.1 sec to start the recovery of glance behavior to the Road and 1.5 sec for trend-stationary proportions of at least one hand on the steering wheel. Conclusions Automation-initiated disengagements triggered substantial changes in driver glance behavior including shorter on-road glances and frequent transitions between Road and Instrument Cluster glance locations. This information seeking behavior may capture drivers’ search for information related to the disengagement or the automation state and is likely shaped by the automation design. The study findings can inform the design of more effective driver-centric information displays for smoother transitions and faster recovery.

Lessons in Sanctions-Proofing from Russia

2025-12-04T03:14:33Z

Lessons in Sanctions-Proofing from Russia Glenn, Caileigh Government actors and other observers across Europe and the United States called the multilateral sanctions imposed on Russia in early 2022 “unprecedented.”Footnote1 Even Russian President Vladimir Putin acknowledged their severity when he stressed “the need to counter economic restrictions that were imposed on us, which are truly unprecedented without any exaggeration.”Footnote2 Part of the response to the Russian invasion of Ukraine, these financial and trade sanctions—imposed on Russia by Western governments—target key firms in the financial and energy sectors, debt financing, technology, Russia’s foreign currency reserves, and more recently, most Russian oil and transportation insurers.

Why Haven’t We Applied the Lessons from Lean to Innovation?

2025-12-04T03:14:32Z

Why Haven’t We Applied the Lessons from Lean to Innovation? Wright, Randall S. Yes, I know. People have been doingLean innovation—increasing efficiencyby capturing customer feedback earlyand often and minimizing waste in theproduct development cycle—for the last10 years.I’m not talking about applying Leanprinciples to innovation. I’m talkingabout how American business leadershad the humility to admit their firmsneeded to learn Lean from Japaneseculture to master globally competitiveoperations, and why they need now tolearn innovation from the culture ofuniversities to master globally compet-itive innovation.

Becoming Infrastructure: A Critical Realist Account of the Evolution of DHIS2 as Digital Public Health Infrastructure in Sierra Leone

2025-12-04T03:14:04Z

Becoming Infrastructure: A Critical Realist Account of the Evolution of DHIS2 as Digital Public Health Infrastructure in Sierra Leone Ndubuisi-Obi, Innocent; Chen, Nuole; Tsai, Lily Today, the District Health Information System 2 (DHIS2) has become the de-facto standard for open-source health management information systems and Sierra Leone's status as the first country in sub-Saharan Africa to implement DHSI2 makes it a productive place for researchers interested in understanding the end-to-end process of infrastructuring in a low-resource bureaucratic setting. In this article, we examine its design, implementation, and maintenance in Sierra Leone over a period of 14 years - from 2008 to 2022. We present an intensive case study discretized by three morphogenetic cycles (decentralization, centralization, and fragmentation) and furnished with explanatory account's of DHIS2's evolution using a critical realist research methodology to describe the emergence of DHIS2 as digital public health infrastructure. These accounts highlight the structural and cultural systems of DHIS2, their elaborations, and their interaction with agents over successive periods of DHIS2's evolution. Our study finds that, despite its continued use in Sierra Leone, the increasing generativity in the structural and cultural systems of DHSI2 and Sierra Leone’s public health system engenders a persistent instability that requires continuous resolution. Though we find that extant literature aids in our understanding of DHIS2’s evolution, we proffer two mechanisms, infrastructural capture and socio-technical debt, which aid our explanation of events observed in our case study. Our work makes a case for more ontologically-diverse theorizing of bureaucracy-aware computing systems.

Factorization in additive monoids of evaluation polynomial semirings

2025-12-04T03:14:18Z

Factorization in additive monoids of evaluation polynomial semirings Ajran, Khalid; Bringas, Juliet; Li, Bangzheng; Singer, Easton; Tirador, Marcos For a positive real α, we can consider the additive submonoid M of the real line that is generated by the nonnegative powers of α. When α is transcendental, M is a unique factorization monoid. However, when α is algebraic, M may not be atomic, and even when M is atomic, it may contain elements having more than one factorization (i.e., decomposition as a sum of irreducibles). The main purpose of this paper is to study the phenomenon of multiple factorizations inside M. When α is algebraic but not rational, the arithmetic of factorizations in M is highly interesting and complex. In order to arrive to that conclusion, we investigate various factorization invariants of M, including the sets of lengths, sets of Betti elements, and catenary degrees. Our investigation gives continuity to recent studies carried out by Chapman et al. in 2020 and by Correa-Morris and Gotti in 2022.

The Argentella scandal: why French officials did not make Corsica a nuclear test site in 1960

2025-12-04T03:14:29Z

The Argentella scandal: why French officials did not make Corsica a nuclear test site in 1960 Cooper, Austin R. Top French officials made plans in early 1960 to transform an abandoned silver mine in Corsica, called the Argentella Massif, into an underground site for nuclear explosions. By June 1960, they had canceled these plans. This article shows how a mass movement on the Mediterranean island forced their hand, and it explains why Corsicans of diverse political affiliations took to the streets. The Argentella project—and the health, environmental, and strategic risks that it entailed—looked in Corsica like evidence that Paris saw the islanders as second-class citizens, even residents of an internal colony. French police intelligence, which maintained surveillance on the Corsican anti-nuclear movement, feared that this movement might have drawn inspiration from the contemporaneous struggle for national liberation in Algeria, where French nuclear explosions began. The Argentella protests illustrated national disagreements about French nuclear ambitions that previous scholarship, proposing official consensus, has minimized. They show how, in a nuclear-armed democracy, local officials, political activists, and ordinary citizens can shape nuclear-weapons policy. But Corsican anti-nuclear action in 1960 did not demand disarmament. These protests also illuminate a longer trajectory in French nuclear history, which involved atmospheric explosions in colonized territories in Algeria and Polynesia until the 1970s, despite local and international resistance.

Towards Foundation Model for Spatiotemporal Data Analysis

2025-12-04T03:13:57Z

Towards Foundation Model for Spatiotemporal Data Analysis Wu, Yuankai; Chen, Xinyu; Zhuang, Dingyi Spatiotemporal data modeling has long been a fundamental task across disciplines such as climate & environmental science, and transportation engineering. A typical goal is to estimate unknown information at specific spatiotemporal points based on partially observed data—for example, interpolating weather conditions at unmeasured locations, reconstructing missing historical records, or forecasting the future trajectories of financial markets. These are all core tasks within the broader scope of spatiotemporal modeling. This tutorial (1 hours) introduces a cohesive view of spatiotemporal data modeling, tracing the evolution from traditional statistical approaches to modern deep learning paradigms. We begin by revisiting Kriging and time series decomposition to highlight the essential assumptions and strengths of these classical methods. Next, we explore low-rank matrix and tensor completion techniques, which leverage the structured patterns of spatiotemporal data. We then elaborate on spatiotemporal graph neural networks, which characterize complex dependencies by integrating graph structures with dynamic temporal features. Finally, we discuss recent advances in applying large foundation models to spatiotemporal tasks, including their capabilities and current limitations. Throughout the tutorial, we emphasize how lessons from traditional methods—such as the importance of locality, periodicity, and smoothness priors—can inspire new directions for developing and fine-tuning foundation models in the spatiotemporal domain. We conclude by outlining key challenges and opportunities in bridging classical wisdom with emerging AI capabilities. SSTD ’25, Osaka, Japan

Gender gaps in South Korea’s labour market: children explain most of the gender employment gap, but little of the gender wage gap

2025-12-04T03:14:15Z

Gender gaps in South Korea’s labour market: children explain most of the gender employment gap, but little of the gender wage gap Stansbury, Anna; Kirkegaard, Jacob Funk; Dynan, Karen South Korea’s gender wage and employment gaps are among the largest in the OECD. Using labour force survey data over 2010–19, we estimate gender wage and employment gaps, and child earnings penalties, for women aged 25–54. We show (i) that the large gender gaps in South Korea’s labour market are mostly not a function of differential sorting by gender along education, occupation, or industry lines, (ii) that caring for children (and, perhaps increasingly, for the elderly) is the major factor inhibiting women’s labour force participation, and (iii) that large gender wage gaps exist even for women without care responsibilities. These findings suggest that improving opportunities for work–family balance is crucial to helping increase women’s labour force participation, but may do little to close gender wage gaps: other major obstacles also appear to stand in the way of Korean women’s full inclusion in the labour force.

Madman or Mad Genius? The International Benefits and Domestic Costs of the Madman Strategy

2025-12-04T03:14:13Z

Madman or Mad Genius? The International Benefits and Domestic Costs of the Madman Strategy Schwartz, Joshua A. According to the “Madman Theory” outlined by Daniel Ellsberg and Thomas C. Schelling, and embraced by Presidents Richard Nixon and Donald Trump, being perceived as mad can help make seemingly incredible threats—such as starting a nuclear war—more credible. However, recent research has largely concluded that the Madman Theory does not work. In this study, I theorize that the international benefits of the Madman Theory have been underestimated, but also that there are significant domestic barriers associated with adopting such a strategy that undermine its effectiveness. Through a series of five novel survey experiments, I find evidence that perceived madness provides limited advantages in coercive bargaining vis-à-vis foreign adversaries, but it also entails significant domestic costs that potentially erode its efficacy. Overall, this study provides clearer support for the Madman Theory than most previous literature has found, but also breaks new theoretical ground by analyzing the domestic politics of perceived madness.

Purrfect Pitch: Exploring Pitch Interval Learning through an Audio-Haptic Interface

2025-12-04T03:13:59Z

Purrfect Pitch: Exploring Pitch Interval Learning through an Audio-Haptic Interface Chin, Sam; Fang, Cathy Mengying; Singh, Nikhil; Ibrahim, Ibrahim; Paradiso, Joe; Maes, Pattie We introduce Purrfect Pitch, a system consisting of a wearable haptic device and a custom-designed learning interface for musical ear training. We focus on the ability to identify pitch intervals (sequences of two musical notes), a perceptually ambiguous task that usually requires rote training. With our system, users hear two tones while simultaneously receiving two corresponding vibrotactile stimuli on the back. Providing haptic feedback on the back makes the auditory distance between tones salient, and the back-worn design is comfortable and unobtrusive. During training, users receive multi-sensory feedback from our system and input their guessed interval value on our web-based learning interface. Our study with 18 participants shows that our system enables novice learners to identify intervals more accurately and consistently than those who only received audio feedback, even after removing the haptic feedback. We also share further insights on designing a multisensory learning system. AHs 2025, Masdar City, Abu Dhabi, United Arab Emirates

Teaching AI to Feel: A Collaborative, Full-Body Exploration of Emotive Communication

2025-12-04T03:13:56Z

Teaching AI to Feel: A Collaborative, Full-Body Exploration of Emotive Communication Lemus, Lissette; Pilcher, Kris; Sprengel, Holger; Sabater-Mir, Jordi; Tütüncü, Esen K. Commonaiverse is an interactive installation exploring human emotions through full-body motion tracking and real-time AI feedback. Participants engage in three phases: Teaching, Exploration and the Cosmos Phase, collaboratively expressing and interpreting emotions with the system. The installation integrates MoveNet for precise motion tracking and a multi-recommender AI system to analyze emotional states dynamically, responding with adaptive audiovisual outputs. By shifting from top-down emotion classification to participant-driven, culturally diverse definitions, we highlight new pathways for inclusive, ethical affective computing. We discuss how this collaborative, out-of-the-box approach pushes multimedia research beyond single-user facial analysis toward a more embodied, co-created paradigm of emotional AI. Furthermore, we reflect on how this reimagined framework fosters user agency, reduces bias, and opens avenues for advanced interactive applications. MM ’25, October 27–31, 2025, Dublin, Ireland

Personalized Animations for Affective Feedback: Generative AI Helps to Visualize Skin Conductance

2025-12-04T03:13:49Z

Personalized Animations for Affective Feedback: Generative AI Helps to Visualize Skin Conductance Scheirer, Jocelyn; Picard, Rosalind; Cantrell, Aubrey Biofeedback interfaces traditionally rely on abstract visualizations, tones, or haptics to convey physiological states—but these often lack personal relevance, emotional salience, and engagement. In this paper, we present a novel system that bridges wearable sensing and generative AI to create real-time, personalized animated biofeedback experiences. Users describe emotionally meaningful objects or scenes to a language model in our system, which outputs generate customized Processing animations. These animations are then dynamically driven by electrodermal activity (EDA) signals from a wrist sensor. We co-design and evaluate the system with autistic adults, many of whom have unique “special interests” that are likely to engage them more than a one-sized-fits-all visualization. Many of these individuals also have difficulty with interoception -- feeling or sensing their own internal and physiological state changes. We built this tool to transform passive physiological monitoring into an interactive multimedia experience, where the visual representation of the body is authored by the user. We introduce a prompt-engineered GPT-based interface that streamlines code generation, sensor mapping, and iterative refinement, requiring no prior coding expertise. The technical pipeline we built includes signal filtering, dynamic parameter mapping, and natural language-based customization— delivering a real-time, visually immersive feedback loop. We report on initial case studies with 12 autistic adults using the system, which highlight both the expressive potential and individual variability of user responses, reinforcing the need for adaptable multimedia frameworks in health technologies. By merging realtime physiological data with generative animation and natural language interaction, this work expands the creative frontier of personalized affective biofeedback. We also address ethical challenges arising from using AI with physiological sensors. MRAC '25, October 27–28, 2025, Dublin, Ireland

Hands-on Strategies for Teaching Social and Societal Impacts of Computing

2025-12-04T03:13:47Z

Hands-on Strategies for Teaching Social and Societal Impacts of Computing Kurkovsky, Stan; Nnamani, Manee Ngozi; Hunter, Aaron; Sobomehin, Olatunde; Braught, Grant; Goldweber, Michael The topic of hands-on strategies for teaching the social and societal impacts of computing is of growing interest to the computer science education community because it addresses a critical gap in traditional CS curricula [7]. While technical skills remain central, educators increasingly recognize the need to prepare students for the ethical, social, and human-centered challenges posed by modern computing technologies. From AI-driven decision-making to digital accessibility and data privacy, computing profoundly affects individuals and communities, making it essential for students to engage with these issues through experiential learning [12]. Different viewpoints on this topic emerge based on pedagogical approaches, disciplinary perspectives, and technological optimism or skepticism. Some educators advocate for integrating servicelearning and community-based projects, arguing that real-world engagement fosters empathy and ethical awareness. Others emphasize case studies and simulations, providing structured exposure to societal challenges without the unpredictability of external partnerships. Additionally, viewpoints may diverge on the role of AI: while some see AI tools as an opportunity to enhance social good, others worry they may exacerbate biases and reduce human agency in computing. Despite these differences, there is broad agreement that computing education must go beyond technical training to include a deeper understanding of computing’s role in society. CompEd 2025, October 21–25, 2025, Gaborone, Botswana

Interactive Storybooks for Early AI Literacy

2025-12-04T03:09:39Z

Interactive Storybooks for Early AI Literacy Pu, Isabella As artificial intelligence (AI) becomes increasingly present in children's everyday environments, there is an urgent need for developmentally appropriate tools that help young learners understand and shape these technologies. To be effective, these tools must not only successfully convey complex concepts but also engage children in ways that are meaningful, accessible, and fun. This thesis introduces the Interactive Storybooks for Early AI Literacy, a series of ten interactive storybooks for children ages 6–9 that combine narrative, mini-games, and scaffolded creative AI interactions to teach core AI and robotics concepts. The storybooks follow an overarching narrative featuring a friendly robot, Doodlebot, who must learn creative tasks with the child's help, framing the child as an AI designer and introducing them to the concept of training AI models through the narrative. The storybooks additionally contain interactive games and activities which help keep kids excited and engaged, while providing structured opportunities to experiment with and explore AI creation tools. First, a pilot study was conducted at a community summer camp with four Interactive Storybooks. Children expressed joy and pride in their AI creations, used the characters as emotional anchors for learning, and began to successfully articulate key AI concepts. Four engagement archetypes emerged: the Reader, the Gamer, the Showcaser, and the Social Connector, each representing a distinct way children interacted with the storybooks. However, despite behavioral signs of engagement, many children described the narrative portions as boring and claimed to prefer games. To explore this tension, a home deployment study compared two versions of the system: a "Books" condition with the full narrative and a "Games" condition with only instructional text. Both conditions included the same mini-games and AI interactions. While children in both groups reported similar levels of enjoyment, those in the Books condition showed significantly higher learning gains, greater increases in perceived knowledge and confidence, and stronger connections to the characters. Children in the Books condition also more frequently referenced the narrative when describing AI concepts and demonstrated more creative and iterative behavior during and after gameplay. Overall, these findings suggest that combining storytelling, gameplay, and creative AI interactions is an effective and engaging approach to teaching AI and robotics to young children. Narrative context appears to support concept recall, deepen emotional investment, and promote thoughtful experimentation, even with complex concepts for this age group, like AI and robotics. Based on insights from both studies, this thesis concludes with six design recommendations for creating developmentally appropriate, emotionally resonant AI education tools for early learners using narrative and play.

Decentralized Machine Learning over Fragmented Data

2025-12-04T03:07:38Z

Decentralized Machine Learning over Fragmented Data Singh, Abhishek The remarkable scaling of data and computation has unlocked unprecedented capabilities in text and image generation, raising the question: Why hasn’t healthcare seen similar breakthroughs? This disparity stems primarily from healthcare data being fragmented across thousands of institutions, each safeguarding patient records in regulatory-compliant silos. The problem is not limited to healthcare but extends to other industries with fragmented data across institutions and individuals. Instead of centralizing various datasets to solve the fragmentation problem, which raises regulatory and ethical concerns, this thesis proposes systems and algorithms to decentralize the machine learning pipeline. Current approaches in this area have centered around Federated Learning (FL), which enables model training over distributed data. However, FL’s dependence on central coordination and inflexibility with heterogeneous systems limit its applicability in healthcare settings. Motivated by these challenges, I explore the following three core themes: 1) Coordination – Today’s coordination algorithms typically rely on static rules or randomized communication, approaches that turn out to be sub-optimal when data heterogeneity is high. I present a new system and a benchmark framework that enables systematic assessment of different coordination algorithms. Next, I propose an adaptive coordination algorithm that leverages historical performance and learning dynamics to improve coordination. 2) Heterogeneity – Data owners can vary significantly in their data distributions, computational resources, and privacy requirements. To address this heterogeneity, I turn the focus from the traditionally protected training phase to securing the critical inference process. Next, I develop techniques for distributed training that adapt to heterogeneous computational capabilities across different agents. 3) Scalability – Enabling scaling in decentralized ML requires addressing three key challenges: parallelization, synchronization, and self-scaling. While parallelization has advanced significantly, the other two remain challenging. I present a framework for offline collaboration through sanitized, synthetic datasets that eliminates constant synchronization needs while preserving privacy. This thesis identifies and addresses some of the bottlenecks along these three core themes through a complementary set of solutions: adaptive coordination, heterogeneity-aware training, and scalable collaboration. Together, these building blocks can enable a practical framework for unlocking data silos across institutions.

Interplay between spatial structure and competition in ecological communities

2025-12-04T03:07:23Z

Interplay between spatial structure and competition in ecological communities Swartz, Daniel W. Ecology, much like physics, has a long history of theoretical contribution. In this thesis, we take a physics approach to describing ecological communities, searching for simple, emergent features that can generalize beyond specific models of community dynamics. Unifying all of the models we study is an underlying spatial structure, leading to a richer set of possible behaviors than a typical well-mixed model. We first study the case of a metapopulation, a collection of smaller communities linked by dispersal. We find that when the environment is allowed to fluctuate stochastically, new growth laws emerge at the single species level, and high diversity is achieved in the case with many species. We then study the case of pathogen evolution, again in the metapopulation framework. We find that intermediate dispersal can act as a strong driver of pathogen evolution. We also study what happens as a population of microbes expands into unexplored territory, known as a range expansion. We find that a simple model can capture all morphological phases observed in experiments and predict invasion fitness as a function of local and global competitive ability. We also break a standard assumption in microbial ecology, the isotropy of space, and find that a new sector morphology emerges.

The Algorithmic Cookbook of Quantum Science: Quantum and Classical Recipes for Computation

2025-12-04T03:07:14Z

The Algorithmic Cookbook of Quantum Science: Quantum and Classical Recipes for Computation Martyn, John Michael Since the dawn of science, computation and physics have evolved alongside each other, both driven by a shared quest to solve problems and calculate properties of the natural world. Today, this symbiotic relationship is epitomized in quantum information science, which proposes to use quantum mechanics to solve hard computational problems and develop new paradigms of communication and cryptography. Yet often absent from these developments is a clear, human-interpretable understanding, with many quantum protocols built from inherently quantum concepts (e.g., entanglement, superposition) that defy our classical line of thought and muddle the search for efficient quantum algorithms. Here we show that this search need not be so opaque: simple mathematical tools, namely polynomials and their fundamental theorems, in unison with concepts from classical computing, provide a powerful framework for the design of quantum algorithms. We develop this framework and use it to construct an assortment of quantum algorithms, including methods for quantum simulation, parallel computing, randomized algorithms, and continuousvariable quantum hardware. In illuminating this framework, we find a striking bidirectional flow: just as classical concepts inspire new quantum algorithms, so too can quantum mechanical insights bring about novel methods of classical computing. In this reverse direction, we adopt inherently quantum concepts, such as random compilation and bosonic symmetry, to develop new classical methods, with applications in simulating quantum systems and designing robust neural networks. In aggregate, this thesis provides a compendium of algorithmic techniques for probing quantum systems and solving hard problems, using both quantum and classical tools—an “algorithmic cookbook”—predicated on deep connections between these two domains. The recipes presented here aim to demystify black boxes of quantum information science, and provide a valuable resource for future developments.

The Death of Quasiparticles: Strongly Interacting Gapless Phases with Fermi Surfaces and Fractional Statistics

2025-12-04T03:07:02Z

The Death of Quasiparticles: Strongly Interacting Gapless Phases with Fermi Surfaces and Fractional Statistics Shi, Zhengyan The emergence of quasiparticles at low temperature provides a powerful organizing principle for many quantum phases of matter, ranging from conventional magnets and superconductors to exotic insulators with topological order. In this thesis, I describe my research in gapless quantum phases in which the framework of quasiparticles breaks down. The main characters are two categories of gapless phases that feature the interplay between strong interactions and two additional ingredients – Fermi surfaces and fractional statistics. Chapter 2 through Chapter 5 focus on strongly interacting metals with Fermi surfaces. The most salient examples are a class of Hertz-Millis models describing the onset of spontaneous symmetry breaking in a metallic environment. At the quantum critical point, gapless order parameter fluctuations destroy quasiparticles living on the Fermi surface, giving rise to a strongly coupled non-Fermi liquid metal. A key result of these chapters is the identification of an infinite-dimensional symmetry that survives in these non-Fermi liquid metals despite the death of quasiparticles. This infinite-dimensional symmetry and its quantum anomaly lead to a series of non-perturbative results on thermodynamics and transport, which are confirmed by perturbative diagrammatic calculations in special examples. Chapter 6 through Chapter 8 explore quantum phases in which anyonic quasiparticles with fractional statistics play an essential role. When parameters in the system are tuned to close the anyon energy gaps, the original anyons lose their coherence and a variety of novel phases emerge. A highlight in this direction is a new mechanism for topological superconductivity in itinerant abelian and non-abelian anyon fluids, which could make contact with experiments on doped fractional quantum anomalous Hall states in the near future.

Particles Inside Particles: The Flow of Energy in Quarks, Gluons, and Jets

2025-12-04T03:07:34Z

Particles Inside Particles: The Flow of Energy in Quarks, Gluons, and Jets Alipour-fard, Samuel This thesis presents the author’s work in developing probes of the inner structure of jets in high-energy particle collisions. We begin by introducing QCD and the scattering of partons (quarks and gluons), discussing jets as theoretical and experimental proxies for partonic physics, and presenting the partonic cascade model of jet formation and jet substructure. Noting the ubiquitous presence of low-energy pollution in particle collision events, in the forms of hadronization, detector effects, the underlying event (UE), and pileup (PU), we then move towards the modern research area of developing pollution-insensitive probes of jet substructure. Pollution-insensitive features of jet substructure are often accessed theoretically either through jet grooming or energyweighted correlation functions. We present the basics of the modern theory of jet grooming as well as the work of the author in developing the Piranha paradigm for continuous jet grooming, introduced by the author in Ref. [1], and explore the formal and phenomenological benefits of continuous grooming techniques as pollutioninsensitive probes of jet substructure. We introduce the basics of the simplest energy-weighted correlation function – the energy-energy correlator (EEC), which probes angular correlations between particle pairs – and discuss its multi-particle analogues. We focus on the efficient and visually intuitive projected and resolved energy correlators introduced by the author in Ref. [2], which provide computationally-realistic, pollution-insensitive probes of angular many-body correlations in QCD jets. Finally, we exposit the generic theory of energy-weighted observable correlations (EWOCs), introduced by the author in Ref. [3], which utilizes the energy weighting of the EEC to provide pollution-insensitive probes of non-angular correlations within jets.

Solid-state cavity quantum electrodynamics with spin ensembles

2025-12-04T03:07:27Z

Solid-state cavity quantum electrodynamics with spin ensembles Wang, Hanfeng Quantum sensors have the potential to operate at fundamental physical performance limits. Among various quantum sensing platforms, solid-state spin emitters stand out due to advantageous characteristics such as room-temperature spin polarization and readout, atomic-scale spatial resolution, and extended coherence times. Despite these strengths, traditional optical detection methods exhibit low readout fidelity in solid-state ensembles, severely limiting their achievable sensitivity. This thesis addresses this limitation by coupling a solid-state emitter ensemble to a microwave cavity, forming a cavity quantum electrodynamics system. Our approach eliminates the need for photon collection required by conventional optical readout methods, and the resulting strongly coupled system allows efficient cavity-based probing of the solid-state spin ensemble. By exploiting the hybrid quantum system with cavity quantum electrodynamics, we achieve record-high sensitivity for solid-state quantum sensors, representing a substantial advancement toward achieving fundamental sensing limits.

Expanding the Phase Space of Photons in Matter: From High-Throughput Screening to Atom-by-Atom Engineering

2025-12-04T03:07:29Z

Expanding the Phase Space of Photons in Matter: From High-Throughput Screening to Atom-by-Atom Engineering Ghorashi, Ali Focusing on the topological band properties of photonic crystals and the plasmonic properties of two-dimensional metals, we seek to answer the question: what is the phase space of photons in matter? For topology, what are the physical parameters that determine whether a given photonic crystal band hosts Dirac points, a non-zero Chern number, or topologically protected corner states? And for plasmons, what are the experimentally addressable ranges of plasmonic dispersions, phase velocities, confinements, and losses? In particular, is it possible to engineer the elusive lossless plasmon? Using high-throughput screening, artificial intelligence, and atom-by-atom engineering through density functional theory, we determine the topological prevalence of photonic bands, propose two systems that evade plasmonic losses through the electron-phonon interaction, and (re)discover general physical laws that govern the geometries of photonic eigenstates.

On the Sample Efficiency of Data-Driven Decision Making

2025-12-04T03:07:25Z

On the Sample Efficiency of Data-Driven Decision Making Qian, Jian This thesis studies the fundamental problem of decision making under uncertainty through the lens of statistical decision theory. We characterize the minimax risk, which captures the sample efficiency required for effective decision making across three key settings: offline estimation with batch data, online estimation with sequential data, and interactive decision making as exemplified by multi-armed bandits and reinforcement learning. The first part of the thesis develops novel algorithmic and theoretical tools to enhance decision making in these regimes and to bridge the gaps between them. We revisit logistic regression in the offline setting and provide guarantees without restrictive boundedness assumptions. We then propose meta-algorithms that reduce online estimation to offline estimation, enabling any offline estimator to be used effectively in online scenarios. Furthermore, we present general-purpose algorithms for interactive decision making problems by leveraging offline or online estimation techniques. The second part of the thesis introduces a unified approach to understanding the fundamental complexity of interactive decision making. We propose the Decision Making with Structured Observation (DMSO) framework, which encompasses bandits, reinforcement learning, and more general settings. Within this framework, we develop a new complexity measure—the Decision-Estimation Coefficient (DEC)—which captures both upper and lower bounds for minimax regret. DEC extends classical notions such as the modulus of continuity to interactive scenarios by introducing an adaptive variant of Le Cam’s method. Finally, we unify the three classical lower bound techniques—Le Cam’s method, Assouad’s lemma, and Fano’s inequality—through a generalized formulation that also incorporates the DEC, offering a comprehensive understanding of the minimax risk in decision making tasks.

Towards achieving power autonomy in soft-actuated micro aerial robots

2025-12-04T03:07:20Z

Towards achieving power autonomy in soft-actuated micro aerial robots Ren, Zhijian Micro aerial robots with insect-like flight capabilities hold immense promise for various applications, including environmental monitoring, precision agriculture, and infrastructure inspection in confined spaces. However, realizing power autonomy in these miniature robotic platforms presents significant challenges due to weight constraints, power density limitations, and inefficient actuation at small scales. This dissertation presents three essential improvements towards achieving power autonomy in soft-actuated micro aerial robots. Our robotic platform is driven by a dielectric elastomer actuator (DEA) and generates lift force through flapping wings, a similar mechanism found in flying insects. First, we implemented a dynamic model to optimize the robot components for pairing with an improved DEA to generate a higher lift force. The robot achieved a peak lift-to-weight ratio of 4.3 and demonstrated a 20-second hovering flight with position and attitude errors smaller than 2.5 cm and 2◦ . Second, we fabricated a lightweight high-voltage boost converter that transformed a 7 V DC input into an AC waveform of 600 V and 400 Hz to drive the actuator. This is the first onboard boost converter that can drive the soft-actuated micro aerial robot to take off, and it represents a substantial achievement in miniaturizing power electronics for microrobots. Third, we took inspiration from the natural autorotation of maple seeds in their slow descent. We implemented the first samara-inspired mechanism on micro aerial robots, enhancing lift generation while maintaining in-flight attitude stability without feedback control. The 1.22-gram vehicle can stably take off in 1 second with a total input thrust of 1 gram-force. These accomplishments provide a pathway towards achieving power autonomy and open opportunities for developing agile, robust, and autonomous micro aerial robots for diverse applications.

Low-Energy Electron-Photon Interactions in a Scanning Electron Microscope

2025-12-04T03:07:07Z

Low-Energy Electron-Photon Interactions in a Scanning Electron Microscope Simonaitis, John The interaction of free-electrons with matter and light is among the most fundamental of processes in nature. From the use of free-electrons for atomic imaging, to their use in the generation of high-intensity, tunable light in synchrotrons, the physics of unconfined electrons has wide application. In recent years, there has been a new focus on the quantum nature of individual electrons in electron microscopes to enable further improvements in these technologies. This work takes advantage of developments in ultrafast optics, electron spectroscopy, quantum optics, and nanofabrication to explore various electron-electron, electron-photon, and electron-material interactions. In this thesis, we construct a low-energy, ultrafast scanning electron microscope, using it to explore quantum coherent interactions between electrons, light, and matter. In Chapter 1, we review the history of free electron experiments and how advances in nanofabrication, low-dimensional materials, and ultrafast optics have opened new opportunities for electron-light interactions to a degree not previously possible. In Chapter 2 we discuss experimental forms of quantum electron microscopy known as interaction-free measurement and electron multi-passing. Chapter 3 details a general theory of electron-photon interactions, including simulations with quantum two-level systems and extended optical nanostructures. In Chapter 4, we design and construct a second microscope with ultrafast triggering, an electron spectrometer with sub-eV resolution, nanostructured interaction regions, and active beam alignment. Chapter 5 explores various experimental results, demonstrating enhanced loss spectroscopy of 2D materials, energy resolution of gold nanoparticle plasmons, as well as spectroscopy of time-tagged cathodoluminescence from optical fibers. Finally, in chapter 6 we discuss future perspectives of this approach, analyzing the impact a heralded electron source would have on electron microscopy and lithography.

Studies of Jet Modification in Heavy Ion Collisions with the CMS Experiment

2025-12-04T03:07:09Z

Studies of Jet Modification in Heavy Ion Collisions with the CMS Experiment Park, Mary Isabelle In the Compact Muon Solenoid (CMS) detector at the Large Hadron Collider (LHC), lead ions are collided at ultra-relativistic velocities to produce Quark-Gluon Plasma (QGP), a state of matter where quarks and gluons are deconfined and move collectively. Jets are produced in high-momentum transfer parton scatterings prior to and independently of QGP formation, and serve as natural probes of its properties. As the high-energy partons pass through the QGP, they lose energy through medium-induced gluon radiation and elastic scattering, resulting in jets that are modified with respect to the vacuum baseline. In this thesis, jet modification is quantified by measuring the jet production cross section as a function of jet radius in inclusive jets and the jet axis decorrelation in jets recoiling from isolated photons in Lead-Lead (PbPb) and Proton-Proton (pp) collisions. Both measurements indicate that effects of medium-induced jet broadening may be balanced by survivor bias in PbPb collisions, potentially due to differences in the magnitude of quenching of wide versus narrow jets. The results underline the importance of constraining the initial jet kinematics with bosons, which are unmodified by the QGP.

Understanding Drivers of Stratospheric Ozone Change and Fingerprinting its Recovery

2025-12-24T03:27:16Z

Understanding Drivers of Stratospheric Ozone Change and Fingerprinting its Recovery Wang, Peidong Stratospheric ozone serves as Earth’s natural protective layer, shielding the surface from harmful ultraviolet radiation. The discovery of the Antarctic ozone “hole” in the late 1980s raised significant societal and scientific concern, prompting the rapid regulation of ozonedepleting substances (ODSs) under international treaties. While the signs of ozone recovery have begun, new challenges continue to arise. This thesis investigates three critical factors driving stratospheric ozone changes and influencing the detection of ozone recovery: (1) ODS emissions, (2) chemical chlorine processes, and (3) internal climate variability. With ODS emissions being regulated under the Montreal Protocol and studies now focusing on illicit new production on the order of tens of gigagrams per year, the ocean’s role as both a natural source and sink of ODSs becomes increasingly important. However, these processes have often been overlooked or highly simplified in past ozone assessments. Using a hierarchy of models, from simple box models to global ocean general circulation models, I quantified the ocean’s uptake and release of various ODSs. Chapter 2 examines the ocean’s uptake of chlorofluorocarbons (CFCs), particularly emphasizing its influence on recent illicit CFC emissions estimation. Chapter 3 extends this analysis to include ocean uptake and potential microbial degradation processes, evaluating their effects on emission estimates for various hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs), which are chemical constituents that have been used to replace CFCs. Once these man-made ODSs reach the stratosphere, they are photolyzed to chlorine reservoir species (e.g., HCl and ClONO2), which, through heterogeneous reactions, can transform into reactive chlorine that depletes ozone. While heterogeneous chlorine activation on volcanic ash is well understood, the unprecedented 2020 Australian wildfires raised new questions about chemical processes on smoke particles. This knowledge gap existed because only a few wildfires had injected significant amounts of smoke particles into the stratosphere during the satellite era. Leveraging over 30 years of satellite data, I separated chemical and dynamic processes affecting chlorine reservoir species to quantify chemical chlorine activation across different aerosol types. In Chapter 4, I developed a new approach to quantitatively estimate the onset temperature for chemical chlorine activation after the 2020 Australian wildfire using satellite observations. Chapter 5 applies this method to compare the impact of chemical chlorine activation from two independent wildfire events with that from a series of volcanic eruptions of varying magnitudes. Despite emerging challenges such as illicit emissions and recent wildfires and volcanic eruptions, advancements in observational records, our understanding of ozone chemistry, and computational power have significantly enhanced our ability to quantitatively detect and attribute stratospheric ozone changes. In Chapter 6, I applied a pattern-based “fingerprinting” technique to quantitatively separate the contributions of ODS forcing from other external forcings and internal variabilities in satellite observations. This analysis shows that Antarctic ozone increases cannot be explained by climate internal variability alone, providing strong confidence that ozone recovery is underway, primarily driven by human efforts to reduce ODS emissions.

Light-Induced Collective Interactions in Arrays of Quantum Emitters

2025-12-04T03:06:57Z

Light-Induced Collective Interactions in Arrays of Quantum Emitters Rubies-Bigorda, Oriol The interaction between light and matter has captivated physicists for centuries, from early studies of vision and refraction in ancient Greece to the development of quantum mechanics and quantum electrodynamics in the past century. While the response of a single quantum emitter to light is well understood, the radiative properties of an ensemble of closely spaced emitters are far more intricate. Coupling to a shared electromagnetic environment induces coherent and dissipative interactions between emitters, giving rise to a collective response that cannot be captured by treating them independently. In the regime of few excitations, the system hosts delocalized subradiant states, that is, coherent superpositions that are largely decoupled from the electromagnetic field and thus decay at suppressed rates. While this weak coupling makes subradiant states attractive for quantum technologies, it also renders them difficult to manipulate. At higher excitation densities, the intrinsic nonlinearity of emitters and the exponential growth of the Hilbert space make theoretical and numerical descriptions of the system and its dynamics increasingly challenging. This thesis explores two fundamental questions: How can subradiant and dark states be selectively accessed and harnessed for practical applications in quantum technologies? And how can interacting ensembles of quantum emitters be efficiently simulated to uncover their many-body physics? The first part of the thesis develops protocols for controlling and addressing dark states in free-space and waveguide-coupled atomic arrays, demonstrating their utility in quantum storage and the deterministic generation of entangled photonic states. Incorporating atomic motion, we further show that collective subradiant states can enhance cooling in dense atomic arrays, offering new avenues for controlling motional dynamics. In the second part, we introduce cumulant expansions of the equations of motion as a powerful tool to analytically and numerically investigate collective decay in the many-body regime. We first examine the collective decay of fully excited atomic arrays in free space, characterizing the onset and scaling of the superradiant burst across different geometries. In collaboration with experiments on ultracold erbium atoms in optical lattices, we provide the first direct observations of many-body collective effects in free-space ordered arrays, including early-time superradiant bursts, late-time subradiant tails, and the emergence of atomic correlations throughout the dynamics. Finally, we theoretically and numerically explore the transient formation of multi-excitation subradiant states, and demonstrate how the existence of multiple dissipation channels suppresses steady-state superradiance in extended arrays.

Hybrid Core Inductors for High Saturation Capability

2025-12-04T03:06:54Z

Hybrid Core Inductors for High Saturation Capability Yang, Rachel S. Power electronics are critical for any system requiring electricity and often impact the performance of these systems. In many cases, the performance of power electronics is limited by lossy and large inductors that are constrained by the saturation of their magnetic core material. Such saturation-limited inductors are typically found in power electronics applications where the inductor sees large dc current with relatively small ac ripple, such as EMI filters or converters operating in continuous conduction mode. This thesis investigates two types of inductor designs that can achieve higher saturation capability by combining multiple materials in a single core, enabling these designs to achieve greater energy storage or lower loss than conventional single-material cores. The first design combines a permanent magnet with a soft magnetic material (e.g. ferrite) to form a PM hybrid core. This core achieves higher saturation capability by directing PM flux to oppose winding flux in the ferrite. First-order models, design processes, and other theory for the PM hybrid core are developed in this thesis, and different geometries for this core are explored. Additionally, two PM hybrid core prototypes are presented, one using a pot core geometry and one using a modified E core geometry. The PM hybrid pot core prototype achieves 70% more energy storage or 50% of the dc loss versus comparable ferrite prototypes, while the PM hybrid E core prototype achieves 30% more energy storage or a minimum of 52% of the total loss versus comparable ferrite prototypes. The second design pairs a low-frequency, high-saturation material (e.g. steel) with a low-saturation, highfrequency material (e.g. ferrite) to form a steel hybrid core. This core achieves higher saturation capability by directing most of the dc flux to the steel and all of the ac flux to the ferrite, enabling the core to leverage both materials’ advantages while avoiding their detriments. First-order models and design processes for the steel hybrid core are developed in this thesis. An example steel hybrid core design using a pot core is also presented. This design can achieve 220% more energy storage versus a comparable ferrite prototype, and it may achieve lower loss. Its performance, though, is sensitive to manufacturing and assembly imperfections. In this thesis, both the PM hybrid and steel hybrid cores are demonstrated to have great potential in achieving high saturation capability. By leveraging these hybrid cores, inductor designs can achieve greater energy storage density or lower loss and thus enable higher performance power electronics.

Quantum Gas Microscopy of Bosonic Correlations in the Continuum

2025-12-04T03:06:49Z

Quantum Gas Microscopy of Bosonic Correlations in the Continuum Xiang, Jinggang This thesis details the complete upgrade and renovation of an existing experimental platform into a high-resolution quantum gas microscope for ultracold 87Rb atoms. Quantum gas microscopes enable site-resolved imaging, providing unprecedented access to quantum statistical effects and many-body phenomena. While such instruments are often employed to study physics in optical lattices, we have innovatively adapted our apparatus to investigate bulk system behavior. A major part of this project involved upgrading the scientific apparatus and retrofitting the previous system. We introduced new optical components, including a high-NA objective, and improved the vacuum system for better optical access. Extensive lab renovations, from upgrading the optical table to reorganizing the laser and imaging setups, were carried out to enhance mechanical and thermal stability. Rigorous optical benchmarking confirmed that the objective achieves diffractionlimited imaging, which is critical for resolving single atoms. This capability allowed us to detect density fluctuations at the scale of the thermal de Broglie wavelength in a quasi-two-dimensional gas of 87Rb atoms. In an experiment resembling Hanbury Brown and Twiss interferometry, we measured a 30% enhancement in the second-order correlation function in situ, demonstrating strong bosonic bunching. This outcome underscores the microscope’s precision and the importance of high-resolution imaging in capturing subtle quantum statistical effects. The successful realization of this apparatus demonstrates the utility of quantum gas microscopes in probing bulk systems. With this new platform in place, future studies can explore critical phenomena, many-body correlations, matter-wave emission, and quantum simulations with cold atoms.

Measurement of Cosmic Ray Lithium Isotopes Using the Alpha Magnetic Spectrometer

2025-12-04T03:06:51Z

Measurement of Cosmic Ray Lithium Isotopes Using the Alpha Magnetic Spectrometer LaVecchia, Gianni The study of cosmic rays and their properties provides insight into the origins of our universe and is a unique lens on the nuclear physics of the cosmos. The identification of cosmic ray isotopes poses a particular challenge, as it requires the measurement of multiple observables to a high degree of accuracy for the deduction of nuclear mass. Using the unique detection capabilities of the Alpha Magnetic Spectrometer (AMS), the isotope fluxes of cosmic ray lithium in the rigidity range of 1.92 to 25 GV are presented. This work is based on 0.97 million ⁶Li and 1.04 million ⁷Li nuclei collected by the AMS over a 12.5 year period, and improves the error and extent of existing measurements by a factor of 10. These results lead to the conclusion that there is no sizable primary component in cosmic ray ⁷Li. The improvements to the AMS velocity measurement establishes the groundwork for future cosmic ray isotope measurements.

Metrics, Muons, Moments, Models, Machine Learning, Measurements, and More: A Manifesto on Collider Physics

2025-12-04T03:05:49Z

Metrics, Muons, Moments, Models, Machine Learning, Measurements, and More: A Manifesto on Collider Physics Gambhir, Rikab The interface between particle theory and particle experiments is essential to improving our understanding of the Standard Model and looking for new physics beyond it. At this interface lies a complicated web of complex and expensive simulations that cannot fully be trusted, experimental and theoretical uncertainties, overwhelmingly large amounts of data, all while we have yet to find any deviations from the Standard Model. In this thesis, we propose strategies for improving the theory ↔ experiment pipeline at all stages. We first show how modern Machine Learning and statistical techniques can be used to improve the calibration and resolution of particle detectors in a robust way, which can lead to improved measurement precision. We then develop brand new classes of measurable observables based on the principle of infrared-and-collinear-safety, geometry, and machine learning, which come with guarantees about their theoretical calculability and interpretability, in turn motivating measurements at collider experiments. Finally, we then present two complementary approaches to search for new physics: one, in the form of an experimental proposal for a muon beam dump experiment that is viable alongside a full future collider program; and the other, in the form of machine-learning based anomaly detection to search for subtle signals in already-published data.

Practical Algorithms for Modeling Causality to Accelerate Scientific Discovery

2025-12-04T03:06:44Z

Practical Algorithms for Modeling Causality to Accelerate Scientific Discovery Wu, Menghua Scientific research revolves around the discovery and validation of causal relationships between variables. Machine learning has the potential to increase the efficiency of this process by proposing novel hypotheses from data observations, or by designing experiments that maximize success rate. This thesis addresses these problems through pragmatic approaches, designed to model large systems and incorporate rich domain knowledge. These algorithms are applied to use cases in molecular biology and drug discovery, which highlight their potential to inform efficient experiment design and to automate the analysis of experimental results.

Recycling and Regeneration of Spent Perfusion Media via Ion Concentration Polarization

2025-12-04T03:06:09Z

Recycling and Regeneration of Spent Perfusion Media via Ion Concentration Polarization Wynne, Eric Michael The widespread adoption of monoclonal antibody therapies is often constrained by their high prices, which can limit accessibility, particularly for patients in low- and middle-income countries. Addressing this economic barrier is crucial to ensure that life-saving treatments can reach all who need them. We present a series of bioprocessing innovations designed to reduce the cost of monoclonal antibody manufacturing and improve global access to these critical therapeutics. The work focuses on developing technologies for media regeneration and recycling, with the goal of reducing the economic and environmental impact of cell culture media in perfusion mammalian cell culture. We demonstrate a microfluidic separation device engineered to selectively remove metabolic waste products—specifically ammonia and lactate—from spent media using ion concentration polarization. Building on this foundation, a scalable millifluidic system was developed to enable higher-throughput waste removal. We characterized the impact of media recycling upon batch and perfusion cell cultures. We devised a nutrient supplementation strategy to create ‘regenerated’ media that minimized any effect on cell growth and productivity compared to fresh media. To support continuous manufacturing, a perfusion culture system incorporating a microfluidic spiral cell retention device and continuous cell bleed was established, and stable performance was maintained over extended durations. A further innovation introduced a multi-stage waste recovery system that increased media regeneration yield to 87.5%. This recovery rate enabled a self-recycling perfusion bioreactor in which 75% of the media feed was regenerated, without significant impact on cell growth, productivity, or product quality. Together, these advances establish a novel biomanufacturing platform that combines electrokinetic waste removal with media regeneration and recycling. The approach is broadly adaptable to mammalian cell culture processes and offers a promising path toward more sustainable, cost-effective, and environmentally responsible production of monoclonal antibodies and other biologics.

Scaling Cooperative Intelligence via Inverse Planning and Probabilistic Programming

2025-12-04T03:06:42Z

Scaling Cooperative Intelligence via Inverse Planning and Probabilistic Programming Zhi-Xuan, Tan How can we build cooperative machines that model and understand human minds — machines that assist us with our goals, coordinate on plans, infer the intentions behind our words, and even learn our norms and values? This thesis presents a scalable model-based approach to building such systems via inverse planning and probabilistic programming. First, we introduce a probabilistic programming architecture that implements a Bayesian theory of mind. This architecture, Sequential Inverse Plan Search (SIPS), performs online inference of human goals and plans by inverting a Bayesian model of incremental human planning. By combining high-performance symbolic planners with sequential Monte Carlo (SMC) inference, SIPS achieves faster-than-real-time speed, while scaling to hundreds of possible goals, and remaining robust to human mistakes due to boundedly-rational planning. Second, we present Cooperative Language-guided Inverse Plan Search (CLIPS), a system that integrates SIPS with large language models (LLMs) to model communicative cooperation. By using LLMs as likelihood functions within probabilistic programs, CLIPS can infer human goals from ambiguous instructions, then provide uncertainty-aware assistance with much higher levels of reliability than LLMs can on their own. In addition, CLIPS can be used to infer the shared intentions of communicating agents from their actions and words. Third, we show how inverse planning can model the acquisition of social normativity, formalizing norm-guided societal behavior as a norm-augmented stochastic game (NSG). In NSGs, agents assume that society follows a shared set of social norms, and infer these norms from the actions of other agents. By doing so, agents can rapidly learn cooperative social norms using orders of magnitude less data than model-free approaches. Finally, we present advances in probabilistic programming infrastructure that have enabled architectures such as SIPS and CLIPS. Through interfaces for programmable SMC and probabilistic programming with LLMs, developers can readily compose modeling and inference subroutines when designing probabilistically coherent intelligent systems. Together, these innovations demonstrate the feasibility and scalability of rational AI engineering for cooperatively intelligent machines, while illuminating the computational and algorithmic foundations of human cooperative intelligence.

Atomistic Study of Traveling Skyrmions in Multi-Sublattice Magnetic Materials

2025-12-04T03:05:19Z

Atomistic Study of Traveling Skyrmions in Multi-Sublattice Magnetic Materials Tremsina, Elizaveta A. The development of novel energy-efficient computing hardware is imperative for the reduction of the carbon footprint and for the extension of computing, mobile and wearable device lifespan. Recent advances have been focused on turning to novel material systems, and one such avenue is magnetic thin films. Bits of information can be encoded by magnetic twisted textures called skyrmions, which can be efficiently driven by applying electrical current. Recently, emphasis has been placed on investigating antiferromagnetic and ferrimagnetic skyrmions, as opposed to the single-sublattice ferromagnetic ones studied earlier, due to their potential for more rapid dynamics and magnetic stability. However, there is a pressing need for a thorough and detailed understanding of the intricacies of skyrmion motion, in particular, limiting velocity, optimization of trajectory, controlled mobility and, notably, the observed dynamic distortions of skyrmion profiles. For this reason, experimental studies are simply not enough to provide a complete picture, since the material parameter space for systems hosting skyrmions is quite large. We perform a comprehensive study combining simulation-based as well as analytical approaches, of the spin-orbit torque motion of skyrmions in a wide host of magnetic materials, ranging from crystalline antiferromagnetic to ferrimagnetic, to ferromagnetic. We systematically analyze the relationship between physical distortions of the skyrmion profiles, based on the action of local Thiele forces, and internal elastic tension forces, providing a quantitative and nuanced explanation of these effects. These results expand the understanding of fundamental properties of magnetic skyrmions, as well as their potential use in spintronics applications.

Transformative Lenses: Empowering Learners with New Perspectives Using Generative AI and Augmented Reality

2025-12-04T03:06:32Z

Transformative Lenses: Empowering Learners with New Perspectives Using Generative AI and Augmented Reality Leong, Joanne Sau Ling Learning is a fundamental human drive that has been shaped by technological advancements over the years. The emergence of generative AI marks a profound shift—its capacity to produce text, images, and video challenges long‐held beliefs about what only humans could create. This shift creates new opportunities for learning, including enabling the design of more customized and personalized learning experiences. Recognizing that learning is deeply influenced by our perceptions of ourselves, others, and our materials and environments, I propose creating transformative lenses powered by generative AI and augmented reality (AR) to adapt what learners perceive, as a means to empower them with new perspectives. I design and implement a set of novel interactive systems and experiences as case studies that address factors including creativity, communication, and motivation. Studying the use of these systems, I gather early evidence that such lenses can help people to overcome their own limiting thoughts and emotions to move towards realizing their full potential. Reflecting on these case studies, I distill key considerations for designing and applying transformative lenses. Finally, I discuss the broader implications of this work at the evolving intersection of generative AI and learning.

Language Models as Opinion Models: Techniques and Applications

2025-12-04T03:06:52Z

Language Models as Opinion Models: Techniques and Applications Brannon, William Real-time social media platforms now host the news cycle and shape public opinion, while large language models (LLMs) give us new tools to observe and predict those shifts. This dissertation links the new affordances of social media with the predictive power of LLMs to explain -- and forecast -- opinion change. We first quantify the dynamics of news on an influential social platform, then develop LLM-based tools to forecast persuasion and predict heterogeneous treatment effects (HTEs). Study I — Media tempo and tone. Using 518,000 hours of U.S. talk-radio broadcasts and 26.6 million tweets from elite and mass users, we show that Twitter discourse (i) moves faster at both take-off and fade-out stages of a news event and (ii) sustains greater outrage than radio – despite radio’s often explicitly outrage-focused business model. To our knowledge, this is the first large-scale, data-driven comparison between Twitter and traditional media of both outrage levels and the rate of decay of attention to news. Study II — Zero-shot persuasion forecasting. Across a diverse set of 28 randomized experiments, LLM-based methods outperform an ensemble of strong baselines at predicting HTEs and deliver good performance at predicting average treatment effects (ATEs) — all without any experiment-specific fine-tuning. Study III — Transfer and scaling. Fine-tuning LLMs on contemporaneous news coverage boosts HTE (and ATE) prediction performance greatly, to more than 3x baseline performance. A new minibatch-moment-matching (M3) objective lets us train a 400M-parameter model to nearly match the HTE prediction performance of an 8B model at a fraction of the inference cost. Transfer, however, falters out of distribution on held-out experiments and demographic groups, lending support to contextual theories of persuasion. Overall, we (i) quantify how platform affordances shape the tone and tempo of public discourse, (ii) introduce LLM-based methods that make causal experiments more sample-efficient, and (iii) chart the limits of transfer learning for opinion prediction. Our findings provide practical tools for HTE prediction and help researchers anticipate persuasion dynamics in a media landscape shaped by both humans and machines.

Across the Scales of the Nucleus: Understanding Short Range Correlations from Medium Modification to Probe Independence

2025-12-04T03:06:28Z

Across the Scales of the Nucleus: Understanding Short Range Correlations from Medium Modification to Probe Independence Denniston, Andrew W. The atomic nucleus presents an intricate system due to the non-linear forces described by Quantum Chromodynamics (QCD) that govern its structure. The range of scales involved is remarkable; the most massive nuclei weigh approximately five orders of magnitude more than the quarks that compose them. The nucleus can be analyzed at various levels, from quarks to hadrons to the nucleus as a whole. Short-Range Correlations (SRCs) within the nucleus play a significant role that spans these diverse scales. At the most fundamental level, SRCs influence the interaction between nucleons. The nucleon-nucleon (NN) interaction, arising from QCD, is crucial in determining nuclear properties. SRCs serve as valuable probes for measuring this NN interaction, as the nucleons within SRCs become effectively decoupled from the rest of the nucleus. Multiple experimental techniques, including electron scattering, have been employed to investigate the NN interaction through SRCs. However, our first project demonstrates that inclusive measurements alone are inadequate to constrain this interaction fully. Moving to the scale of the nucleus, SRCs contribute to the high-momentum tail of the nuclear spectral function. While the low-momentum region is characterized by nucleons exhibiting bulk properties, nucleons begin to pair into SRCs at higher momenta. Our research aims to bridge the understanding between the mean-field portion of the nucleus and its high-momentum SRC components. Additionally, SRCs affect the quark structure of protons, as evidenced by the EMC effect, which indicates that quarks behave differently when protons are embedded within a nucleus—an effect referred to as medium modification. This thesis explores the correlation between SRCs and medium modification across various experimental setups. Finally, we seek to establish an interpretation of the nuclear ground-state. Accomplishing this requires demonstrating that our SRC observables are independent of the probe’s scale and scheme. The concluding project of this thesis illustrates how we utilize triple coincidence quasi-elastic scattering across a range of (Q2 ) values to develop a model-dependent framework for understanding SRC distributions within the nucleus’s ground-state wavefunction.

Discovering and Engineering the Computation Underlying Large Intelligent Agents

2025-12-04T03:06:47Z

Discovering and Engineering the Computation Underlying Large Intelligent Agents Sharma, Pratyusha The richness of language and intelligent behavior has often been attributed to latent compositional structure. Can we build tools for discovering how deep networks learn and represent this latent structure implicitly? And more importantly, can we use this knowledge to improve generalization in largely structure-less general purpose models or refine our understanding of the world they describe? In this dissertation, I present three perspectives to answer these questions. First, I present experimental methods to functionally characterize the space of learnt solutions in LLMs and demonstrate how this understanding can be used to improve their empirical generalization in a gradient free manner, sometimes by as much as 30% points on language understanding benchmarks. Following that, I show how to decipher the structure of another (black box) language-like system, the naturally arising communication system of sperm whales in the wild, discovering for the first time a unique combinatorial communication system. Finally, I apply insights from these results to equip embodied agents with a latent language of thought—hierarchical and compositional—and show how it can enable long-horizon reasoning and planning in these systems. This dissertation ultimately aims to bridge the gap between natural and artificial intelligence, offering new insights into both the fundamental nature of communication in complex biological organisms in the wild and the development of more powerful, and improved AI systems. A key pattern in the discoveries in this thesis has been how simple structures enable complex externalized behaviors in both biological organisms and AI systems.

Volume Mount Devices

2025-12-04T03:09:35Z

Volume Mount Devices Han, Alan As Moore's Law ends and AI demands increasingly tax our climate and resources, the limitations of two-dimensional electronics integration have become critical bottlenecks. Surface-mount devices (SMDs) remain entrenched in industry practice despite being insufficient for today's computing challenges and sustainability needs. This thesis introduces the volume mount device (VMD), a three-dimensional electronics packaging standard that bypasses the traditional die-to-server stack while offering a scalable, reversible framework inspired by natural ecosystems' circularity. The VMD approach embeds both electrical function and mechanical structure into modular elements that assemble freely in 3D space. Rather than building circuits on planar PCBs, this system constructs functional circuits by linking components into a self-constraining lattice architecture. My current implementation leverages existing supply chains by incorporating SMD components on small tile PCBs, while establishing a pathway toward eventually replacing SMDs at the IC packaging level. I developed a hybrid assembly system combining 3D printing and pick-and-place automation to build multi-layered electronic assemblies efficiently. Where prior work achieved only tens of parts at hundreds of components per hour (CPH), my system demonstrates automated assembly of hundreds of integrated elements at approximately 1000 CPH. I evaluate various geometric configurations, assess performance overhead compared to conventional approaches, and develop cost-effective, self-aligning connector interfaces for reliable joints—creating a foundation for electronics systems that can be assembled, disassembled, and reassembled as needed while improving resilience against supply chain disruptions.

Generative Diffusion Models Towards De Novo Protein Design

2025-12-04T03:06:39Z

Generative Diffusion Models Towards De Novo Protein Design Yim, Jason De novo protein design aims to generate proteins with desired functions by rationally engineering novel protein structures and sequences. The structure requires modeling continuous 3D coordinates of atoms with rigid biochemical constraints of the polymer chain while the sequence is a series of discrete amino acids that should fold into a plausible structure. Understanding the protein function-structure-sequence relationship necessary for protein design is complex, but deep learning has proven promising to learn the relationship from large protein datasets. This thesis aims to develop deep learning models that generate novel structures and sequences that can be guided towards desired functions. We first describe novel generative models that learn to generate protein structures and sequences by developing diffusion models over general state spaces including Riemannian manifolds and discrete tokens. The resulting methods – FrameDiff, FrameFlow, and MultiFlow – demonstrate the ability of diffusion models to extrapolate beyond the training data to generate novel and diverse protein structures and sequences that pass in silico protein design filters. Next, we apply diffusion models to practical protein design challenges by collaborating with experimental and computational biologists to develop RoseTTAFold Diffusion (RFdiffusion). By combining the structure prediction capabilities of RoseTTAFold and diffusion modeling principles, RFdiffusion can generate functional proteins with in vitro validated properties such as high-affinity binders and symmetric protein assemblies. De novo protein design aims to generate proteins with desired functions by rationally engineering novel protein structures and sequences. The structure requires modeling continuous 3D coordinates of atoms with rigid biochemical constraints of the polymer chain while the sequence is a series of discrete amino acids that should fold into a plausible structure. Understanding the protein function-structure-sequence relationship necessary for protein design is complex, but deep learning has proven promising to learn the relationship from large protein datasets. This thesis aims to develop deep learning models that generate novel structures and sequences that can be guided towards desired functions. We first describe novel generative models that learn to generate protein structures and sequences by developing diffusion models over general state spaces including Riemannian manifolds and discrete tokens. The resulting methods – FrameDiff, FrameFlow, and MultiFlow – demonstrate the ability of diffusion models to extrapolate beyond the training data to generate novel and diverse protein structures and sequences that pass in silico protein design filters. Next, we apply diffusion models to practical protein design challenges by collaborating with experimental and computational biologists to develop RoseTTAFold Diffusion (RFdiffusion). By combining the structure prediction capabilities of RoseTTAFold and diffusion modeling principles, RFdiffusion can generate functional proteins with in vitro validated properties such as high-affinity binders and symmetric protein assemblies.

Learning to Solve Long-Horizon Robot Manipulation Problems

2025-12-04T03:06:06Z

Learning to Solve Long-Horizon Robot Manipulation Problems Yang, Zhutian If we want mobile robots that perform multi-step tasks in visually diverse and geometrically complex environments, we need them to quickly decide what to do and how to do it. Manipulating multiple objects in environments with movable and articulated obstacles over time requires the robot to satisfy constraints like collision-freeness, reachability, and action feasibility. For problems with large state spaces, continuous action spaces, and long decision horizons, the hybrid constraint satisfaction problems induced by planners become combinatorially difficult to solve. In this thesis, I will discuss strategies for using offline learning to speed up deploymenttime planning, i.e., using a plan feasibility predictor, a subgoal generator, or a compositional joint continuous constraint solver. I will also present strategies for chaining policies learned from demonstrations using conditional inputs, such as key poses and natural language, for generalization in real-world environments. With the resulting efficient long-horizon manipulation planning system, we can solve complex robotic manipulation problems faster at deployment time. It can also be used to generate diverse large-scale whole-body trajectories as part of the data mixture for training robot foundation models in embodied reasoning, planning, and acting.

Building small domain-specific masked language models vs. large generative models for clinical decision support and their effects on users.

2025-12-04T03:06:17Z

Building small domain-specific masked language models vs. large generative models for clinical decision support and their effects on users. Sergeeva, Elena The frequently adopted definition of knowledge defines it as “justified true belief”. As one may notice this definition presents some issues when applied to AI: it is unclear to which degree it is justified to use “humanizing” vocabulary like “belief” or “justification” when describing the performance of an AI system. Traditional explicit knowledge-representation based AI involves reasoning over symbolic representation of statements standing for such “justified true beliefs” [1], the modern connectionist methodology however replaces explicit reasoning with making a prediction based on a set of computations done over weighted continuous representations of the inputs. The continuous representations learned by such systems remain “black box-like”, where the only elements directly understandable by the human user are the model inputs and outputs. In the first part of this thesis I introduce a set of Masked-Language model transformer based models for a diverse set of medical natural language processing tasks including Named Entity Recognition, Negation Extraction and Relation extraction that perform as well or better than bigger prompt-and-generate transformer-based causal language models. In the second part of the thesis, I discuss the modern “prompt-and-generate” approach to natural language processing where both the inputs and the outputs of the model are word-like elements commonly referred to as “tokens”. I explore the nature of token based representation of the input and look at the way token “meaning” is refined at each layer of the successive transformer computation. With respect to the outputs, I explore how people engage with AI generated sequences of tokens that people perceive as “explained” predictions.

Language-Centric Medical Image Understanding

2025-12-04T03:05:51Z

Language-Centric Medical Image Understanding Wang, Peiqi This thesis advances medical image understanding by leveraging the multifaceted roles of language: as supervision, prior knowledge, and a medium for communication. We introduce three main contributions: (1) a weakly supervised framework that uses language in clinical reports to guide fine-grained alignment between image regions and textual descriptions, (2) an adaptive debiasing method that uses language prior to improve the robustness of learning algorithms under noisy supervision, and (3) a novel approach for calibrating linguistic expressions of diagnostic certainty, enabling more reliable communication of clinical findings. Together, these methods lead to more accurate, robust, and reliable machine learning systems, ultimately streamlining clinical workflows and improving patient care.

Exploring spin physics with ultracold atoms

2025-12-04T03:06:24Z

Exploring spin physics with ultracold atoms Lee, Yoo Kyung The dynamics of many interacting spins is an active frontier of research; not only can they explain magnetic phenomena, but they also provide paradigmatic models with deep connections to high-T_c superconductivity, optimization problems, neural networks, and more. Experiments with ultracold alkali atoms in optical lattices have realized spin models with great success. In particular, the isotropic Heisenberg model---the XXX model---was realized more than a decade ago. The ⁷Li apparatus described here was the first to realize a tunable, anisotropic Heisenberg model, also known as the XXZ model. In this thesis, I will describe how the capabilities of this apparatus were harnessed to characterize the spin models we realize, employ them to observe new resonances, and to contribute to studies in spin squeezing and fundamental physics. First, I will discuss how we prepared and observed phantom helix states: eigenstates of the XXZ Hamiltonian. Our understanding of the contact interactions and the phantom helix states enabled us to observe long-predicted lattice-induced resonances, whose effects can be leveraged as another knob to tune the XXZ Hamiltonian. Furthermore, our control over the spin system allowed us to generate spin-squeezed states, a paradigmatic form of entanglement for spin ensembles. This is the first time squeezed states were realized with nearest-neighbor contact interactions in a lattice. Finally, our control over the spin degree of freedom and defects in our state preparation allowed us to create pristine periodic lattices with which to study gedankenexperiments in light scattering.

Probing the Diversity of Fast Radio Bursts with CHIME/FRB

2025-12-04T03:05:31Z

Probing the Diversity of Fast Radio Bursts with CHIME/FRB Shin, Kaitlyn Fast radio bursts (FRBs) are extremely bright extragalactic radio transients that flash for microseconds to milliseconds at a time, most never to repeat again. Encoded in every observed FRB is information from burst propagation effects, giving us clues about their mysterious origins as well as the environments they traveled through. With inferred all-sky rates of hundreds per day, FRBs have held great interest for those interested in extreme astrophysical processes as well as those interested in cosmological properties of the Universe. The Canadian Hydrogen Intensity Mapping Experiment (CHIME)/FRB project has revolutionized the FRB field with its field-leading discovery rate. With CHIME/FRB, we can start to carry out population-level studies of FRBs to constrain their origins and inform their use as cosmological probes. I present the first population-level studies of CHIME/FRB-observed FRBs using the CHIME/FRB Catalog 1 data release and the injections system to account for observational biases. I discover that CHIME/FRB is likely observationally biased against bursts originating from turbulent local environments, and constrain the energy and distance distributions of FRBs. I also present the Catalog 1 dataset updated with channelized raw voltage (“baseband”) data (“BaseCat1”), for which I played a pivotal role. The CHIME/FRB baseband localization pipeline can localize FRBs to arcminute-precision as long as the signal is bright enough to trigger the saving of offline baseband data. I then discuss two single source-studies enabled by the baseband localization pipeline — one discovering repeaters during phases of unusually heightened burst activity, and one using the burst properties of an unusual FRB to probe the properties of its sightline. In the latter study, I constrain the electron density content of a diffuse filamentary structure on the outskirts of the Virgo Cluster, demonstrating the power of FRBs as probes of diffuse media.

Controlling for the Ionospheric and Baseline-Offset Uncertainties in the CHIME/FRB Outriggers VLBI Network for Milliarcsecond Precision

2025-12-04T03:09:33Z

Controlling for the Ionospheric and Baseline-Offset Uncertainties in the CHIME/FRB Outriggers VLBI Network for Milliarcsecond Precision Willis, Jacob Fast radio bursts (FRBs) are a novel form of radio transients discovered in 2007. These bright, extragalactic radio signals have an inferred all-sky rate of hundreds of detections per day. The properties of FRBs hold valuable clues about the extreme physical processes driving them while also holding information about the astrophysical plasmas they traverse on their journey to Earth. The Canadian Hydrogen Intensity Mapping Experiment (CHIME)/FRB project has led the field with the hundreds of FRB detections the collaboration has published to date. However, these detections typically have localization regions so large that we cannot identify a single host galaxy, never mind its local environment. To improve upon this, CHIME/FRB has been transformed into a very long baseline interferometry (VLBI) array, drastically increasing the angular resolution of CHIME/FRB from arcminute to sub-arcsecond precision. In this work, I present my contributions to commissioning the CHIME/FRB VLBI Outrigger station located at the Green Bank Observatory (GBO) in West Virginia. This includes measuring and validating GBO's exact position to enable the localization of FRBs to sub-arcsecond precision. For VLBI networks spanning thousands of kilometers, the difference in the local ionospheric environments is significant and leads to errors in the CHIME/FRB Outrigger localizations. I present a thin shell model of the ionosphere to parameterize the local ionospheric environment for each VLBI station. This model may be used to interpolate the error induced by the ionosphere in FRB observations.

Using Z-hadron correlations to probe the medium response in PbPb and pp collisions at √ˢNN = 5.02 TeV

2025-12-04T03:09:22Z

Using Z-hadron correlations to probe the medium response in PbPb and pp collisions at √ˢNN = 5.02 TeV Chou, Pin-Chun The first measurement of Z-hadron two-particle correlation function are reported in PbPb collisions at √ˢNN = 5.02 TeV, using the PbPb collision data taken in 2018. The integrated luminosity of the PbPb data is 1.67 ±0.03 nb⁻¹ which made the analysis possible for the first time. Collision data with at at least one Z boson with 40

DePUDS: Decentralized Prosocial Urban Development System

2025-12-04T03:06:16Z

DePUDS: Decentralized Prosocial Urban Development System Zhang, Yan Urban areas face severe socio-economic and environmental challenges like housing crises, inequity, and environmental degradation, often worsened by traditional zoning practices. These are typically rigid, inefficient, outdated, and susceptible to obstruction by narrow special interests (NIMBYism), failing to engage the broader community or adapt to evolving needs. This dissertation proposes the Decentralized Prosocial Urban Development System (DePUDS), a novel governance framework designed to overcome these shortcomings by empowering informed collective consensus and including the often-silent majority. DePUDS integrates decentralized technologies like blockchain and smart contracts with structured economic incentives, facilitated through an accessible user-friendly Decentralized Application (DApp) to encourage broad participation. This system fosters transparent, inclusive, and equitable urban development. Its core mechanism, adaptive incentive-based zoning, dynamically aligns developer profitability with community-endorsed priorities—such as affordable housing, public amenities, and sustainability—providing flexibility absent in traditional zoning. Employing advanced agent-based simulations enhanced by large language models (LLMs), this research rigorously assesses DePUDS's effectiveness across two distinct case studies: Kendall Square in Cambridge, MA (a dynamic innovation hub) and the Inner Richmond District in San Francisco, CA (a culturally rich but housing-constrained neighborhood). Simulation results demonstrate DePUDS significantly aligns development outcomes with community preferences. In Kendall Square, targeted incentives substantially increased affordable housing and public amenities without hindering private investment. In the Inner Richmond, substantial community-driven incentives successfully unlocked constrained development, markedly reducing displacement risks, boosting affordable housing, enhancing amenity access, lowering carbon emissions via density, and preserving local cultural assets. The comparative analysis underscores DePUDS's versatility, showing its potential to enhance growth in active markets and stimulate development in constrained ones. Key policy implications point towards structured DApp-based community participation, adaptive incentive zoning, and dedicated funding. While acknowledging practical implementation hurdles (legal, economic, technological), the findings affirm the feasibility, effectiveness, and transformative potential of decentralized, incentive-driven urban governance. This dissertation offers significant theoretical contributions, practical policy guidelines, and future research pathways to foster more inclusive, sustainable, and resilient urban communities.

Materializing Light: Real-time, Handheld Fabrication of Programmable Structural Color

2025-12-04T03:09:31Z

Materializing Light: Real-time, Handheld Fabrication of Programmable Structural Color Myers, Paris G. Structural color is nature’s programmable color palette. While pigments and dyes absorb light to produce color, structural color uses nanoscale, light-reflecting structures to appear iridescently colored. We present MorphoChrome, an optical device for real-time, handheld, programmable structural color fabrication. Analogous to painting with light, MorphoChrome creates multicolor, structurally colored designs by exposing a commercially available holographic photopolymer film to user-controlled wavelengths. Within the device, red, green, and blue laser diodes go through an optical prism, combining light and producing mixed color outputs on the film. Additionally, we introduce a resin-based process to adhere and integrate the structurally-colored film with flexible and rigid objects and diverse making processes. In this thesis, we focus on the device optical design and fabrication, color-mixing, color output UI controller, device aperture tips, and holographic photo-polymer film adherence process. We evaluate the available color space and color resolution, and demonstrate creative fabrication applications.

Biophysical specializations supporting efficiency in neural networks

2025-12-04T03:06:19Z

Biophysical specializations supporting efficiency in neural networks Toloza, Enrique H.S. Neuroscience and artificial intelligence (AI) research have long enjoyed a synergistic relationship. AI has drawn key inspiration from the organization and function of the brain, while our understanding of the biological processes underlying computation has been profoundly enriched by studying the behavior of artificial systems. As breakthroughs in generative AI continue to transform our world, and as the need for more sustainable artificial neural systems becomes more urgent, the neuro-AI feedback loop has never been more important. AI needs ever more powerful and efficient systems, and neuroscience needs further insight into how our brains work. The development of more brain-like AI promises solutions to both of these problems. Unfortunately, this has thus far been stymied by two critical challenges: 1) how do we identify the features that make a system brain-like and 2) how do we incorporate these features into artificial networks in a useful and interpretable way? To address the first of these challenges, I will use the remarkable structural and biophysical diversity of the brain as an introduction into what it means for a system to be “brain-like.” This will lead us to a discussion of dendrites, the tree-like structures implicated at virtually every length scale of neural computation. Dendrites will thereafter act as the focal point for our study of brain-like computation. Specifically, I will trace how relatively simple biophysical features defined at the subcellular level can transform the computational landscape of large networks of neurons. To address the second of these challenges, it is necessary to discuss several enduring problems in computational neuroscience, broken down as chapters in this thesis. In Chapter 2, I will present the development of a new model of single-neuron dynamics that is realistic enough to capture the rich dynamics of dendritic spiking but efficient enough for use in simulations of thousands of neurons, thereby filling a long unmet need in the field. In Chapter 3, I will describe a solution to the general problem of training neural networks with arbitrary differentiable dynamics, thus opening the door for the study of countless biophysical phenomena in the context of networks that can learn to perform computations. In Chapter 4, I will use these tools to test several longstanding hypotheses regarding the utility of different biophysical features in neurons, performing first-of-their-kind fair comparisons of the computational performance of spiking networks, rate-based networks, and networks with nonlinear and linear dendrites. Finally, in Chapter 5, I will use insights gained from studying dendrites at the network level to provide a new perspective as to how the structural and biophysical diversity of the brain could emerge from a complex interplay of functional pressures (e.g., task demands) and physical constraints (e.g., space and energy). Together, the chapters of this thesis outline a general quantitative framework for building more brain-like AI for use in both AI research and neuroscience. This framework illustrates how biophysical specializations arising at the level of single neurons shape the emergent dynamics of the brain.

Next Week Tonight: Simulating Counterfactual Narratives of the future using Agentic Knowledge Graphs

2025-12-04T03:09:32Z

Next Week Tonight: Simulating Counterfactual Narratives of the future using Agentic Knowledge Graphs Agarwal, Gauri Understanding the ripple effects of events—both real and speculative—is essential for navigating complex futures. Large Language Models (LLMs) have emerged as powerful tools that offer a user-friendly and narrative experience for question answering and reasoning across large corpuses of unstructured data [15, 96]. While LLMs can respond to complex ‘what-if’ questions, they typically provide single, unverifiable answers. Even with retrievalaugmented generation (RAG) that grounds LLM responses on external sources, the opacity of reasoning pathways undermines trust in model outputs [97]. Next Week Tonight builds on the narrative and reasoning capability of LLMs further by enhancing the exploration of what-if futures and making it more transparent and evidencebased. NWT exposes the underlying knowledge graph, allowing users to inspect inference pathways directly. This also enables the generation of multiple, diverse scenarios from a single condition—each following different but explainable causal chains. In testing 15 counterfactual prompts that span diverse news topics, NWT produced scenario narratives that were rated as significantly more causally coherent, transparent, and easier to audit than standard chat completions. Beyond technical performance, NWT reinvents scenario planning as an interactive narrative experience - encouraging curiosity, critical thinking, and deeper engagement with the complexities of future events. By surfacing not only what could happen but why and how, NWT aims to empower analysts, policymakers, and the public to navigate uncertainty with greater clarity and confidence. Github: https://github.com/viral-medialab/next-week-tonight

On the Learnability of General Reinforcement-Learning Objectives

2025-12-04T03:05:33Z

On the Learnability of General Reinforcement-Learning Objectives Yang, Cambridge Reinforcement learning enables agents to learn decision-making policies in unknown environments to achieve specified objectives. Traditionally, these objectives are expressed through reward functions, enabling well-established guarantees on learning near-optimal policies with a high probability — a property known as probably approximately correct (PAC) -learnability. However, reward functions often serve as imperfect surrogates for true objectives, leading to reward hacking and undermining these guarantees. This thesis formalizes the specification and learnability of general reinforcement-learning objectives beyond rewards, addressing fundamental questions of expressivity and policy learnability. I examine three increasingly expressive classes of objectives: (1) Linear Temporal Logic (LTL) objectives, which extend conventional scalar rewards to temporal specifications of behavior and have garnered recent attention, (2) Computable objectives, encompassing a broad class of structured, algorithmically definable objectives and (3) Non-computable objectives, representing general objectives beyond the computable class. For LTL objectives, I prove that only finitary LTL objectives are PAC-learnable, while infinite-horizon LTL objectives are inherently intractable under the PAC-MDP framework. Extending this result, I establish a general criterion: an objective is PAC-learnable if it is continuous and computable. This criterion facilitates the establishment of PAC-learnability for various existing classes of objectives with unknown PAC-learnability and informs the design of new, learnable objective specifications. Finally, for non-computable objectives, I introduce limit PAC-learnability, a practical relaxation where a sequence of computable, PAC-learnable objectives approximates a non-computable objective. I formalize a universal representation of non-computable objectives using nested limits of computable functions and provide sufficient conditions under which limit PAC-learnability holds. By establishing a theoretical foundation for general RL objectives, this thesis advances our understanding of which objectives are learnable, how they can be specified, and how agents can effectively learn policies to optimize them. These results contribute to the broader goal of designing intelligent agents that align with expressive, formally defined objectives—moving beyond the limitations of reward-based surrogates.

Solid-State Quantum Memories for Near-Term Quantum Repeaters

2025-12-04T03:06:12Z

Solid-State Quantum Memories for Near-Term Quantum Repeaters Sutula, Madison M. Over the past decade, quantum computers have emerged as a promising technology to enable transformational advances in information processing and communication and solve problems that are intractable to classical computers. While there is great promise in linking quantum computers together over long distances via quantum channels, these technologies are still under development. Solid-state emitters with coherent spin-photon interfaces, long spin lifetimes, and narrow optical transitions are a leading platform for use as quantum memories in networked quantum repeaters. However, while such emitters have already enabled advanced quantum networking demonstrations in laboratory settings, applying these devices as useful memory devices at scale is a key outstanding challenge. In this thesis, we experimentally investigate solid-state quantum memories for quantum information applications. First, we develop experimental techniques to characterize solid-state emitters with high throughput, enabling both better understanding of the distribution of emitter properties and improved feedback on material preparation and device fabrication. Next, we implement quantum frequency conversion to create a coherent spin-photon interface between silicon-vacancy centers in diamond and optical photons in the low-loss telecom band. Finally, we investigate color centers in other engineering materials, including silicon and silicon carbide, to better understand the fundamental trade space of requirements for solid-state hosts. Together, these efforts represent a significant advance in creating, controlling, and deploying telecom-compatible spin interfaces, paving the way for memory-enabled quantum repeaters.

Inferring Clonal Dynamics in Blood using Single-Cell Measurements

2025-12-04T03:09:25Z

Inferring Clonal Dynamics in Blood using Single-Cell Measurements Perry, Andrea N. In this work, we uniquely tag hematopoietic (blood) stem cells with genetic barcodes and follow their progeny over time to ask whether clonally related cells in myeloproliferative neoplasms (MPNs) favor particular blood cell fates. Myeloproliferative neoplasms are clonal disorders driven most frequently by the JAK2-V617F mutation, which arises in a single hematopoietic stem cell (HSC) and ultimately dominates the normal process of blood cell production. Although all patients carry the same driver mutation, they still branch into three distinct disease forms—essential thrombocythemia (ET), polycythemia vera (PV), or primary myelofibrosis (PMF)—and the reason for this variation remains unknown. One compelling hypothesis is that the JAK2-V617F mutation may arise in HSC subsets with intrinsic biases toward platlet-producing cells (as in ET) or red blood cell precursors (PV). To investigate this question, we analyzed bone-marrow cKit⁺ cells from mice engineered for inducible MPN disease and CRISPR array repair lineage tracing (CARLIN), using single-cell RNA sequencing. Our gene expression analysis shows that the mutation keeps key signaling and stress-response genes switched on and boosts growth-promoting enzymes, collectively pushing blood production toward the myeloid line. At the resolution of individual CARLIN clones (i.e. cells grouped by a shared progenitor), however, we observe no robust mutation-induced lineage bias—an outcome attributable to limited clone recovery and inter-mouse variability. Crucially, this work establishes a scalable analysis pipeline for future, higher-yield CARLIN experiments. Enhancing lineage-tracing sensitivity, barcode diversity, and biological replication will be essential to test whether these interferon-/stress-response and kinase programs manifest as subtle, clone-level fate biases in JAK2-driven MPN.

Principled Approaches for Latency Reduction in Networking Systems

2025-12-04T03:05:20Z

Principled Approaches for Latency Reduction in Networking Systems Pit-Claudel, Benoit Modern networks face unprecedented challenges due to exponential growth in traffic demands, driven by AI workloads in datacenters and the ubiquitous adoption of cloud services across the internet. This dissertation addresses three critical challenges in network systems: efficient scheduling of inference tasks, performance optimization in hybrid networks, and memory-efficient load balancing in datacenters. First, we introduce Nona, a stochastic scheduling framework that leverages queueing theory to optimize task placement in datacenter environments. By employing randomized algorithms and considering both network and compute constraints, Nona demonstrates multiple orders of magnitude improvements in job completion times while maintaining implementation simplicity. Nona proposes stochastic scheduling, in which the complexity of the scheduling problem is moved to an offline phase. When handling jobs online, stochastic schedulers are oblivious to the instantaneous state of the network and only rely on predetermined allocation probabilities to make lightning-fast decisions. Second, we present LINC, an in-network coding solution designed for hybrid backbone networks. Through comprehensive mathematical analysis and simulation, we highlight the benefits of network coding in cases where no modifications of the end-hosts are possible. Finally, we develop Sirona, a memory-efficient version of a reactive subflow spraying mechanism suited for hardware deployment. We show that Sirona can achieve competitive performance in homogeneous and heterogeneous datacenter networks while keeping a low memory footprint.

Forward Modeling for Bolometry and Disruption Mitigation in Tokamaks or How to Kill Your Plasma With Confidence, Style, and Pizzazz

2025-12-04T03:06:08Z

Forward Modeling for Bolometry and Disruption Mitigation in Tokamaks or How to Kill Your Plasma With Confidence, Style, and Pizzazz Stein-Lubrano, Benjamin The tokamak is a promising approach to magnetic confinement fusion. Tokamak functionality is threatened by plasma disruption events, which can damage critical machine components. Disruption damage can be mitigated by high-Z impurities, delivered by Massive Gas Injection (MGI) or Shattered Pellet Injection (SPI). Impurities radiate energy out of the plasma and onto the first wall. Evenly distributed radiation causes less damage than unmitigated disruption pathways, which deliver concentrated heat loads. In order to successfully develop and deploy mitigation systems, it is important to accurately measure and characterize disruption radiation. Accurate measurement is challenged by fast disruption timescales and highly asymmetric radiation patterns, which push the time and spatial resolution limits of radiant heat sensors. Previous radiation analysis approaches are typically limited to two dimensions or less by the highly under-determined nature of tomographic reconstruction and limited spatial resolution of sensors. Two dimensional analysis is often inaccurate for disruption radiation, which can be highly three dimensional as a result of localized impurity sources and fast 3D MHD events. In this thesis, I present a new algorithm for 3D radiation analysis in tokamak disruptions, called Emis3D. When Emis3D is applied to mitigated disruptions on the JET tokamak, a significant injection plume radiation effect in mitigated disruptions is revealed. When this effect is included in radiated energy calculations, the mitigated radiation fraction of plasmas with high thermal energy content is significantly improved, indicating that thermal mitigation is more effective than previously thought. Emis3D can also be used as a design tool to evaluate potential radiant heat sensor layouts. When applied to the SPARC tokamak, Emis3D demonstrates that toroidally skewed sensor sightlines improve spatial resolution and reduce blind spots, allowing more accurate measurement.

Understanding the Milky Way with Stars

2025-12-04T03:05:44Z

Understanding the Milky Way with Stars Ou, Xiaowei "How do galaxies form?" is one of the most important questions in modern astrophysics. Hierarchical growth, the most plausible theory behind galaxy formation, suggests that galaxies, including the Milky Way, assemble through the accretion of smaller systems, over a scaffolding of the invisible Dark Matter. Such growth is evidenced by the differences in stellar structures found in the Galaxy over the last few decades, accelerated most recently by the Gaia space mission. Yet, we still lack a full picture of the formation of the Milky Way and its stellar components, and we are even further in understanding its underlying Dark Matter distribution. For the latter, discrepancies between observations and predictions from CDM model at galactic scales have sparked debate about how well this model accounts for the evolution of the Milky Way. Stellar tracers provide a powerful tool for examining these discrepancies, helping us explore the hierarchical assembly of galaxies in the Local Group and test different models for dark matter. At the same time, cosmological simulations and machine learning techniques offer a bridge between the theory and observations. In this thesis, I combine observation of stellar kinematics and chemistry with cosmological simulations to understand the formation and evolution of the Milky Way and its satellite dwarf galaxies. I map the dark matter distributions in the Milky Way and one of its ultra-faint dwarf galaxies using stellar dynamics, combining simulations of tidal disruption with observational data to study ongoing merger events and how hierarchical assembly shaped the Milky Way today. I conduct robust machine learning searches of kinematic substructures from disrupted dwarf galaxy debris in the Milky Way and utilize stellar heavy element abundances to probe the galaxies that merged with the Milky Way in the past. Lastly, I develop synthetic surveys from simulations to bridge gaps between theory and observation, testing the robustness of current and future methodologies in understanding how the Milky Way came to be.

Coating Thermal Noise in Gravitational-Wave Detectors

2025-12-04T03:05:13Z

Coating Thermal Noise in Gravitational-Wave Detectors Demos, Nicholas The direct detection of gravitational waves, originating from cataclysmic events such as black hole and neutron star mergers, has ushered in a new era of observational astronomy. These signals offer unique insights into astrophysical phenomena and fundamental physics, but fully realizing their potential requires continued improvements in detector sensitivity. A primary factor limiting the performance of current ground-based interferometers like Advanced LIGO and Advanced Virgo is thermal noise arising from the highly reflective multilayer coatings on the test mass mirrors. Reducing this coating thermal noise, particularly its Brownian component, while simultaneously maintaining exceptionally low optical absorption and scatter is necessary to advance detector capabilities. This thesis addresses this challenge through the characterization and development of alternative coating materials and designs. Central to this work is a dedicated experimental apparatus employing a high-finesse folded optical cavity and a multimode co-resonance technique. This system enables direct, high-precision measurements of coating thermal noise in the frequency band relevant to gravitational-wave detectors and allows for relatively rapid evaluation of candidate coatings, providing timely feedback for materials development. Coating materials such as niobia-based oxides, hafnia-tantala mixtures, and substoichiometric silica, were explored employing strategies like compositional optimization, post-deposition annealing, and multimaterial designs with buried layers. Progress toward lower-noise coatings is demonstrated. Highly reflective coatings based on optimized titania-silica, titania-germania, and ternary silicon nitride structures achieved thermal noise levels approximately 75% that of current detector coatings. These coatings also exhibited exceptionally low optical absorption, reaching levels near 1 part-per-million following appropriate heat treatment. While challenges related to defect formation during annealing and discrepancies between different noise measurement methodologies were identified, ongoing research, particularly on defect mitigation in materials like titania-germania, continues to advance the field. The findings presented here contribute to the materials science foundation for improving current gravitational-wave detectors and guiding the design of future observatories.

Time-Domain Astrophysics with the Transiting Exoplanet Survey Satellite

2025-12-04T03:05:59Z

Time-Domain Astrophysics with the Transiting Exoplanet Survey Satellite Jayaraman, Rahul The Transiting Exoplanet Survey Satellite (TESS) is conducting an all-sky survey with the primary aim of detecting planets orbiting nearby stars. However, its large field of view and 200 s imaging cadence are useful for other science cases, ranging from stellar astrophysics to transient science. This thesis focuses on using TESS to study both the circumstellar environment and stellar interiors, as well as using the satellite to detect and characterize optical emission from gamma-ray bursts (GRBs). Chapter 2 focuses on the discovery of HD 135348, a "rigidly rotating magnetospheric" star–wherein the stellar magnetic field traps dust in a co-rotating orbit and leads to complex periodic photometric modulations–using solely photometric data. Chapter 3 focuses on the discovery of a long-period subdwarf B (sdB) star using 20 s cadence TESS data and proposes a novel formation mechanism for long-period sdB stars that relies upon stable, nonconservative mass transfer. Chapters 4 and 5 focus on pulsating stars in close binaries, and the evolutionary insights that these "tidally tilted" pulsations enable. In particular, we focus on developing models to track the amplitude and phase of these pulsations as a function of orbital phase, as well as tools to perform physically-motivated modeling of the binary components. Chapters 6-7 focus on the optical signatures of gamma-ray bursts in TESS, and analyze the prompt optical flash that is often observed contemporaneously with the high-energy emission from these bursts. Chapter 7, in particular, aims to connect the prompt optical flash to the high-energy spectral energy distribution (SED), and explains the suppression of the optical flash (compared to the extrapolation of the high-energy SED) by invoking dust extinction in the host galaxy. This thesis represents a significant step forward in both stellar and transient astrophysics; throughout this work, we emphasize the use of an unconventional tool–TESS–to pursue timely scientific questions.

Building Intelligence that can Interact with the Physical World

2025-12-04T03:05:25Z

Building Intelligence that can Interact with the Physical World Wang, Tsun-Hsuan (Johnson) Recent advances in Artificial Intelligence (AI) have demonstrated remarkable success in parsing, reasoning, and generating digital content across modalities such as natural language, speech, images, videos, and 3D data. However, these breakthroughs have yet to extend meaningfully beyond the digital realm into the physical world. Developing AI for physical interaction poses challenges such as limited grounding, scarce physical data, and high reliability demands in safety-critical settings. This thesis takes a holistic approach to building intelligence that can interact with the physical world – through the lenses of data, brain, and body. Data is the fuel powering highly capable AI systems. We present methods for data-driven simulation that synthesize sensor measurements from physical processes, and knowledge-driven simulation that leverages large language models to generate actor behaviors and scenarios. By reverse engineering the generative processes behind physical data, we address data scarcity while enabling scalable and effective evaluation. The brain, driven by data, demands a deep understanding of the physical world and reliable interaction with it. We introduce methods to bridge the internet-scale knowledge of digital AI with the physical world to improve generalization and interpretability. For greater reliability, we integrate control-theoretic modules into AI models to enable certifiability. Beyond the behavioral intelligence, the body plays a crucial role in physical interaction. We demonstrate how morphological intelligence can emerge from computation and show how pre-trained generative AI models (brain), when augmented with physics-based simulation that provides feedback on generated data, can be applied to robot design. To this end, this thesis explores how digital AI can be extended into the physical world through a comprehensive investigation of data, brain, and body – laying the groundwork for building physical AI.

Biomolecular Modeling at Scale

2025-12-04T03:05:04Z

Biomolecular Modeling at Scale Wohlwend, Jeremy Predicting the structure and interactions of biomolecules is a fundamental problem in computational biology, with broad implications for disease understanding and drug discovery. Advances in deep learning have enabled remarkable progress, but scaling these approaches to the varied and complex realities of biology is a persistent challenge. This work introduces deep learning methods for biomolecular modeling at scale, designed for efficiency, adaptability, and accessibility. The early chapters present models developed in the general molecular domain, including prediction of structure and interactions for proteins, nucleic acids, and small molecules. To demonstrate how these methods extend to specific biological problems, the latter portion of this work focuses on modeling T cell receptor recognition. As a key immunological mechanism, it highlights the promise of scalable models, but also their present limitations in capturing fine-grained molecular selectivity. Together, these contributions define a framework for bridging foundational models and domain-specific applications, with the potential to scale, and meet the demands of increasingly complex biological systems.

Multimessenger signatures of compact binaries

2025-12-04T03:06:02Z

Multimessenger signatures of compact binaries Mo, Geoffrey Kwan Lok Gravitational waves and electromagnetic observations provide complementary views into some of the most extreme objects in the Universe. In this thesis, I present studies of multimessenger compact binaries from two angles: electromagnetic follow-up of gravitational-waves, and gravitational-wave follow-up of electromagnetic sources. I first describe technical and computational efforts to enable the distribution of alerts of kHz gravitational-wave sources as a member of the LIGO--Virgo--KAGRA collaboration, and to improve localizations of these events by folding in galaxy catalog information. I then detail work to enable electromagnetic follow-up observations of binary neutron star and neutron star--black hole mergers with two telescopes, the Transiting Exoplanet Survey Satellite (TESS) and the Wide-field Infrared Transient Explorer (WINTER). Approaching multimessenger observations from the opposite direction, I describe a search for gravitational waves coincident with fast radio bursts from the only Galactic fast radio burst source. Lastly, I perform an electromagnetic study of Type Ia supernovae in the mid-infrared, whose white dwarf binary progenitors will be mHz gravitational-wave sources for the future LISA space mission.

Efficient Network Systems Design for Machine Learning

2025-12-04T03:05:27Z

Efficient Network Systems Design for Machine Learning Yang, Mingran Machine learning (ML) is transforming modern life by powering a diverse range of groundbreaking applications. As ML models and datasets expand, the scale of training and inference workloads in modern datacenters is increasing at an unprecedented pace. As the demand for computing resources grows, the need for low-latency and energy-efficient network systems becomes increasingly urgent. This thesis introduces efficient network systems designed to support machine learning workloads. It presents three key systems: Trio-ML, which accelerates ML training; Lightning, which enhances ML inference efficiency; and on-fiber photonic computing, a forward-looking vision for next-generation computing systems. The first system, Trio-ML, accelerates data-parallel distributed ML training by leveraging in-network computing on Juniper Networks' programmable chipset Trio. Trio-ML features two key designs: in-network aggregation, which utilizes Trio packet processing threads to aggregate gradients directly inside the network, and in-network straggler mitigation, which utilizes Trio timer threads to detect and address stragglers. We prototype Trio-ML on a testbed with three real DNN models (ResNet50, DenseNet161, and VGG11) to demonstrate its effectiveness in mitigating stragglers while performing in-network aggregation. Our evaluations show that when stragglers occur in the cluster, Trio-ML outperforms today's state-of-the-art in-network aggregation solutions by up to 1.8x. The second system, Lightning, is the first reconfigurable photonic-electronic smartNIC to serve real-time ML inference requests. Lightning uses a fast datapath to feed traffic from the NIC into the photonic domain without creating digital packet processing and data movement bottlenecks. To do so, Lightning leverages a novel reconfigurable count-action abstraction that keeps track of the required computation operations of each inference packet. Our count-action abstraction decouples the compute control plane from the data plane by counting the number of operations in each task and triggers the execution of the next task(s) without interrupting the dataflow. We evaluate Lightning's performance using four platforms: prototype, chip synthesis, emulations, and simulations. Our simulations with large DNN models show that compared to Nvidia A100 GPU, A100X DPU, and Brainwave smartNIC, Lightning accelerates the average inference serve time by 337x, 329x, and 42x, while consuming 352x, 419x, and 54x less energy, respectively. Building on the in-network computing and photonic computing concepts discussed in Trio-ML and Lightning, we present a forward-looking vision for future computing systems. We argue that pluggable transponders are a prime platform for performing photonic computing inside the network without having to replace networking switches and routers. Optical transponders are ubiquitous in today's wide-area and datacenter networks, giving us a unique opportunity to re-purpose them for photonic computing. To this end, we introduce on-fiber photonic computing, explore key research challenges in bringing this vision to reality, and discuss real-world applications.

Wireless, Battery-Free, High-Sensitivity 5G RF Energy Harvesters for Next Generation IoT Sensor Tags

2025-12-04T03:05:17Z

Wireless, Battery-Free, High-Sensitivity 5G RF Energy Harvesters for Next Generation IoT Sensor Tags Yildirim, Deniz Umut The Internet of Things (IoT) is revolutionizing various industries, enabling a new wave of smart applications such as automated asset tracking in warehouses, substation monitoring in smart grids, and precision agriculture. However, as IoT devices proliferate, powering these devices in a sustainable and maintenance-free manner has become a critical challenge. Traditional IoT systems rely on batteries, which present issues of limited lifespan, environmental impact, and maintenance costs, especially in large-scale deployments. As a result, the development of battery-free IoT devices powered by ambient energy harvesting has gained significant attention. Among various energy-harvesting technologies, radio frequency (RF) energy harvesting has emerged as a promising solution for powering IoT devices. By harvesting energy from ambient RF signals in licensed frequency bands, RF energy-harvesting systems eliminate the need for batteries and allow for continuous, maintenance-free operation. This is especially crucial in environments where battery replacement is impractical or impossible, such as in large industrial warehouses, remote infrastructure, and hazardous environments. However, achieving high sensitivity and reliable operation in RF energy-harvesting systems poses several challenges. High-sensitivity rectifiers are required to capture and convert weak RF signals into usable energy, but integrating these rectifiers with ultra-low power baseband data processing circuits remains a significant hurdle. Moreover, antenna-rectifier matching calibration must be compatible with the duty-cycled operation of these tags, where brief communication periods are followed by long charging intervals. Additionally, the antenna system must be robust to detuning when placed on various objects, ensuring that the system can operate effectively in diverse environments. This thesis presents two integrated circuits to work towards these goals. The first chip is designed with the goal of minimizing the charging time as much as possible, which is critical in scenarios such as inventory management in warehouses, and tamper detection. The goal was to achieve < 1-minute charging time while maintaining sensitivity competitive with the state-of-the-art. Unlike previous harvesters that either focused solely on sensitivity without integrating baseband processing and communication, or included those features but considered continuous communication at low sensitivity, the IC developed in this work achieves a sensitivity of −31 dBm and is capable of backscattering data approximately 18 seconds after a cold start. It also provides a detailed description of the difficulty of achieving higher sensitivities at higher 5G frequencies. The second chip in this thesis builds upon the first one and integrates an analog front-end to convert sensor data for environmental monitoring. We implemented an antenna-rectifier calibration method that is maintained as long as there if RF power, even though the tag goes into long charging periods. Even though the charging time, or the data readout interval, for these tags is more relaxed compared to the inventory management applications, we have also developed a design methodology to minimize the energy required to generate a data packet for backscattering, through which we were able to keep the charging time at 4 minutes while having additional functionalities and backscattering at a higher data rate compared to the first chip. Finally, a simple shielding method was implemented to enable the tags to be placed on any objects without resonance frequency detuning. All of these were achieved while still obtaining a sensitivity of −30 dBm, competitive with the state of the art. In addition, the third project investigates the use of heterogeneously integrated “beyondCMOS” devices to enhance overall rectifier performance. These emerging devices, fabricated by Palacios Group at MIT, show promise in overcoming sensitivity limitations commonly found in rectifiers, thereby extending the range and coverage of energy-harvesting IoT systems. We conduct a detailed characterization of these devices, highlighting their unique physical behaviors not present in standard CMOS technology, and provide system-level design guidelines for building improved rectifiers. Preliminary simulation results show that rectifiers using negative-capacitance field-effect transistors (NCFETs) can harvest up to four times as much power than their CMOS-based counterparts, while maintaining the same sensitivity. This thesis outlines the design, implementation, and evaluation of all three systems. The two aforementioned ICs are tested both in simulation and in real-world scenarios such as a typical office environment. Meanwhile, the novel device technologies are explored through simulation, demonstrating their significant potential for next-generation rectifier design.

Analysis of Driver and Pedestrian Gesture Use in the Boston Area. Automated Vehicles May Need More Than Kinematics in Ambiguous Situations

2025-12-04T03:13:50Z

Analysis of Driver and Pedestrian Gesture Use in the Boston Area. Automated Vehicles May Need More Than Kinematics in Ambiguous Situations Weibert, Alexander; Manstetten, Dietrich; Reimer, Bryan; Gershon, Pnina; Mehler, Bruce; Abdenebaoui, Larbi; Hatice Şahin, İppoliti Roadways, despite their formal regulations, are dynamic spaces where humans interact beyond formal rules to resolve conflicts. In ambiguous situations, the right of way is often unclear. Self-driving vehicles in urban traffic introduce challenges to their coexistence with humans, indicating a need for greater social awareness in these vehicles. To investigate social interactions among roadway users, we analyzed a naturalistic driving dataset focusing on instances where drivers yielded to pedestrians, by noting gestures. Video analysis showed that gestures were more common in ambiguous situations than in regulated scenarios. Drivers used gestures to navigate the right of way efficiently, while pedestrians used them to express gratitude. These findings highlight the importance of understanding social expressions in designing socially aware self-driving vehicles. AutomotiveUI Adjunct ’25, Brisbane, QLD, Australia

Characterizing and Optimizing Realistic Workloads on a Commercial Compute-in-SRAM Device

2025-12-04T03:13:46Z

Characterizing and Optimizing Realistic Workloads on a Commercial Compute-in-SRAM Device Zhang, Niansong; Zhu, Wenbo; Golden, Courtney; Ilan, Dan; Chen, Hongzheng; Batten, Christopher; Zhang, Zhiru Compute-in-SRAM architectures offer a promising approach to achieving higher performance and energy efficiency across a range of data-intensive applications. However, prior evaluations have largely relied on simulators or small prototypes, limiting the understanding of their real-world potential. In this work, we present a comprehensive performance and energy characterization of a commercial compute-in-SRAM device, the GSI APU, under realistic workloads. We compare the GSI APU against established architectures, including CPUs and GPUs, to quantify its energy efficiency and performance potential. We introduce an analytical framework for general-purpose compute-in-SRAM devices that reveals fundamental optimization principles by modeling performance trade-offs, thereby guiding program optimizations. Exploiting the fine-grained parallelism of tightly integrated memory-compute architectures requires careful data management. We address this by proposing three optimizations: communicationaware reduction mapping, coalesced DMA, and broadcast-friendly data layouts. When applied to retrieval-augmented generation (RAG) over large corpora (10GB–200GB), these optimizations enable our compute-in-SRAM system to accelerate retrieval by 4.8×–6.6× over an optimized CPU baseline, improving end-to-end RAG latency by 1.1×–1.8×. The shared off-chip memory bandwidth is modeled using a simulated HBM, while all other components are measured on the real compute-in-SRAM device. Critically, this system matches the performance of an NVIDIA A6000 GPU for RAG while being significantly more energy-efficient (54.4×-117.9× reduction). These findings validate the viability of compute-in-SRAM for complex, real-world applications and provide guidance for advancing the technology. MICRO ’25, Seoul, Republic of Korea

Voice to Vision: A Sociotechnical System for Transparent Civic Decision-Making

2025-12-04T03:13:39Z

Voice to Vision: A Sociotechnical System for Transparent Civic Decision-Making Hughes, Margaret; Overney, Cassandra; Kamra, Ashima; Tepale, Jasmin; Hamby, Elizabeth; Jasim, Mahmood; Roy, Deb Communities frequently report sending feedback “into a void” during community engagement processes like neighborhood planning, creating a critical disconnect between public input and decision-making. Voice to Vision addresses this gap with a sociotechnical system that comprises three integrated components: a flexible data architecture linking community input to planning outputs, a sensemaking interface for planners to analyze and synthesize feedback, and a community-facing platform that makes the entire engagement process transparent. By creating a shared information space between stakeholders, our system demonstrates how structured data and specialized interfaces can foster cooperation across stakeholder groups, while addressing tensions in accessibility and trust formation. Our CSCW demonstration will showcase this system’s ability to transform opaque civic decision-making processes into collaborative exchanges, inviting feedback on its potential applications beyond urban planning. CSCW Companion ’25, Bergen, Norway

Augmenting Collaborative Problem-Solving: Exploring the Design and Use of GenAI for Groupwork

2025-12-04T03:13:44Z

Augmenting Collaborative Problem-Solving: Exploring the Design and Use of GenAI for Groupwork Johnson, Janet; Rick, Steven; Gr?nb?k, Jens Emil; Wong, Emily; Yin, Ming; Nebeling, Michael; Klein, Mark; Ackerman, Mark; Malone, Thomas Complex problem-solving and creative work in the real world are rarely individual endeavors and typically unfold within teams and group settings. While advancements in generative artificial intelligence (GenAI) have shown promise in augmenting creativity and productivity, these tools are primarily designed for individual use and overlook group dynamics and the collaborative aspects of teamwork. This workshop will provide a platform for researchers and practitioners to explore the design of future human-AI groups across four key themes: (1) the role of GenAI in group settings, (2) collaborative and multimodal interactions with GenAI, (3) evaluating GenAI’s influence within groups and designing for appropriate reliance, and (4) evolving group practices in the presence of GenAI. We hope to build a community and construct alignment across participants around how to pursue research that understands how GenAI can augment, undermine, or bring new practices to collaborative settings and groupwork. CSCW Companion ’25, Bergen, Norway

Development of systematic uncertainty-aware neural network trainings for binned-likelihood analyses at the LHC

2025-12-04T03:14:35Z

Development of systematic uncertainty-aware neural network trainings for binned-likelihood analyses at the LHC CMS Collaboration We propose a neural network training method capable of accounting for the effects of systematic variations of the data model in the training process and describe its extension towards neural network multiclass classification. The procedure is evaluated on the realistic case of the measurement of Higgs boson production via gluon fusion and vector boson fusion in the τ τ decay channel at the CMS experiment. The neural network output functions are used to infer the signal strengths for inclusive production of Higgs bosons as well as for their production via gluon fusion and vector boson fusion. We observe improvements of 12 and 16% in the uncertainty in the signal strengths for gluon and vector-boson fusion, respectively, compared with a conventional neural network training based on cross-entropy.

Designs Related Through Projective and Hopf Maps

2025-12-04T03:14:37Z

Designs Related Through Projective and Hopf Maps Lindblad, Ayodeji We verify a construction which, for K the reals, complex numbers, quaternions, or octonions, builds a spherical t-design by placing a spherical t-design on each K -projective or K -Hopf fiber associated to the points of a ⌊ t / 2 ⌋ -design on a quotient projective space K P n ≠ O P 2 or sphere. This generalizes work of König and Kuperberg, who verified the K = C case of the projective settings, and of Okuda, who (inspired by independent observation of this construction by Cohn, Conway, Elkies, and Kumar) verified the K = C case of the generalized Hopf settings.

A generative deep learning approach to de novo antibiotic design

2025-12-03T06:24:53Z

A generative deep learning approach to de novo antibiotic design Krishnan, Aarti; Anahtar, Melis N.; Valeri, Jacqueline A.; Jin, Wengong; Donghia, Nina M.; Sieben, Leif; Luttens, Andreas; Zhang, Yu; Modaresi, Seyed Majed; Hennes, Andrew; Fromer, Jenna; Bandyopadhyay, Parijat; Chen, Jonathan C.; Rehman, Danyal; Desai, Ronak; Edwards, Paige; Lach, Ryan S.; Aschtgen, Marie-Stéphanie; Gaborieau, Margaux; Gaetani, Massimiliano; Palace, Samantha G.; Omori, Satotaka; Khonde, Lutete; Moroz, Yurii S.; Blough, Bruce; Jin, Chunyang; Loh, Edmund; Grad, Yonatan H.; Saei, Amir Ata; Coley, Connor W.; Wong, Felix; Collins, James J. The antimicrobial resistance crisis necessitates structurally distinct antibiotics. While deep learning approaches can identify antibacterial compounds from existing libraries, structural novelty remains limited. Here, we developed a generative artificial intelligence framework for designing de novo antibiotics through two approaches: a fragment-based method to comprehensively screen >107 chemical fragments in silico against Neisseria gonorrhoeae or Staphylococcus aureus, subsequently expanding promising fragments, and an unconstrained de novo compound generation, each using genetic algorithms and variational autoencoders. Of 24 synthesized compounds, seven demonstrated selective antibacterial activity. Two lead compounds exhibited bactericidal efficacy against multidrug-resistant isolates with distinct mechanisms of action and reduced bacterial burden in vivo in mouse models of N. gonorrhoeae vaginal infection and methicillin-resistant S. aureus skin infection. We further validated structural analogs for both compound classes as antibacterial. Our approach enables the generative deep-learning-guided design of de novo antibiotics, providing a platform for mapping uncharted regions of chemical space.

Frontiers of biological material intelligence

2025-12-03T06:25:07Z

Frontiers of biological material intelligence Marom, Lee; Buehler, Markus J. Biological materials exhibit a form of intelligence that enables them to sense, adapt, and self-optimize in response to their environments. Unlike synthetic materials, which are often designed for singular, static functions, natural material systems integrate sensing, memory, and feedback directly into their architectures. As industries face increasing demands for resilience, sustainability, and efficiency, the development of intelligent materials has become a promising step toward the future of material innovation. Advances in artificial intelligence and machine learning, along with mathematical frameworks spanning graph theory and category theory, provide powerful tools to uncover the underlying design principles of intelligent biological materials. Simultaneously, digital fabrication methods, including additive manufacturing and biofabrication, allow the scalable realization of adaptive material systems. As the integration of deep biological insight, computational modeling, and advanced fabrication continues to evolve, it sets the stage for a profound shift in how we conceive, create, and deploy materials. Advancing this convergence will accelerate the development of intelligent systems that are capable of autonomous adaptation, long-term resilience, and embedded functionality across scales and environments.

Leveraging community engagement and human-centered design to develop multilevel implementation strategies to enhance adoption of a health equity intervention

2025-12-03T06:25:10Z

Leveraging community engagement and human-centered design to develop multilevel implementation strategies to enhance adoption of a health equity intervention Price, Maggi A.; Mulkern, Patrick J.; Condon, Madelaine; Rakhilin, Marina; Johansen, Kara; Lyon, Aaron R.; Saldana, Lisa; Pachankis, John; Woodward, Sue A.; Roeder, Kathryn M.; Moran, Lyndsey R.; Jerskey, Beth A. Background Health equity intervention implementation (which promotes positive health outcomes for populations experiencing disproportionately worse health) is often impeded by health-equity-specific barriers like provider bias; few studies demonstrate how to overcome these barriers through implementation strategies. An urgent health equity problem in the U.S. is the mental health of transgender youth. To address this, we developed Gender-Affirming Psychotherapy (GAP), a health equity intervention comprising best-practice mental health care for transgender youth. This paper details the identification of implementation determinants and the development of targeted strategies to promote provider adoption of GAP. Methods This study represents part of a larger study of mental health provider adoption of GAP. Here we describe the first 2 stages of the 3-stage community-engaged and human-centered design process – Discover, Design/Build, and Test – to identify implementation determinants of adoption and develop implementation strategies with transgender youth, their parents, and mental health providers. This process involved collecting data via focus groups, design meetings, usability testing, and champion meetings. Data were analyzed using rapid and conventional content analysis. Qualitative coding of implementation determinants was guided by the Health Equity Implementation Framework, and implementation strategy coding was facilitated by the ERIC Implementation Strategy Compilation. Results We identified 15 determinants of GAP adoption, and all were specific to the transgender population (e.g., inclusive record system, anti-transgender attitudes). Seventeen implementation strategies were recommended and 12 were developed, collectively addressing all identified determinants. Most strategies were packaged into an online self-paced mental health provider training (implementation intervention) with 6 training tools. Additional inner-setting strategies were designed to support training uptake (e.g., mandate training) and GAP adoption (e.g., change record system). Conclusions Community-engaged and human-centered design methods can identify health equity intervention implementation determinants and develop targeted strategies. We highlight five generalizable takeaways for health equity implementation scientists: (1) implementer bias may be a key barrier, (2) experience with the health equity population may be an important facilitator, (3) stakeholder stories may be an effective training tool, (4) inner-setting-level implementation strategies may be necessary, and (5) teaching implementers how to build implementation strategies can overcome resource-constraints. Trial registration November 11, 2022; NCT05626231.

Breeding of microbiomes conferring salt tolerance to plants

2025-12-03T06:25:08Z

Breeding of microbiomes conferring salt tolerance to plants Guilherme Pereira, Caio; Edwards, Joseph A.; Khasanova, Albina; Carlson, Alexis; Brisson, Vanessa; Schaefer, Estelle; Glavina del Rio, Tijana; Tringe, Susannah; Vogel, John P.; Des Marais, David L.; Juenger, Thomas E.; Mueller, Ulrich G. Background Microbiome breeding through host-mediated selection is a technique to artificially select for microbiomes conferring beneficial properties to plants. Using a systematic selection protocol that maximises the heritability of microbiome effects, transmission fidelity, and microbiome stability through multiple selection cycles, we previously developed root-associated microbial communities conferring sodium and aluminium tolerance to Brachypodium distachyon, a model for cereal crops. Here, we explore the physiological mechanisms underlying our selected microbiomes’ effect on plant fitness and analyse how our selection protocol shaped the composition and structure of these microbiomes. We analysed the effects of our selected microbiomes on plant fitness and tissue-nutrient concentration, then used 16S rRNA amplicon sequencing to examine microbial community composition and co-occurrence network patterns. Results Our sodium-selected microbiomes reduced leaf sodium concentration by ~ 50%, whereas the aluminium-selected microbiomes had no effect on leaf-tissue nutrient concentration, suggesting different mechanisms underlying the microbiome-mediated stress tolerance. By testing the selected microbiomes in a cross-fostering experiment, we show that our artificially selected microbiomes attained (a) ecological robustness contributing to transplantability (i.e. inheritance) of microbiome-encoded effects between plants; and (b) network features identifying key bacteria promoting salt-stress tolerance. Conclusions Combined, these findings elucidate critical mechanisms underlying host-mediated artificial selection as a framework to breed microbiomes with targeted benefits for plants under salt stresses, with significant implications for sustainable agriculture.

Measurement of ψ(2S) to J/ψ cross-section ratio as function of multiplicity in pPb collisions at √sNN = 8.16 TeV

2026-03-08T03:32:05Z

Measurement of ψ(2S) to J/ψ cross-section ratio as function of multiplicity in pPb collisions at √sNN = 8.16 TeV Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J.; Alessio, F.; The LHCb collaboration The production ratio of ψ(2S) to J/ψ charmonium states is presented as a function of multiplicity in proton-lead collisions at a centre-of-mass energy of s NN = 8.16 TeV, for both prompt and nonprompt sources. The total luminosity recorded by the LHCb experiment corresponds to 13.6 nb−1 for pPb collisions and 20.8 nb−1 for Pbp collisions, where the first particle corresponds to the particle traveling towards the detector. Measurements are performed in the dimuon final state at forward (backward) centre-of-mass rapidity, with respect to the proton direction, 1.5 < y* < 4.0 (−5.0 < y* < −2.5) for pPb (Pbp) collisions. A multiplicity dependence of the prompt production ratio is observed in pPb collisions, whereas no dependence is found in nonprompt production, nor in either prompt or nonprompt production in Pbp collisions. These results suggest that in the Pb-going direction additional suppression mechanisms beyond comover effects may be present, possibly related to the formation of quark-gluon plasma. This highlights a transition from small to large collision systems and provides important insight into the suppression of charmonia in proton-nucleus collisions.

CURENet: combining unified representations for efficient chronic disease prediction

2026-03-08T03:32:04Z

CURENet: combining unified representations for efficient chronic disease prediction Dao, Cong-Tinh; Phan, Nguyen M. T.; Ding, Jun-En; Wu, Chenwei; Restrepo, David; Luo, Dongsheng; Zhao, Fanyi; Liao, Chun-Chieh; Peng, Wen-Chih; Wang, Chi-Te; Chen, Pei-Fu; Chen, Ling; Ju, Xinglong; Liu, Feng; Hung, Fang-Ming Electronic health records (EHRs) are designed to synthesize diverse data types, including unstructured clinical notes, structured lab tests, and time-series visit data. Physicians draw on these multimodal and temporal sources of EHR data to form a comprehensive view of a patient’s health, which is crucial for informed therapeutic decision-making. Yet, most predictive models fail to fully capture the interactions, redundancies, and temporal patterns across multiple data modalities, often focusing on a single data type or overlooking these complexities. In this paper, we present CURENet, a multimodal model (Combining Unified Representations for Efficient chronic disease prediction) that integrates unstructured clinical notes, lab tests, and patients’ time-series data by utilizing large language models (LLMs) for clinical text processing and textual lab tests, as well as transformer encoders for longitudinal sequential visits. Curenet has been capable of capturing the intricate interaction between different forms of clinical data and creating a more reliable predictive model for chronic illnesses. We evaluated CURENet using the public MIMIC-III and private FEMH datasets, where it achieved over 94% accuracy in predicting the top 10 chronic conditions in a multi-label framework. Our findings highlight the potential of multimodal EHR integration to enhance clinical decision-making and improve patient outcomes.

Effective field theory factorization for diffraction

2026-03-08T03:32:03Z

Effective field theory factorization for diffraction Lee, Kyle; Schindler, Stella T.; Stewart, Iain W. We derive a factorization formula for coherent and incoherent ep diffraction using the soft collinear effective theory, utilizing multiple power expansion parameters to handle different kinematic regions. This goes beyond the known hard-collinear diffractive factorization to address the small-x Regge dynamics and Pomeron exchange from first principles. The effective field theory analysis also uncovers and factorizes an important irreducible incoherent background generated by color-nonsinglet exchange, dubbed “quasi-diffraction”, for which we calculate the associated Sudakov suppression. For unpolarized scattering we show that there are four diffractive structure functions at leading power, and point out the importance of studying F 3 , 4 D through asymmetries, in addition to F 2 , L D . For the quasi-diffractive background, we make model independent predictions for ratios of the corresponding structure functions in a perturbative kinematic region. Our analysis also makes predictions for six leading-power spin-dependent structure functions. Finally, we provide connections to diffractive parton distributions, and assess the Ingelman-Schlein model. Our work lays a path for further QCD-based studies of diffraction.

Coherent photoproduction of ρ0, ω and excited vector mesons in ultraperipheral PbPb collisions

2026-03-08T03:32:06Z

Coherent photoproduction of ρ0, ω and excited vector mesons in ultraperipheral PbPb collisions Aaij, R.; Abdelmotteleb, A. S. W.; Abellan Beteta, C.; Abudinén, F.; Ackernley, T.; Adefisoye, A. A.; Adeva, B.; Adinolfi, M.; Adlarson, P.; Agapopoulou, C.; Aidala, C. A.; Ajaltouni, Z.; Akar, S.; Akiba, K.; Albicocco, P.; Albrecht, J. The invariant-mass distribution for the coherent photoproduction of dipions in ultraperipheral PbPb collisions is measured using data, corresponding to an integrated luminosity of 224.6 ± 9.6μb−1, collected by the LHCb experiment in 2018 at a nucleon-nucleon centre-of-mass energy s NN = 5.02 TeV. In the mass range from 400 to 1200 MeV, the results are consistent with previous experiments, with the spectrum dominated by the ρ0 meson, which interferes with a nonresonant component, together with a smaller ω meson contribution. In an extended mass range up to 2300 MeV, models previously used do not fit the data and a consistent description requires the introduction of two resonances at masses of 1350 ± 20 MeV and 1790 ± 20 MeV with widths of about 300 MeV. The cross-section for each meson is measured differentially in twelve bins of rapidity from 2.05 to 4.90. The ρ0 cross-section increases with rapidity from about 400 to 600 mb and is measured with a typical precision of 8%, while the cross-section times branching fraction for the ω, ρ′ and ρ′′, with the statistical precision of the data, do not have a pronounced rapidity dependence and are between 0.5 and 1.5mb, with uncertainties up to 30%. A large nuclear suppression is observed for the ρ0 meson compared to expectations based on photoproduction on the proton that use the impulse approximation. Significant suppression is also observed compared to that predicted by elastic scattering described in the Glauber approach, or with the addition of inelastic scattering in a Gribov-Glauber model.

Forcing with Invariant Measures

2026-03-08T03:32:08Z

Forcing with Invariant Measures Ackerman, Nathanael; Freer, Cameron; Golshani, Mohammad; Mirabi, Mostafa; Patel, Rehana This paper introduces a model-theoretic generalization of the notion of forcing with random reals, in which forcing gives rise to random generic structures. Specifically, we consider forcing with κ -Borel probability measures on the space of L -structures with a (possibly uncountable) infinite set X, focusing on those that are invariant under the action of the symmetric group Sym ( X ) . We demonstrate how any Sym ( X ) -invariant measure where X is countable can be uniquely extended to a Sym ( Y ) -invariant measure where Y is uncountable, and prove that forcing with such measures satisfies the countable chain condition. We also show that we can uniformly distinguish between these random generic structures and the Cohen generic structures that arise from forcing with a strong Fraïssé class: There is a κ -Borel set of low complexity that contains every Cohen generic structure that is not highly homogeneous but contains no random generic structure, implying that a structure that is not highly homogeneous cannot be both Cohen generic and random generic. Finally, we answer an open question of Kostana in the case of ω 1 , by establishing a connection between forcing with a strong Fraïssé class and Cohen forcing.

Gnotobiotic growth and phosphorus limitation of Arabidopsis thaliana and co-occurring microbes on phosphated iron oxides

2026-03-08T03:31:52Z

Gnotobiotic growth and phosphorus limitation of Arabidopsis thaliana and co-occurring microbes on phosphated iron oxides Mackie, Amanda M.; Schuler, Christopher J.; McRose, Darcy L. The macronutrient phosphorus is vital for sustaining cellular processes in all life forms. Due to its frequent adsorption on iron minerals, phosphorus bioavailability is low in many soils. While the abiotic adsorption of phosphate on iron minerals has been well studied, the direct effects of this process on bioavailability to plants and microbes has not been thoroughly investigated in a simplified laboratory system. We developed a hydroponic growth system that uses hydrous ferric oxide (HFO) to induce phosphorus limitation and can enable both plant and microbial cultivation as well as gnotobiotic co-culture. We demonstrate that this system can be used for phosphorus-limited growth of the model plant Arabidopsis thaliana as well as two root-associated bacterial isolates (from the genera Rhizobium and Pseudomonas). Elemental analysis of phosphorus and iron concentration in A. thaliana shoots reveals that the addition of increasing amounts of HFO leads to a progressive decrease in phosphorus concentration but does not affect iron quotas. We also report that phosphorus concentrations in both bacterial isolates decrease when cultivated in media supplemented with HFO. We further show that A. thaliana can be co-cultured with a Rhizobium isolate in our phosphorus-limited hydroponic system with bacteria relying on plant photosynthate as their sole carbon source. Our work provides a controlled demonstration of the effects of mineral adsorption on phosphorus bioavailability and a tool for further investigation of how plants and microbes access phosphorus in the environment.

Beyond submodular maximization via one-sided smoothness

2026-03-08T03:32:05Z

Beyond submodular maximization via one-sided smoothness Ghadiri, Mehrdad; Santiago, Richard; Shepherd, Bruce The multilinear framework for submodular maximization was developed to achieve a tight 1 - 1 / e approximation for maximizing a monotone submodular function subject to a matroid constraint, including as special case the submodular welfare problem. The framework has a continuous optimization step (solving the multilinear extension of a submodular function) and a rounding part (rounding a fractional solution to an integral one). We extend both parts to provide a framework for a wider array of applications. The continuous part works for a more general class of continuous functions parameterized by a new smoothness parameter σ . A twice differential function F is called σ -one-sided-smooth ( σ -OSS) if its second derivatives are bounded as follows: 1 2 u T ∇ 2 F ( x ) u ≤ σ · ‖ u ‖ 1 ‖ x ‖ 1 u T ∇ F ( x ) for all u , x ≥ 0 , x ≠ 0 . For σ = 0 this includes previously studied continuous DR-Submodular functions as well as quadratics defined by copositive matrices. We give a modification of the continuous greedy algorithm which finds a solution for maximizing a monotone σ -OSS F over a polytope in the non-negative orthant; the solution approximates the optimum to within factors which are functions of σ which depend on additional properties. Interestingly, σ -OSS functions arise as the multilinear extensions of set functions associated with several well-studied diversity maximization problems: max f ( S ) = ∑ i , j ∈ S A ij : | S | ≤ k . For instance, when A ij defines a σ -semi-metric, its extension is σ -OSS. In these settings, we also develop rounding schemes to approximate the discrete problem.

Report to the President for year ended June 30, 2025, Dean, School of Engineering

2025-12-02T03:09:52Z

Report to the President for year ended June 30, 2025, Dean, School of Engineering Gallagher, Mary Beth This report contains the following sections: Administrative Initiatives, Personnel Information, Educational Activities, Strategic Initiatives, Entrepreneurship, Leadership, and Innovation Activities.

Report to the President for year ended June 30, 2025, MIT Washington Office

2025-12-02T03:09:51Z

Report to the President for year ended June 30, 2025, MIT Washington Office Zuber, Maria This report contains the following sections: Personnel, Communications, Federal advocacy, Priority areas, MIT in DC, and Student engagement and mentorship.

Report to the President for year ended June 30, 2025, Department of Brain and Cognitive Sciences

2025-12-02T03:09:47Z

Report to the President for year ended June 30, 2025, Department of Brain and Cognitive Sciences Fee, Michale This report contains the following sections: Introduction: Our Mission and Approach, The Building 46 Community, Strategic Planning, Leadership, Faculty, Research Centers, Academics, Finances and Funding, and Research Highlights.

Shotgun Metagenomics of Gastric Biopsies Reveals Compositional and Functional Microbiome Shifts in High- and Low-Gastric-Cancer-Risk Populations from Colombia, South America

2026-03-08T03:32:01Z

Shotgun Metagenomics of Gastric Biopsies Reveals Compositional and Functional Microbiome Shifts in High- and Low-Gastric-Cancer-Risk Populations from Colombia, South America Mannion, Anthony; Sheh, Alexander; Shen, Zeli; Dzink-Fox, JoAnn; Piazuelo, M Blanca; Wilson, Keith T; Peek, Richard; Fox, James G Along with Helicobacter pylori infection, the gastric microbiota is hypothesized to modulate stomach cancer risk in susceptible individuals. Whole metagenomic shotgun sequencing (WMS) is a sequencing approach to characterize the microbiome with advantages over traditional culture and 16S rRNA sequencing including identification of bacterial and non-bacterial taxa, species/strain resolution, and functional characterization of the microbiota. In this study, we used WMS to survey the microbiome in extracted DNA from antral gastric biopsy samples from Colombian patients residing in the high-risk gastric cancer town Túquerres (n = 10, H. pylori-positive = 7) and low-risk town of Tumaco (n = 10, H. pylori-positive = 6). Kraken2/Bracken was used for taxonomic classification and abundance. Functional gene profiles were inferred by InterProScan and KEGG analysis of assembled contigs and gene annotation. The most abundant taxa represented bacteria, non-human eukaryota, and viral genera found in skin, oral, food, and plant/soil environments including Staphylococus, Streptococcus, Bacillus, Aspergillus, and Siphoviridae. H. pylori was the predominant taxa present in H. pylori-positive samples. Beta diversity was significantly different based on H. pylori-status, risk group, and sex. WMS detected more bacterial taxa than 16S rRNA sequencing and aerobic, anaerobic, and microaerobic culture performed on the same gastric biopsy samples. WMS identified significant differences in functional profiles found between H. pylori-status, but not risk or sex groups. H. pylori-positive samples were significantly enriched for H. pylori-specific genes including virulence factors such as vacA, cagA, and urease, while carbohydrate and amino acid metabolism genes were enriched in H. pylori-negative samples. This study shows WMS has the potential to characterize the taxonomy and function of the gastric microbiome as risk factors for H. pylori-associated gastric disease. Future studies will be needed to compare and validate WMS versus traditional culture and 16S rRNA sequencing approaches for characterization of the gastric microbiome.

Resonance Scattering Treatment with the Windowed Multipole Formalism

2026-03-08T03:32:00Z

Resonance Scattering Treatment with the Windowed Multipole Formalism Ridley, Gavin; Forget, Benoit; Burke, Timothy A new method for directly sampling the resonance upscattering effect is presented. Alternatives have relied on inefficient rejection sampling techniques or large tabular storage of relative velocities. None of these approaches, which require pointwise energy data, are particularly well suited to the windowed multipole cross-section representation. The new method, called multipole analytic resonance scattering, overcomes these limitations by inverse transform sampling from the target relative velocity distribution where the cross section is expressed in the multipole formalism. The closed-form relative speed distribution contains a novel special function we deem the incomplete Faddeeva function, and we present the first results on its efficient numerical evaluation.

Assessing the Structure-Based Turbulence Model Performance for Thermal Striping Applications Using Symmetric Jet Experiments

2026-03-08T03:31:59Z

Assessing the Structure-Based Turbulence Model Performance for Thermal Striping Applications Using Symmetric Jet Experiments Pham, Monica; Petrov, Victor; Manera, Annalisa; Baglietto, Emilio Turbulent mixing of coolant streams can result in an oscillatory mixing phenomenon called thermal striping. These fluctuations have the potential to lead to anticipated thermal fatigue failures in advanced nuclear reactors. To predict thermal striping, robust and computationally affordable modeling tools that are capable of accurately representing complex turbulence are needed. Hybrid turbulence approaches, such as detached-eddy simulation and scale-adaptive simulation, have shown some success in resolving complex unsteady turbulence for massively separated flows, however the applicability of these models to internal flows is limited. A STRUCTure-based (STRUCT) second-generation Unsteady Reynolds-Averaged Navier–Stokes turbulence model was recently proposed at the Massachusetts Institute of Technology to robustly extend the applicability of hybrid closures. In this work, the STRUCT model is evaluated using experimental data taken at the Reactor Cavity Cooling System separate-effects test facility at the University of Michigan. The experiments observed the interaction of parallel symmetric rectangular jets, and include measurements for mean profiles of velocity and Reynolds stresses. In the present work, the simulation results are assessed against mean profiles of velocity and Reynolds stresses, demonstrating the ability to reproduce the unsteadiness of the jets in close agreement with the measurements at considerably reduced computational cost.

Deep-learning models for forecasting financial risk premia and their interpretations

2026-03-08T03:32:02Z

Deep-learning models for forecasting financial risk premia and their interpretations Lo, Andrew W; Singh, Manish The measurement of financial risk premia, the amount that a risky asset will outperform a risk-free one, is an important problem in asset pricing. The noisiness and non-stationarity of asset returns makes the estimation of risk premia using machine learning (ML) techniques challenging. In this work, we develop ML models that solve the problems associated with risk premia forecasting by separating risk premia prediction into two independent tasks, a time series model and a cross-sectional model, and using neural networks with skip connections to enable their deep neural network training. These models are tested robustly with different metrics, and we observe that our models outperform several existing standard ML models. A known issue with ML models is their ‘black box’ nature, i.e. their opaqueness to interpretability. We interpret these deep neural networks using local approximation-based techniques that provide explanations for our model's predictions.

Machine Learning Distinguishes Plant Bioelectric Recordings with and Without Nearby Human Movement

2026-03-08T03:31:52Z

Machine Learning Distinguishes Plant Bioelectric Recordings with and Without Nearby Human Movement Gloor, Peter A.; Weinbeer, Moritz Background: Quantitatively detecting whether plants exhibit measurable bioelectric differences in the presence of nearby human movement remains challenging, in part because plant signals are low-amplitude, slow, and easily confounded by environmental factors. Methods: We recorded bioelectric activity from 2978 plant samples across three species (basil, salad, tomato) using differential electrode pairs (leaf and soil electrodes) sampling at 142 Hz. Two trained performers executed three specific eurythmic gestures near experimental plants while control plants remained isolated. Random Forest and Convolutional Neural Network classifiers were applied to distinguish the control from treatment conditions using engineered features including spectral, temporal, wavelet, and frequency domain characteristics. Results: Random Forest classification achieved 62.7% accuracy (AUC = 0.67) distinguishing differences in recordings collected near a moving human from control conditions, representing a statistically significant 12.7 percentage point improvement over chance. Individual performer signatures were detectable with 68.2% accuracy, while plant species classification achieved only 44.5% accuracy, indicating minimal species-specific artifacts. Temporal analysis revealed that the plants with repeated exposure exhibited consistently less negative bioelectric amplitudes compared to single-exposure plants. Innovation: We introduce a data-driven approach that pairs standardized, short-window bioelectric recordings with machine-learning classifiers (Random Forest, CNN) to test, in an exploratory manner, whether plant signals differ between human-moving-nearby and isolation conditions. Conclusions: Plants exhibit modest but statistically detectable bioelectric differences in the presence of nearby human movement. Rather than attributing these differences to eurythmic movement itself, the present design can only demonstrate that plant recordings collected within ~1 m of a moving human differ, modestly but statistically, from recordings taken ≥3 m away. The underlying biophysical pathways and specific contributing factors (airflow, VOCs, thermal plumes, vibration, electromagnetic fields) remain unknown. These results should therefore be interpreted as exploratory correlations, not mechanistic evidence of gesture-specific plant sensing.

Analysis of Regional Surface CO2 Fluxes Using the MEGA Satellite Data Assimilation System

2026-03-08T03:32:07Z

Analysis of Regional Surface CO2 Fluxes Using the MEGA Satellite Data Assimilation System Hu, Liting; Hu, Xiaoyi; Jiang, Fei; He, Wei; Deng, Zhu; Fang, Shuangxi; Fang, Xuekun Understanding the dynamics of terrestrial carbon sources and sinks is crucial for addressing climate change, yet significant uncertainties remain at regional scales. We developed the Monitoring and Evaluation of Greenhouse gAs Flux (MEGA) inversion system with satellite data assimilation and applied it to China using OCO-2 V11.1r XCO2 retrievals. Our results show that China’s terrestrial ecosystems acted as a carbon sink of 0.28 ± 0.15 PgC yr−1 during 2018–2023, consistent with other inversion estimates. Validation against surface CO2 flask measurements demonstrated significant improvement, with RMSE and MAE reduced by 30%–46% and 24–44%, respectively. Six sets of prior sensitivity experiments conclusively demonstrated the robustness of MEGA. In addition, this study is the first to systematically compare model-derived and observation-based background fields in satellite data assimilation. Ten sets of background sensitivity experiments revealed that model-based background fields exhibit superior capability in resolving seasonal flux dynamics, though their performance remains contingent on three key factors: (1) initial fields, (2) flux fields, and (3) flux masks (used to control regional flux switches). These findings highlight the potential for further refinement of the atmospheric inversion system.

ZeoSyn: A Comprehensive Zeolite Synthesis Dataset Enabling Machine-Learning Rationalization of Hydrothermal Parameters

2025-11-27T05:20:45Z

ZeoSyn: A Comprehensive Zeolite Synthesis Dataset Enabling Machine-Learning Rationalization of Hydrothermal Parameters Pan, Elton; Kwon, Soonhyoung; Jensen, Zach; Xie, Mingrou; Gómez-Bombarelli, Rafael; Moliner, Manuel; Román-Leshkov, Yuriy; Olivetti, Elsa Zeolites, nanoporous aluminosilicates with well-defined porous structures, are versatile materials with applications in catalysis, gas separation, and ion exchange. Hydrothermal synthesis is widely used for zeolite production, offering control over composition, crystallinity, and pore size. However, the intricate interplay of synthesis parameters necessitates a comprehensive understanding of synthesis-structure relationships to optimize the synthesis process. Hitherto, public zeolite synthesis databases only contain a subset of parameters and are small in scale, comprising up to a few thousand synthesis routes. We present ZeoSyn, a dataset of 23,961 zeolite hydrothermal synthesis routes, encompassing 233 zeolite topologies and 921 organic structure-directing agents (OSDAs). Each synthesis route comprises comprehensive synthesis parameters: 1) gel composition, 2) reaction conditions, 3) OSDAs, and 4) zeolite products. Using ZeoSyn, we develop a machine learning classifier to predict the resultant zeolite given a synthesis route with >70% accuracy. We employ SHapley Additive exPlanations (SHAP) to uncover key synthesis parameters for >200 zeolite frameworks. We introduce an aggregation approach to extend SHAP to all building units. We demonstrate applications of this approach to phase-selective and intergrowth synthesis. This comprehensive analysis illuminates the synthesis parameters pivotal in driving zeolite crystallization, offering the potential to guide the synthesis of desired zeolites. The dataset is available at https://github.com/eltonpan/zeosyn_dataset.

One-Pot Synthesis of CHA/ERI-Type Zeolite Intergrowth from a Single Multiselective Organic Structure-Directing Agent

2025-11-27T05:20:32Z

One-Pot Synthesis of CHA/ERI-Type Zeolite Intergrowth from a Single Multiselective Organic Structure-Directing Agent Kwon, Soonhyoung; Bello-Jurado, Estefanía; Ikonnikova, Evgeniia; Lee, Hwajun; Schwalbe-Koda, Daniel; Corma, Avelino; Willhammar, Tom; Olivetti, Elsa A; Gomez-Bombarelli, Rafael; Moliner, Manuel; Román-Leshkov, Yuriy We report the one-pot synthesis of a chabazite (CHA)/erionite (ERI)-type zeolite intergrowth structure characterized by adjustable extents of intergrowth enrichment and Si/Al molar ratios. This method utilizes readily synthesizable 6-azaspiro[5.6]dodecan-6-ium as the exclusive organic structure-directing agent (OSDA) within a potassium-dominant environment. High-throughput simulations were used to accurately determine the templating energy and molecular shape, facilitating the selection of an optimally biselective OSDA from among thousands of prospective candidates. The coexistence of the crystal phases, forming a distinct structure comprising disk-like CHA regions bridged by ERI-rich pillars, was corroborated via rigorous powder X-ray diffraction and integrated differential-phase contrast scanning transmission electron microscopy (iDPC S/TEM) analyses. iDPC S/TEM imaging further revealed the presence of single offretite layers dispersed within the ERI phase. The ratio of crystal phases between CHA and ERI in this type of intergrowth could be varied systematically by changing both the OSDA/Si and K/Si ratios. Two intergrown zeolite samples with different Si/Al molar ratios were tested for the selective catalytic reduction (SCR) of NO_x with NH₃, showing competitive catalytic performance and hydrothermal stability compared to that of the industry-standard commercial NH₃-SCR catalyst, Cu-SSZ-13, prevalent in automotive applications. Collectively, this work underscores the potential of our approach for the synthesis and optimization of adjustable intergrown zeolite structures, offering competitive alternatives for key industrial processes.

Recommendations for improving rigor and reproducibility in site specific characterization

2025-11-27T05:20:43Z

Recommendations for improving rigor and reproducibility in site specific characterization Wrasman, Cody J; Bell, Alexis T; Chandler, Bert D; Harris, James W; Kwon, Stephanie; Ball, Madelyn R; Krishna, Siddarth H; Khatib, Sheima J; Bollini, Praveen; Román-Leshkov, Yuriy; “Bean” Getsoian, Andrew; Weber, Robert S; Lercher, Johannes A; Liu, Dongxia; Resasco, Daniel E; Bates, Jason S; Hall, Jacklyn N; Lebrón-Rodríguez, Edgard A; Paz Herrera, Laura; Notestein, Justin M; Schaidle, Joshua A Heterogeneous catalysis is driven by the interaction of reactant molecules and the catalyst surface. The locus of this interaction as well as the surrounding ensemble of atoms is referred to as the catalyst active site. Active site characterization attempts to distinguish active catalytic sites from inactive surface sites, to elucidate the structural and chemical nature of active sites, and to quantify active site concentration. Numerous techniques have been demonstrated to provide compositional and structural information about the active sites within a catalyst. However, each technique has its own limitations and experimental pitfalls that can lead to data misinterpretation or irreproducible results. This work aims to provide an overview of the types of data that can be collected, to outline common experimental challenges and how to avoid them, and to assemble relevant references for the most used active site characterization techniques. More broadly, we aim to outline best practices for researchers to collect, interpret, and report active site characterization data in a way that provides the most benefit to the broader catalysis community. Increasing the rigor and reproducibility of active site characterization offers a strategy to better link properties with catalytic performance and to enable the community to develop consensus concerning these relationships.

Design and Validation of a High-Throughput Reductive Catalytic Fractionation Method

2025-11-27T05:20:36Z

Design and Validation of a High-Throughput Reductive Catalytic Fractionation Method Kenny, Jacob K; Neefe, Sasha R; Brandner, David G; Stone, Michael L; Happs, Renee M; Kumaniaev, Ivan; Mounfield, William P; Harman-Ware, Anne E; Devos, Katrien M; Pendergast, Thomas H; Medlin, J Will; Román-Leshkov, Yuriy; Beckham, Gregg T Reductive catalytic fractionation (RCF) is a promising method to extract and depolymerize lignin from biomass, and bench-scale studies have enabled considerable progress in the past decade. RCF experiments are typically conducted in pressurized batch reactors with volumes ranging between 50 and 1000 mL, limiting the throughput of these experiments to one to six reactions per day for an individual researcher. Here, we report a high-throughput RCF (HTP-RCF) method in which batch RCF reactions are conducted in 1 mL wells machined directly into Hastelloy reactor plates. The plate reactors can seal high pressures produced by organic solvents by vertically stacking multiple reactor plates, leading to a compact and modular system capable of performing 240 reactions per experiment. Using this setup, we screened solvent mixtures and catalyst loadings for hydrogen-free RCF using 50 mg poplar and 0.5 mL reaction solvent. The system of 1:1 isopropanol/methanol showed optimal monomer yields and selectivity to 4-propyl substituted monomers, and validation reactions using 75 mL batch reactors produced identical monomer yields. To accommodate the low material loadings, we then developed a workup procedure for parallel filtration, washing, and drying of samples and a ¹H nuclear magnetic resonance spectroscopy method to measure the RCF oil yield without performing liquid-liquid extraction. As a demonstration of this experimental pipeline, 50 unique switchgrass samples were screened in RCF reactions in the HTP-RCF system, revealing a wide range of monomer yields (21-36%), S/G ratios (0.41-0.93), and oil yields (40-75%). These results were successfully validated by repeating RCF reactions in 75 mL batch reactors for a subset of samples. We anticipate that this approach can be used to rapidly screen substrates, catalysts, and reaction conditions in high-pressure batch reactions with higher throughput than standard batch reactors.

Electrifying Hydroformylation Catalysts Exposes Voltage-Driven C–C Bond Formation

2025-11-27T05:20:26Z

Electrifying Hydroformylation Catalysts Exposes Voltage-Driven C–C Bond Formation Zeng, Joy S; Cosner, Emma L; Delgado-Kukuczka, Spencer P; Jiang, Chenyu; Adams, Jason S; Román-Leshkov, Yuriy; Manthiram, Karthish Electrochemical reactions can access a significant range of driving forces under operationally mild conditions and are thus envisioned to play a key role in decarbonizing chemical manufacturing. However, many reactions with well-established thermochemical precedents remain difficult to achieve electrochemically. For example, hydroformylation (thermo-HFN) is an industrially important reaction that couples olefins and carbon monoxide (CO) to make aldehydes. However, the electrochemical analogue of hydroformylation (electro-HFN), which uses protons and electrons instead of hydrogen gas, represents a complex C-C bond-forming reaction that is difficult to achieve at heterogeneous electrocatalysts. In this work, we import Rh-based thermo-HFN catalysts onto electrode surfaces to unlock electro-HFN reactivity. At mild conditions of room temperature and 5 bar CO, we achieve Faradaic efficiencies of up to 15% and turnover frequencies of up to 0.7 h^-1. This electro-HFN rate is an order of magnitude greater than the corresponding thermo-HFN rate at the same catalyst, temperature, and pressure. Reaction kinetics and operando X-ray absorption spectroscopy provide evidence for an electro-HFN mechanism that involves distinct elementary steps relative to thermo-HFN. This work demonstrates a step-by-step experimental strategy for electrifying a well-studied thermochemical reaction to unveil a new electrocatalyst for a complex and underexplored electrochemical reaction.

Enabling Lignin Valorization Through Integrated Advances in Plant Biology and Biorefining

2025-11-27T05:20:40Z

Enabling Lignin Valorization Through Integrated Advances in Plant Biology and Biorefining Dixon, Richard A; Puente-Urbina, Allen; Beckham, Gregg T; Román-Leshkov, Yuriy Despite lignin having long been viewed as an impediment to the processing of biomass for the production of paper, biofuels, and high-value chemicals, the valorization of lignin to fuels, chemicals, and materials is now clearly recognized as a critical element for the lignocellulosic bioeconomy. However, the intended application for lignin will likely require a preferred lignin composition and form. To that end, effective lignin valorization will require the integration of plant biology, providing optimal feedstocks, with chemical process engineering, providing efficient lignin transformations. Recent advances in our understanding of lignin biosynthesis have shown that lignin structure is extremely diverse and potentially tunable, while simultaneous developments in lignin refining have resulted in the development of several processes that are more agnostic to lignin composition. Here, we review the interface between in planta lignin design and lignin processing and discuss the advances necessary for lignin valorization to become a feature of advanced biorefining.

Reducing Solvent Consumption in Reductive Catalytic Fractionation through Lignin Oil Recycling

2025-11-27T05:20:28Z

Reducing Solvent Consumption in Reductive Catalytic Fractionation through Lignin Oil Recycling Jang, Jun Hee; Callejón Álvarez, Júlia; Neuendorf, Quinn S; Román-Leshkov, Yuriy; Beckham, Gregg T Reductive catalytic fractionation (RCF) enables the simultaneous valorization of lignin and carbohydrates in lignocellulosic biomass through solvent-based lignin extraction, followed by depolymerization and catalytic stabilization of the extracted lignin. Process modeling has shown that the use of exogenous organic solvent in RCF is a challenge for economic and environmental feasibility, and previous works proposed that lignin oil, a mixture of lignin-derived monomers and oligomers produced by RCF, can be used as a cosolvent in RCF. Here, we further explore the potential of RCF solvent recycling with lignin oil, extending the feasible lignin oil concentration in the solvent to 100 wt %, relative to the previously demonstrated 0-19 wt % range. Solvents containing up to 80 wt % lignin oil exhibited 83-93% delignification, comparable to 83% delignification with a methanol-water mixture, and notably, using lignin oil solely as a solvent achieved 67% delignification in the absence of water. In additional experiments, applying the RCF solvent recycling approach to ten consecutive RCF reactions resulted in a final lignin oil concentration of 11 wt %, without detrimental impacts on lignin extraction, lignin oil molar mass distribution, aromatic monomer selectivity, and cellulose retention. Overall, this work further demonstrates the potential for using lignin oil as an effective cosolvent in RCF, which can reduce the burden on downstream solvent recovery.

A Career in Catalysis: Mark E. Davis

2025-11-27T05:20:39Z

A Career in Catalysis: Mark E. Davis Arhancet, Juan P; Chen, Cong-Yan; Cybulskis, Viktor J; Gounder, Rajamani; Hong, Suk Bong; Jones, Christopher W; Kang, Jong Hun; Kubota, Yoshihiro; Lee, Hyunjoo; Orazov, Marat; Román-Leshkov, Yuriy; Schmidt, Joel E Mark E. Davis led an independent research program from 1981 to 2023, beginning at the Virginia Polytechnic Institute and State University (VPI) and then transitioning to the California Institute of Technology (Caltech). His research program was marked by exceptional creativity, breadth, and depth. With classical training in reaction engineering, Davis developed expertise in experimental heterogeneous catalysis and led work in this discipline for more than 40 years. His name is synonymous with zeolites, and today, he is one of the most widely recognized experts in zeolite synthesis, characterization, and catalysis in the world. Early work at the VPI focused on zeolites and catalysis with supported metal coordination complexes. His creativity was evident at the earliest stages of his career, with the development of supported aqueous phase catalysts and the world’s first crystalline, extra-large pore molecular sieve, both reported in the late 1980s. A move to Caltech saw a significant expansion of his zeolite synthesis program and the rapid acceleration of a multidecade collaboration with Dr. Stacey I. Zones of Chevron. At Caltech, his work expanded to include studies of molecular recognition and catalysis with organic/inorganic hybrid materials, and he developed a large, parallel program in drug delivery. His work on catalysis heavily emphasized zeolite catalysis, including major thrusts on the conversion of sugars in the liquid phase and methanol in the gas phase. Numerous new zeolites and molecular sieves were discovered throughout the four decades of the Davis laboratory, highlighted by a successful, multidecade quest to prepare a chiral zeolite with enantioselective catalytic properties. Davis is one of the most decorated researchers of the last four decades. He is one of only 21 living people currently elected to all of the US National Academies (Engineering, Science, Medicine) and elected as a Fellow of the National Academy of Inventors. He was the first engineer to win the NSF’s Alan T. Waterman Award and is one of only two researchers (to date) to win the International Zeolite Association’s Donald Breck Award twice (1989, 2019). Awards from the ACS (Ipatieff, Murphree, and Somorjai Awards), AIChE (Colburn, Professional Progress Awards), and North American Catalysis Society (Emmett Award) are among his accolades.

Plant Bioelectrical Signals for Environmental and Emotional State Classification

2025-11-27T05:20:23Z

Plant Bioelectrical Signals for Environmental and Emotional State Classification Gloor, Peter A. In this study, we present a pilot investigation using a single Purple Heart plant (Tradescantia pallida) to explore whether bioelectrical signals for dual-purpose classification tasks: environmental state detection and human emotion recognition. Using an AD8232 ECG sensor at 400 Hz sampling rate, we recorded 3 s bioelectrical signal segments with 1 s overlap, converting them to mel-spectrograms for ResNet18 CNN (Convolutional Neural Network) classification. For lamp on/off detection, we achieved 85.4% accuracy with balanced precision (0.85–0.86) and recall (0.84–0.86) metrics across 2767 spectrogram samples. For human emotion classification, our system achieved optimal performance at 73% accuracy with 1 s lag, distinguishing between happy and sad emotional states across 1619 samples. These results should be viewed as preliminary and exploratory, demonstrating feasibility rather than definitive evidence of plant-based emotion sensing. Replication across plants, days, and experimental sites will be essential to establish robustness. The current study is limited by a single-plant setup, modest sample size, and reliance on human face-tracking labels, which together preclude strong claims about generalizability.

Within-Subtype HIV-1 Polymorphisms and Their Impacts on Intact Proviral DNA Assay (IPDA) for Viral Reservoir Quantification

2025-11-27T05:20:24Z

Within-Subtype HIV-1 Polymorphisms and Their Impacts on Intact Proviral DNA Assay (IPDA) for Viral Reservoir Quantification Arikatla, Mohith Reddy; Mathad, Jyoti S.; Reddy, Kavidha; Reddy, Nicole; Ndung’u, Thumbi; Dupnik, Kathryn M.; Lee, Guinevere Q. The Intact Proviral DNA Assay (IPDA) is widely used to quantify genome-intact HIV proviruses in people living with HIV, but viral sequence diversity has been observed to cause assay failures due to primer/probe mismatches. Adapted for subtype C, IPDA-BC is a modified version of the IPDA validated on South African HIV-1 subtype C. India is also impacted by subtype C, but IPDA performance within-subtype across geographical regions is not well studied. We analyzed Indian (IN) and South African (ZA) subtype C sequences in silico, hypothesizing that IPDA-BC may underperform with IN viruses. Primer/probe binding was predicted using three increasingly stringent nucleotide mismatch criteria, whose sensitivity and specificity were evaluated against experimental IPDA outcomes. Phylogenetic analyses confirmed that IN and ZA subtype C sequences form distinct clusters with significant compartmentalization (p < 0.003). Across criteria, up to 6–10% decreases in primer/probe binding were observed in IN versus ZA, with the env forward primer being the most affected. These criteria showed low sensitivity (18–53%) and variable specificity (67–100%) in predicting experimental outcomes. In conclusion, even within subtype, HIV-1 variation across geographical regions may impact IPDA performance, underscoring the need for improved predictive models to guide assay design for global HIV cure research.

An Investigation of the Modulating Effects of Sensory Stimulation and Transcranial Magnetic Stimulation on Memory-Related Brain Activity

2025-11-27T05:20:21Z

An Investigation of the Modulating Effects of Sensory Stimulation and Transcranial Magnetic Stimulation on Memory-Related Brain Activity Nikolin, Stevan; Wang, Matthew; Moffa, Adriano; Huang, Haijing; Xu, Mei; Pande, Siddhartha Raj; Martin, Donel Background/Objectives: As the global population ages, the prevalence of disorders associated with memory dysfunction (e.g., Alzheimer’s disease) continues to increase. There is a need for novel interventions that can enhance memory and support affected individuals. Non-invasive brain stimulation provides a promising approach to engage circuits within the hippocampal network, a group of brain regions critical for episodic memory, and thereby improve cognition. Methods: Twenty healthy participants completed a single-blind, within-subject crossover study over four sessions. In each session, they received one of four interventions whilst viewing pictures of real-world objects: 40 Hz synchronised audiovisual stimulation (AVS), theta burst stimulation (TBS), a combination of synchronised 5 Hz repetitive transcranial magnetic stimulation with AVS (rTMS + AVS), or sham rTMS. Electroencephalography (EEG) was recorded to measure associated brain activity changes. Following each intervention, participants completed a recognition memory task. Results: Mixed-effect repeated measure models (MRMMs) revealed no significant differences in recognition memory performance or theta (5 Hz) activity across conditions. However, both TBS and rTMS + AVS significantly increased gamma (40 Hz) activity compared to sham rTMS, and TBS induced a widespread increase in theta-gamma phase-amplitude coupling during picture viewing. Conclusions: While the neuromodulatory interventions did not enhance memory performance, the observed increase in gamma activity, particularly following rTMS-based stimulation, suggests potential engagement of neural processes associated with memory. These findings warrant further investigation into the role of gamma oscillations in memory and cognitive enhancement.

Shaping In-Vehicle Behaviours through Activity-Centered Design

2025-11-27T05:20:06Z

Shaping In-Vehicle Behaviours through Activity-Centered Design Patel, Ankit; Gershon, Pnina; Habibovic, Azra; Novakazi, Fjoll?; Akahoshi, Sakura; Alsaid, Areen; Cha, Kyungjoo In today’s fast-paced society, most individuals commute either by personal vehicle or public transportation. User preferences and requirements are crucial, with design playing a significant role. The nature of design should be such that it is both inclusive and assimilative, and its purpose is to propel innovation and progress while also improving the quality of life of the user. That is why a general focus was given to the user-centered design approach while developing vehicles, especially, cabin (cockpit) design. With prioritizing the user activities, it is interesting to explore how users’ experience and behavior vary through the application of different design approaches. Nevertheless, existing literature has significantly overlooked the impact of design approaches on “human activity". Therefore, the main objective of the workshop is to examine the relationships between activity-centered design and user behavior. AutomotiveUI Adjunct ’25, Brisbane, QLD, Australia

Scale, Engage, or Both?: Potential and Perils of Applying Large Language Models in Interview and Conversation-Based Research

2025-11-27T05:20:01Z

Scale, Engage, or Both?: Potential and Perils of Applying Large Language Models in Interview and Conversation-Based Research Hwang, Angel Hsing-Chi; Aubin Le Qu?r?, Marianne; Schroeder, Hope; Cuevas, Alejandro; Dow, Steven; Kapania, Shivani; Rho, Eugenia An increasing number of studies apply tools powered by large language models (LLMs) to interview and conversation-based research, one of the most commonly used research methods in CSCW. This panel invites the CSCW community to critically debate the role of LLMs in reshaping interview-based methods. We aim to explore how these tools might (1) address persistent challenges in conversation-based research, such as limited scalability and participant engagement, (2) introduce novel methodological possibilities, and (3) surface additional practical, technical, and ethical concerns. The panel discussion will be grounded on the panelists’ prior experience applying LLMs to their own interview and conversation-based research. We ask whether LLMs offer unique advantages to enhance interview research, beyond automating certain aspects of the research process. Through this discussion, we encourage researchers to reflect on how applying LLM tools may require rethinking research design, conversational protocols, and ethical practices. CSCW Companion ’25, Bergen, Norway

Yeast Display Reveals Plentiful Mutations That Improve Fusion Peptide Vaccine-Elicited Antibodies Beyond 59% HIV-1 Neutralization Breadth

2025-11-27T05:20:42Z

Yeast Display Reveals Plentiful Mutations That Improve Fusion Peptide Vaccine-Elicited Antibodies Beyond 59% HIV-1 Neutralization Breadth França, Camila T; Pletnev, Sergei; Madan, Bharat; Katsamba, Phinikoula S; McKee, Krisha; Morano, Nicholas C; Zhang, Baoshan; Bahna, Fabiana; Bylund, Tatsiana; Lin, Bob C; Louder, Mark K; Mannepalli, Seetha; Nimrania, Rajani; O’Dell, Sijy; Doria-Rose, Nicole A; Kwong, Peter D; Shapiro, Lawrence; Sheng, Zizhang; Zhou, Tongqing; DeKosky, Brandon J Background/Objectives: Vaccine elicitation of antibodies with high HIV-1 neutralization breadth is a long-standing goal. Recently, the induction of such antibodies has been achieved at the fusion peptide site of vulnerability. Questions remain, however, as to how much anti-fusion peptide antibodies can be improved and whether their neutralization breadth and potency are sufficient to prevent HIV-1 infection. Methods: Here, we use yeast display coupled with deep mutational screening and biochemical and structural analyses to study the improvement of the best fusion peptide-directed, vaccine-elicited antibody, DFPH_a.01, with an initial 59% breadth. Results: Yeast display identified both single and double mutations that improved recognition of HIV-1 envelope trimers. We characterized two paratope-distal light chain (LC) mutations, S10R and S59P, which together increased breadth to 63%. Biochemical analysis demonstrated DFPH-a.01_10R59P-LC, and its component mutations, to have increased affinity and stability. Cryo-EM structural analysis revealed elbow-angle influencing by S10R-LC and isosteric positioning by S59P-LC as explanations for enhanced breadth, affinity, and stability. Conclusions: These results, along with another antibody with enhanced performance (DFPH-a.01_1G10A56K-LC with 64% breadth), suggest that mutations improving DFPH_a.01 are plentiful, an important vaccine insight.

A Probabilistic Perspective on Tiling Sparse Tensor Algebra

2025-11-27T05:20:09Z

A Probabilistic Perspective on Tiling Sparse Tensor Algebra Sharma, Ritvik; Xue, Zi Yu; Zhang, Nathan; Lacouture, Rubens; Kjolstad, Fredrik; Achour, Sara; Horowitz, Mark Sparse tensor algebra computations are often memory-bound due to irregular access patterns and low arithmetic intensity. We present D2T2 (Data-Driven Tensor Tiling), a framework that optimizes static coordinate-space tiling schemes to minimize memory traffic by identifying and leveraging relevant high-level statistics from input operands. For a given tensor algebra computation, D2T2 collects statistics from input tensors, builds a probability distribution-based model of the tensor computation, and uses it to predict traffic for various tiling configurations. It searches over tile shape and size configurations to minimize total traffic. We evaluate D2T2 against Tailors and DRT, two state of the art tiling schemes for sparse tensor algebra. We find that D2T2 achieves, on average, a 2.54 × speedup over Tailors and a 1.13× lower memory bandwidth compared to DRT for sparse-sparse matrix multiplication (SpMSpM). We also achieve 1.22–48.94× lower bandwidth for SpMSpM and up to 34.31× lower bandwidth for tensor operations (TTM and MTTKRP) than conservative static tiling schemes. Unlike prior tiling techniques, D2T2 is deployable without specialized hardware support. On Opal, a 16nm sparse tensor algebra accelerator, D2T2 generated tiling configurations that achieve 1.23–3.34 × speedups compared to their original hand-tuned configurations. MICRO ’25, Seoul, Republic of Korea

HapticHearing: A Haptic Feedback System for Complementing Auditory Speech Perception for Mild-to-Moderate Hearing Loss

2025-11-27T05:20:00Z

HapticHearing: A Haptic Feedback System for Complementing Auditory Speech Perception for Mild-to-Moderate Hearing Loss Chin, Sam; Fitz-Gibbon, Emmie; Huang, Bingjian; Paradiso, Joseph Age-related hearing loss is often caused by cochlear hair cell degradation. This creates a challenge for hearing aids, which rely on sound amplification. Once hearing ability in a specific frequency is lost, amplification alone provides little benefit. Previous haptic systems have tried to solve this with complete sensory substitution, converting audio signals like phonemes to tactile patterns. However, these systems require significant amount of time to learn, and induce high cognitive load in haptic perception. Our system, HapticHearing, takes an alternative approach of leveraging a user’s residual hearing and complementing it with tactile feedback. We present a custom multi-actuator haptic device, designed to translate phonemic information from speech into tactile patterns that are customized to a user’s hearing loss and speech perception abilities. The system consists of a microphone for speech capture, four-band energy envelope extraction with vowel embedding, a custom USB-to-haptic driver PCB, and wearable devices containing eight vibrotactile actuators that deliver personalized tactile feedback based on the user’s audiogram. Psychophysical validation (n=9) showed neck-worn devices achieved better spatial localization (67% vs 53%) while while bracelet and necklace devices had lower detection thresholds than over-ear (thresholds 0.09 vs 0.18). ASSETS ’25, Denver, CO, USA

Resonance: Drawing from Memories to Imagine Positive Futures through AI-Augmented Journaling

2025-11-27T05:20:13Z

Resonance: Drawing from Memories to Imagine Positive Futures through AI-Augmented Journaling Zulfikar, Wazeer; Chiaravalloti, Treyden; Shen, Jocelyn; Picard, Rosalind; Maes, Pattie People inherently use experiences of their past while imagining their future, a capability that plays a crucial role in mental health. Resonance is an AI-powered journaling tool designed to augment this ability by offering AI-generated, action-oriented suggestions for future activities based on the user’s own past memories. Suggestions are offered when a new memory is logged and are followed by a prompt for the user to imagine carrying out the suggestion. In a two-week randomized controlled study (N=55), we found that using Resonance significantly improved mental health outcomes, reducing the users’ PHQ8 scores, a measure of current depression, and increasing their daily positive affect, particularly when they would likely act on the suggestion. Notably, the effectiveness of the suggestions was higher when they were personal, novel, and referenced the user’s logged memories. Finally, through open-ended feedback, we discuss the factors that encouraged or hindered the use of the tool. AHs 2025, Masdar City, Abu Dhabi, United Arab Emirates

A11yShape: AI-Assisted 3-D Modeling for Blind and Low-Vision Programmers

2025-11-27T05:20:04Z

A11yShape: AI-Assisted 3-D Modeling for Blind and Low-Vision Programmers Zhang, Zhuohao (Jerry); Li, Haichang; Yu, Chun Meng; Faruqi, Faraz; Xie, Junan; Kim, Gene; Fan, Mingming; Forbes, Angus; Wobbrock, Jacob; Guo, Anhong; He, Liang Building 3-D models is challenging for blind and low-vision (BLV) users due to the inherent complexity of 3-D models and the lack of support for non-visual interaction in existing tools. To address this issue, we introduce A11yShape, a novel system designed to help BLV users who possess basic programming skills understand, modify, and iterate on 3-D models. A11yShape leverages LLMs and integrates with OpenSCAD, a popular open-source editor that generates 3-D models from code. Key functionalities of A11yShape include accessible descriptions of 3-D models, version control to track changes in models and code, and a hierarchical representation of model components. Most importantly, A11yShape employs a cross-representation highlighting mechanism to synchronize semantic selections across all model representations—code, semantic hierarchy, AI description, and 3-D rendering. We conducted a multi-session user study with four BLV programmers, where, after an initial tutorial session, participants independently completed 12 distinct models across two testing sessions, achieving results that aligned with their own satisfaction. The result demonstrates that participants were able to comprehend provided 3-D models, as well as independently create and modify 3-D models—tasks that were previously impossible without assistance from sighted individuals. ASSETS ’25, Denver, CO, USA

Benthic: Perceptually Congruent Structures for Accessible Charts and Diagrams

2025-11-27T05:20:31Z

Benthic: Perceptually Congruent Structures for Accessible Charts and Diagrams Mei⁎, Catherine; Pollock⁎, Josh; Hajas, Daniel; Zong, Jonathan; Satyanarayan, Arvind Graphical representations — such as charts and diagrams — have a visual structure that communicates the relationship between visual elements. For instance, we might consider two elements to be connected when there is a line or arrow between them, or for there to be a part-to-whole relationship when one element is contained within the other. Yet, existing screen reader solutions rarely surface this structure for blind and low-vision readers. Recent approaches explore hierarchical trees or adjacency graphs, but these structures capture only parts of the visual structure — containment or direct connections, respectively. In response, we present Benthic, a system that supports perceptually congruent screen reader structures, which align screen reader navigation with a graphic’s visual structure. Benthic models graphical representations as hypergraphs: a relaxed tree structure that allows a single hyperedge to connect a parent to a set of children nodes. In doing so, Benthic is able to capture both hierarchical and adjacent visual relationships in a manner that is domain-agnostic and enables fluid (i.e., concise and reversible) traversal. To evaluate Benthic, we conducted a study with 15 blind participants who were asked to explore two kinds of graphical representations that have previously been studied with sighted readers. We find that Benthic’s perceptual congruence enabled flexible, goal-driven exploration and supported participants in building a clear understanding of each diagram’s structure. ASSETS ’25, Denver, CO, USA

Quartz: A Reconfigurable, Distributed-Memory Accelerator for Sparse Applications

2025-11-27T05:19:04Z

Quartz: A Reconfigurable, Distributed-Memory Accelerator for Sparse Applications Golden, Courtney; Feldmann, Axel; Emer, Joel; Sanchez, Daniel Iterative sparse matrix computations lie at the heart of many scientific computing and graph analytics algorithms. On conventional systems, their irregular memory accesses and low arithmetic intensity create challenging memory bandwidth bottlenecks. To overcome such bottlenecks, distributed-SRAM architectures are structured as an array of tiles, each with a processing element (PE) and a small local memory, to achieve very high aggregate memory bandwidth. However, current distributed-SRAM architectures suffer from either poor programmability due to over-specialized PEs or poor compute performance due to inefficient general-purpose PEs. We propose Quartz, a new architecture that uses short dataflow tasks and reconfigurable PEs in a distributed-SRAM system to deliver both high performance and high programmability. Unlike traditional sparse CGRAs or on-die reconfigurable engines, Quartz allows reconfigurable compute to be highly utilized and scaled by (1) providing high memory bandwidth to each processing element and (2) introducing a task-level dataflow execution model that fits this new setting. Our execution model dynamically reconfigures each tile’s PE in response to inter-tile messages to execute tasks on local data. This execution model enables fine-grained data partitioning across tiles. To make execution efficient, we explore novel data partitioning techniques that use graph and hypergraph partitioning to minimize network traffic and balance load in the face of both static-static and static-dynamic operand sparsity. To ensure programmability, we show how a wide range of Einsum-expressible computations and flexible data distributions can be systematically captured in small tasks for execution on Quartz. Quartz’s architecture, data partitioning techniques, and programming model together achieve gmean 21.4 × speedup over a prior state-of-the-art system for six different iterative sparse applications from scientific computing and graph analytics. MICRO ’25, Seoul, Republic of Korea

Converting Spatial to Social: Using Persistent Homology to Understand Social Groups

2025-11-27T05:19:58Z

Converting Spatial to Social: Using Persistent Homology to Understand Social Groups Chen, Valerie; Liang, Claire; Shah, Julie; Andrist, Sean In social settings, people display sophisticated spatial behaviors—for example, one might naturally enter into a conversation by sidling up to a group. Artificial agents will need the ability to reason about spatial representations of social information to understand not only how social groups form, but also how to interact within and around them. Leveraging the insight that people reason about shared space topologically rather than geometrically, we employ techniques from applied topology to introduce a new method for social group analysis that improves quantifiability and enables rigorous analysis of social group structure. We present a novel topological mathematical formalism called the social simplicial complex that provides an equivalence relation for socially analogous configurations of people and is provably robust against small perturbations and noise. Moreover, this formalism suggests quantifiable metrics to assess the confidence of social group existence and the social closeness of people within groups. We further use this formalism to introduce an open-source toolkit for evaluating possible models of social relationships, which we name the Social Topological Analysis (SoTA) Toolkit. Finally, we explore algebraic topology’s potential to serve more generally as a powerful tool for multi-modal social data processing, and its possibilities for further applications in social-spatial analysis. ICMI ’25, Canberra, ACT, Australia

LLMs in Citation Intent Classification: Progress, Precision, and Reproducibility Challenges

2025-11-27T05:20:07Z

LLMs in Citation Intent Classification: Progress, Precision, and Reproducibility Challenges Fogelson, Alex; Thompson, Neil; Trišović, Ana Understanding the intent behind scientific citations is critical for advancing scholarly search and knowledge mapping. This paper reflects on the methodological use of large language models (LLMs) for multi-class citation intent classification. Our experiments evaluating a diverse range of models and approaches reveal striking disagreement among state-of-the-art (SotA) systems. This inconsistency suggests that citation intent classification remains a challenging task for LLMs raising questions about the robustness, reliability and replicability of current methods. Moreover, our findings highlight a concerning dependency on proprietary LLMs that, even with access to compute resources, were necessary to achieve sufficient accuracy. This introduces new challenges, as silent updates, lack of versioning, and opaque training pipelines pose threats to methodological transparency and long-term reproducibility in LLMenabled research. ACM REP ’25, Vancouver, BC, Canada

Ancestral Technology: Inside Colombia’s Hidden Technological Landscape

2025-11-27T05:20:35Z

Ancestral Technology: Inside Colombia’s Hidden Technological Landscape Reynolds-Cuellar, Pedro Luz Marina Burgos’ fingers moved deliberately across the threads, constructing a tšombiach—a ceremonial sash commonly used to protect and strengthen the body. “This is the frog; to us, it represents fertility,” she explained, pointing to an emerging pattern. “This is the sun. Families weave it differently. This is how the tšombiach helps us tell our own story.” What I witnessed in this Colombian village was not simply craft—it was a technology for encoding and transmitting intergenerational knowledge. Passing most of my time between MIT and Harvard created a sense of technology as merely technical or socio-technical systems serving as a means to undetermined progress that only a few seem able to influence or have power over. A sense of relentless push towards the new, often at the expense of the old. Learning from Luz Marina, a traditional weaver from the Quillasinga Indigenous people, helped me make sense of radically different technological values, motivations and purposes. She is part of a centuries-long tradition of sustaining technologies designed for a different purpose entirely: cultural preservation. These technological systems solve immediate problems while maintaining the social fabric that makes problem-solving possible across generations. During five years of fieldwork in Colombia’s rural communities —ultimately leading to my doctoral dissertation— I encountered technologies that function according to entirely different logics than those driving “modern” narratives of innovation. I began —along with my collaborators in Colombia— conceptualizing these as “ancestral technologies”: forms of world-making —some of which take the form of artifacts— that primarily support cultural cohesion, remain rooted in specific geographies and carry their history through collective memory. Unlike modern technologies optimized for profit, efficiency or scale, these ancestral systems optimize for continuity and collective meaning. In an era when predictive technology sells the fantasy that unlimited computational power must be our goal as a society, perhaps the question is not whether we can build more powerful systems, but whether we can build systems that help us preserve what matters most.

Asymmetric linker generates intrinsically disordered metal–organic framework with local MOF-74 structure

2025-11-26T03:11:05Z

Asymmetric linker generates intrinsically disordered metal–organic framework with local MOF-74 structure Dinakar, Bhavish; Oppenheim, Julius J; Vandone, Marco; Torres, Juan F; Iliescu, Andrei; Yang, Zhentao; Román-Leshkov, Yuriy; Dincă, Mircea Here, we report an intrinsically disordered MOF in the MOF-74 family, Mg2x(as-dobpdc) (as-dobpdc4 = 30 ,4-dioxidobiphenyl-3,40 -dicarboxylate). Despite the absence of crystallinity, this material exhibits local ordering consistent with that of its crystalline isomers, maintains porosity, and exhibits a high density of open metal sites.

Sustainable aviation fuels from biomass and biowaste via bio- and chemo-catalytic conversion: Catalysis, process challenges, and opportunities

2025-11-26T03:11:03Z

Sustainable aviation fuels from biomass and biowaste via bio- and chemo-catalytic conversion: Catalysis, process challenges, and opportunities Zhang, Junyan; Webber, Matthew S; Pu, Yunqiao; Li, Zhenglong; Meng, Xianzhi; Stone, Michael L; Wei, Bingqing; Wang, Xueqi; Yuan, Sainan; Klein, Bruno; Seemala, Bhogeswararao; Wyman, Charles E; Ramasamy, Karthikeyan K; Thorson, Mike; Langholtz, Matthew H; Heyne, Joshua S; Koishybay, Aibolat; Adhikari, Shiba; Cao, Sufeng; Sutton, Andrew D; Tuskan, Gerald A; Román-Leshkov, Yuriy; Ragauskas, Arthur J; Ling, Tao; Davison, Brian H Sustainable aviation fuel (SAF) production from biomass and biowaste streams is an attractive option for decarbonizing the aviation sector, one of the most-difficult-to-electrify transportation sectors. Despite ongoing commercialization efforts using ASTM-certified pathways (e.g., lipid conversion, Fischer–Tropsch synthesis), production capacities are still inadequate due to limited feedstock supply and high production costs. New conversion technologies that utilize lignocellulosic feedstocks are needed to meet these challenges and satisfy the rapidly growing market. Combining bio- and chemo-catalytic approaches can leverage advantages from both methods, i.e., high product selectivity via biological conversion, and the capability to build C-C chains more efficiently via chemical catalysis. Herein, conversion routes, catalysis, and processes for such pathways are discussed, while key challenges and meaningful R&D opportunities are identified to guide future research activities in the space. Bio- and chemo-catalytic conversion primarily utilize the carbohydrate fraction of lignocellulose, leaving lignin as a waste product. This makes lignin conversion to SAF critical in order to utilize whole biomass, thereby lowering overall production costs while maximizing carbon efficiencies. Thus, lignin valorization strategies are also reviewed herein with vital research areas identified, such as facile lignin depolymerization approaches, highly integrated conversion systems, novel process configurations, and catalysts for the selective cleavage of aryl C–O bonds. The potential efficiency improvements available via integrated conversion steps, such as combined biological and chemo-catalytic routes, along with the use of different parallel pathways, are identified as key to producing all components of a cost-effective, 100% SAF.

Methods for Carbon Mass Closure in Polyolefin Hydrocracking

2025-11-26T03:11:02Z

Methods for Carbon Mass Closure in Polyolefin Hydrocracking Brenner, Anna E; Drake, Griffin; Beckham, Gregg T; Román-Leshkov, Yuriy Heterogeneous catalytic hydrocracking of polyolefins is a promising approach for the processing of postconsumer plastics, but product quantification methods remain inconsistent across the literature. In systems that generate a large fraction of vapor-phase products, typical product capture methods can result in large carbon balance deficits, exceeding 50%, compromising reported yields and selectivities. Here, we identify the major sources of product loss and develop enhanced capture methods to improve the quantification accuracy. Seven supplemental techniques were evaluated, targeting either increased vapor recovery (by increasing the volatility or system volume) or enhanced retention in the liquid phase (by decreasing volatility). Among these, a flow collection approach using a continuous helium sweep and downstream gas sampling bag capture yielded the highest recovery, achieving a 96 ± 9.2% carbon balance closure. We show that the efficacy of these methods is strongly dependent on product distribution. In general, solvent addition was most effective when condensable species dominate the product distribution, while flow collection was preferred when both condensable species and light gases are present in high concentrations. These results highlight the need for method-specific workup strategies and demonstrate that no single protocol is universally optimal. We provide general guidelines for selecting and implementing robust product capture techniques, enabling accurate yield and selectivity determinations in polyolefin hydrocracking systems.

Lignin Extraction and Condensation as a Function of Temperature, Residence Time, and Solvent System in Flow-through Reactors

2025-11-26T03:11:00Z

Lignin Extraction and Condensation as a Function of Temperature, Residence Time, and Solvent System in Flow-through Reactors Brandner, David G; Gracia Vitoria, Jaime; Kenny, Jacob K; Bussard, Jeremy R; Jang, Jun Hee; Woodworth, Sean P; Vanbroekhoven, Karolien; Román-Leshkov, Yuriy; Beckham, Gregg T Solvolytic extraction of lignin from biomass is a critical step in lignin-first biorefining, including the reductive catalytic fractionation (RCF) process. Key to optimal RCF processing is the ability to rapidly extract lignin from biomass at high delignification extents and transfer the lignin molecules to a catalyst surface in a time frame that minimizes lignin condensation reactions. Here, we use a flow-through reactor to study the effects of temperature (175-250 °C), residence time (9 to 36 min), and solvent composition (methanol and methanol-water) on lignin extraction and condensation. We evaluated three metrics at each condition: total delignification, delignification rate, and extent of condensation, the latter measured by a decrease in monomer yield for batch hydrogenolysis reactions of solvolysis liquor compared to batch RCF reactions. We observe that delignification is predominantly determined by temperature, while residence time dictates the lignin condensation extent. Moreover, the extent of both extraction and condensation increased in the methanol-water solvent system compared to that in the methanol system. Lignin extracted in methanol is stable up to 18-min residence times at or below 225 °C, while a majority of the lignin extracted in methanol-water is condensed with a 9-min residence time at 200 °C. These results can inform reactor designs and solvent selection for lignin-first biorefining processes that aim to physically separate the biomass and catalyst.

Additive Manufacturing of Electrical Machines and Electronic Devices

2025-11-26T03:04:38Z

Additive Manufacturing of Electrical Machines and Electronic Devices Cañada Pérez-Sala, Jorge Recent advancements in the additive manufacture of electronics and electrical machines have led to successful demonstrations of 3D-printed passive (e.g., resistors, capacitors, inductors) and active (e.g., transistors) electronic components, as well as magnetic cores and power transfer devices. However, each new demonstration of 3D-printed functional devices has typically required increasingly specialized and expensive manufacturing hardware. This work opposes that trend by developing a technology capable of fabricating all such devices on a single, affordable machine: a material extrusion 3D printer. Material extrusion stands out among additive manufacturing technologies for its widespread availability and its compatibility with monolithic multi-material manufacturing, essential for the fabrication of functional electromagnetic devices. These attributes, together with its well-established ability to fabricate mechanically functional parts, make material extrusion a promising technology for the single-step fabrication of electronics and electrical machines, and for their monolithic integration into complex devices, such as custom functionalized prostheses, robots, and space exploration hardware. In this research, a desktop 3D printer was transformed into an almost-universal manufacturing machine capable of fabricating a myriad of electrically, magnetically, and mechanically functional devices, using various feedstock formats (e.g., filament, pellets, ink). With this machine, milestones such as the fabrication of the first semiconductorfree, fully 3D-printed logic gates, and that of the first fully 3D-printed motor, have been achieved. Built for under $4000 in parts, the modified 3D printer opens the door to the democratization of electronics and electrical machine manufacturing, empowering institutions and individuals alike, and serving as an educational tool to introduce advanced manufacturing to new generations. Additionally, this work investigates optimization strategies for planar inductors and alternative techniques for the creation of miniaturized, three-dimensional, electrically functional components via two-photon polymerization. By demonstrating novel methods and applications, this thesis advances the state of the art in the additive manufacture of electromagnetic devices and paves the way toward the decentralized fabrication of electrical machines and electronic devices.

A Unified Theory of Representation Learning: How Hidden Relationships Power Algorithms that can Learn without Labels

2025-11-26T03:04:36Z

A Unified Theory of Representation Learning: How Hidden Relationships Power Algorithms that can Learn without Labels Hamilton, Mark T. How does the human mind make sense of raw information without being taught how to see or hear? This thesis presents a unifying theory that describes how algorithms can learn and discover structure in complex systems, like natural images, audio, language, and video - without human input. This class of algorithms has the possibility to extend our own understanding of the world by helping us to see previously unseen patterns in nature and science. At the core of this thesis’ unified theory is the notion that relationships between deep network representations hold the key discover the structure of the world without human input. This work will begin with a few examples of this principle in action; discovering hidden connections that span cultures and millennia in the visual arts, discovering visual objects in large image corpora, classifying every pixel of our visual world, and rediscovering the meaning of words from raw audio, all without human labels. In the latter half of this thesis, we will present two unifying mathematical theories of unsupervised learning. The first will explain why relationships between deep features can rediscover the semantic structure of the natural world by connecting model explainability, cooperative game theory, and deep feature relationships. The second mathematical theory will show that relationships between representations can be used to unify over 20 common machine learning algorithms spanning 100 years of progress in the field of machine learning. In particular, we introduce a single equation that unifies classification, regression, large language modeling, dimensionality reduction, clustering, contrastive learning, and spectral methods. This thesis uses this unified equation as the basis for a “periodic table of representation learning” that predicts the existence of new types of algorithms. We show that one of these predicted algorithms is a state-of-the-art unsupervised image classification technique. Finally, this work will summarize the key findings and share ongoing and future directions.

Score Estimation for Generative Modeling

2025-11-26T03:04:32Z

Score Estimation for Generative Modeling Jayashankar, Tejas Kumar Recent advances in score-based (diffusion) generative models have achieved state-of-the-art sample quality across standard benchmarks. Building on the remarkable property of these models in estimating scores, this thesis presents three core contributions: 1) new objectives to reduce score estimation error, 2) a novel Bayesian-inspired optimization framework for solving inverse problems, and 3) a fast one-step generative modeling framework that is based on a novel amortized score estimation framework. In the first part of this thesis, we introduce two new score estimation objectives with applications to both implicit and diffusion-based generative models. To improve spectralbased non-parametric estimators, we propose a theoretically optimal parametric framework that learns the score by projecting it onto its top-L principal directions. Additionally, inspired by matrix-valued kernel methods, we present a second approach that lifts the score into the space of outer products, and minimizes the distance between the estimated and true scores in this higher-order space. In the second part, we shift focus from score estimation to leveraging diffusion models as data-driven priors for solving inverse problems. Centering our development around the problem of source separation, we introduce a novel algorithm inspired by maximum a posteriori estimation. This approach combines multiple levels of Gaussian smoothing with an α-posterior, enabling effective signal separation using only independent priors for the sources. We demonstrate the effectiveness of this method through its application to interference mitigation in digital communication signals. Finally, we outline how this framework can be naturally extended to tackle a broader class of inverse problems. In the final part, we return to the fundamental challenge of efficient sampling, which is critical for enabling practical data-driven engineering systems. We propose a novel generative modeling framework that enables training a one-step neural sampler from scratch. At the core of this method is a new objective based on multi-divergence minimization, guided by a novel approach for score estimation of mixture distributions. Our framework is simple to implement, stable during training, unifies several existing approaches, and achieves state-of-the-art performance in image generation tasks. Furthermore, we discuss how this framework can be naturally extended to multi-step neural sampling and adapted for fast posterior sampling—an essential component in simulation-based inverse problem solvers.

Superconducting Nanowire Integrated Circuits for Scalable Cryogenic Memory

2025-11-26T03:04:30Z

Superconducting Nanowire Integrated Circuits for Scalable Cryogenic Memory Medeiros, Owen A. Superconducting nanowire integrated circuits (SNICs) are a promising class of cryogenic electronics that harness the zero resistance, high kinetic inductance, and nanoscale geometry of ultrathin superconducting wires to implement logic, memory, amplification, and sensing with minimal energy dissipation. Unlike Josephson-junction-based circuits, SNICs support compact, planar layouts compatible with single-layer fabrication and operation in unshielded cryogenic environments. This thesis develops superconducting nanowire memory (SNM) as a scalable implementation of SNICs. A modular cell architecture is introduced, exploiting hysteretic switching and inductive asymmetry to enable nonvolatile digital state storage with zero static power consumption. A hierarchical design framework is established, combining automated layout generation, electrothermal simulation in LTspice, and microscopic modeling using the time-dependent Ginzburg–Landau (TDGL) formalism. To enable scalable integration, this work implements a row–column SNM array layout and demonstrates fabrication across full 4-inch wafers using a planar, singlelayer process. Cryogenic measurements validate reliable operation in both single cells and multi-cell arrays, confirming the predictive accuracy of the design and modeling framework. Tradeoffs in bias current levels, pulse timing, and read/write conditions are systematically evaluated through cryogenic measurements, revealing their impact on bit error rate, operational margins, and energy efficiency across single cells and arrays. Together, these contributions establish SNICs as a viable and extensible platform for cryogenic memory, providing the tools, models, and infrastructure needed to enable broader adoption in quantum computing, neuromorphic systems, and other energy-constrained cryogenic applications.

Next Generation Operating Systems for the Datacenter

2025-11-26T03:04:08Z

Next Generation Operating Systems for the Datacenter Fried, Joshua Modern datacenters face a fundamental challenge: handling demanding real-time and dataintensive workloads that require both microsecond-scale low latency and high throughput, while simultaneously achieving high resource utilization and efficient multi-tenancy. Traditional operating systems, designed for an era of slower hardware, introduce significant overheads to microsecond-scale I/O that prevent applications from exploiting the full performance of the underlying hardware. Furthermore, their millisecond-scale resource management is ill-equipped to handle the microsecond-level burstiness of modern workloads, leading to costly overprovisioning and idle resources. Recognizing the performance limitations imposed by traditional OSes, a common workaround has emerged: letting applications communicate directly with hardware, bypassing the OS entirely. While this approach offers performance gains by removing the OS from the critical path, existing kernel-bypass solutions require dedicated resources, extensive application rewrites, and provide weak isolation, making them unsuitable for general-purpose, shared environments. This thesis presents a new datacenter operating system, composed of three integrated systems: Shenango, Caladan, and Junction. Together, they preserve the high-performance, low-overhead I/O benefits of kernel bypass, while providing efficient resource management, strong isolation for multi-tenant workloads, and compatibility with unmodified software. First, Shenango enables applications to bypass traditional OS-mediated I/O without dedicating CPU cores solely to polling. Next, Caladan ensures that idle resources can be used productively by other applications by actively managing competition for microarchitectural resources, thereby preserving each application’s high I/O performance and responsiveness. Finally, Junction overcomes several common limitations of kernel-bypass solutions, bringing these benefits to all applications by preserving compatibility with existing software and reducing memory and polling overheads. Collectively, these systems provide the advantages of direct hardware access without sacrificing the flexibility or efficiency of a general-purpose operating system. This work demonstrates that high I/O performance, efficient resource utilization, and broad application compatibility can indeed coexist, paving the way for a new generation of datacenter operating systems.

Systematic Development of Healthcare AI: From Data Curation, Algorithm Optimization, Benchmark Design and Clinical Applications

2025-11-26T03:04:22Z

Systematic Development of Healthcare AI: From Data Curation, Algorithm Optimization, Benchmark Design and Clinical Applications Gao, Mingye Artificial intelligence (AI) has brought transformative changes to healthcare industry in the recent years from various aspects, such as patient care, disease diagnosis and medical research. As healthcare systems worldwide face increasing pressure from aging populations and rising chronic disease rates, there is an urgent need for systematic approaches to develop reliable and safe AI solutions. This thesis advances the systematic development of healthcare AI through four interconnected components: data curation, algorithm optimization, benchmark design, and clinical applications. The primary contribution of this thesis focuses on establishing a comprehensive pipeline for healthcare large language models (LLMs), spanning from data curation to clinical deployment. At the data level, a rule-based filtering framework was developed to select high-quality subsets from the large pre-training corpora, significantly improving both continue pre-training and fine-tuning performance of LLMs. For safety alignment, an automated pipeline was developed for preference learning that includes preference dataset synthesis, rule-based and data-adaptive annotation, and reward model training. Additionally, two novel benchmarks were created to ensure reliability and safety of LLMs in healthcare tasks: one assessing demographic biases of LLMs across common diseases, while another assessing models’ ability to reject illogical requests from users in drug-related scenarios. Finally, LLMs were used to generating patient-friendly educational content for clinical trials, demonstrating their role in improving patient education and engagement in clinical trials. This systematic progression from data to deployment establishes a blueprint for developing safe and effective language models in healthcare settings. Beyond language models, machine learning techniques were applied on an additional healthcare task. In this project, a novel approach combining normalized cross-correlation and attention graph convolutional recurrent networks was developed to realize contactless, continuous and reliable radar-based vital signs monitoring in dynamic home environments. Through systematic data collection and algorithm optimization, the accurate heart rate can be obtained across varying radar-subject distances (2-2.5m) and subject orientations, demonstrating robust performance in real-world conditions through extensive validation in four test houses with six subjects. Collectively, these contributions advance healthcare AI development across 2 fronts: establishing frameworks for safe and effective deployment of language models in healthcare settings and enabling reliable and continuous health monitoring at-home without wearable devices.

SERenaDE: Hardware Acceleration of Cloud Serialization Frameworks

2025-11-26T03:06:36Z

SERenaDE: Hardware Acceleration of Cloud Serialization Frameworks Zarkos, Christos V. Serialization frameworks are a fundamental operation of datacenters, as they enable language- and platform-neutral communication and storage. However, software serialization faces major performance bottlenecks, resulting in a significant fraction of cloud cycles dedicated to this process. Prior work has proposed specialized hardware accelerators to address these overheads. While these proposals achieve considerable speedups, they are expensive in terms of verification, fabrication, and deployment, and often hardcode too many details about the (de)serialization framework in hardware. We propose SERenaDE, a serialization framework designed to integrate general-purpose accelerators currently deployed in datacenters in order to accelerate and offload serialization to hardware. Specifically, we repurpose the Intel In-Memory Analytics Accelerator (IAA), an accelerator engine offering fast compression, to enable fast and transparent to the user serialization and deserialization, completely removing software serialization from the execution pipeline. We evaluate our system on latest-generation production machines, both with synthetic microbenchmarks, and open-source representative fleet-wide benchmarks. Our results show comparable performance in terms of per-request latency across all types of messages, while significantly improving throughput - especially at the tail -, maintaining thread scalability and achieving high compression ratios alongside substantial speedups for larger messages. Under 95th latency percentile latency constraints SERenaDE improves serialization and deserialization throughput by 13% and 30% respectively, while achieving from 0.2x to 6.94x smaller serialized message sizes for messages of a total memory layout larger than 4KB.

Modeling Biomolecular Interactions with Generative Models

2025-11-26T03:04:11Z

Modeling Biomolecular Interactions with Generative Models Corso, Gabriele In 2021, DeepMind’s AlphaFold2 revolutionized single-chain protein structure prediction achieving atomic accuracy, solving a longstanding challenge in biology. However, understanding biomolecular interactions, a critical problem for advancing drug discovery and biological research, remained unsolved. This thesis presents our research to redefine the machine learning approach to this problem, modeling structures with a new generative paradigm and tailoring the neural architectures and learning tasks to the specific challenges that arose. These ideas combined with significant engineering efforts led us to develop a class of open-source models from DiffDock to the recent Boltz-1. These have significantly pushed our ability to understand biomolecular interactions, they have been widely adopted in industry and academia to help with drug development and protein design and they have opened the door to new research paradigms to push biological research further.

Designing Hardware Accelerators for Solving Sparse Linear Systems

2025-11-26T03:03:14Z

Designing Hardware Accelerators for Solving Sparse Linear Systems Feldmann, Axel Solving sparse linear systems is a key primitive that sits at the heart of many important numeric algorithms. Because of this primitive’s importance, algorithm designers have spent many decades optimizing linear solvers for high performance hardware. However, despite their efforts, existing hardware has let them down. State-of-the-art linear solvers often utilize < 1% of available compute throughput on existing architectures such as CPUs and GPUs. There are many different algorithms used to solve sparse linear systems. These algorithms are diverse and often have very different computational bottlenecks. These include low arithmetic intensity, fine-grained parallellism, tight dependences, and sparsity-induced load imbalance. This thesis studies the problem of designing hardware accelerators for sparse linear solvers. We propose three novel architectures that explore different parts of the design space. The accelerators exploit static sparsity as the basis of novel hardware-software co-designed scheduling approaches. First, we introduce Spatula, an architecture designed to accelerate direct solvers. Then, we propose Azul, a hardware accelerator targeted at iterative solvers. Taken together, Spatula and Azul demonstrate significant speedups on both of the main classes of sparse linear solver algorithms. Finally, to show that our techniques are useful for end-to-end applications, we present Ōmeteōtl, an accelerator targeted at applications that use iterative solvers in their inner loop. Ōmeteōtl also shows that the techniques in this thesis generalize to sparse matrix computations beyond linear solvers. These accelerators deliver order-of-magnitude speedups over state-of-the-art GPU baselines, achieving > 100× speedups on many inputs.

Physics-Optimized Design of 3D Shapes with Part-Based Control

2025-11-26T03:06:39Z

Physics-Optimized Design of 3D Shapes with Part-Based Control Zhan, Sean We introduce PhysiOPart, a computational approach for rapid generative design of 3D objects optimized for physical integrity. PhysiOPart enables users to edit and combine object parts to explore a vast design space. To model continuous surfaces of arbitrary resolution without topology restrictions, we parametrize parts with neural implicit representations. However, when parts are assembled to form an object, the resulting geometry is not guaranteed to be functional. Existing generative modeling approaches use task-specific neural predictors to approximate physical behaviors with limited accuracy. We propose an end-to-end differentiable physics simulation pipeline that performs linear static analysis to optimize for user-specified objectives, leveraging learned geometry priors. Our part-based formulation with finite element method is highly customizable, allowing for user-defined per-part materials, loads, and boundary conditions. The optimized designs exhibit improved physical behavior and can be fabricated.

Fast Assembly of Curved Structures from Flat Configuration

2025-11-26T03:06:43Z

Fast Assembly of Curved Structures from Flat Configuration Zaman, Akib Imagine deploying an emergency shelter that transitions seamlessly from a flat configuration to a lifted structure, or a folded robot that is sent through a tunnel and subsequently activated to expand into a larger form at the endpoint, with a single, collective pull of strings. This scenario raises two critical questions: (i) how to decompose the structure into a flat state that encodes the 3D geometry, and (ii) where to place strings through the unit modules to achieve complete actuation. Although these questions have been explored individually, comprehensive solutions remain scarce. To address this challenge, this thesis presents a computational approach for designing freeform structures that can be rapidly assembled from initially flat configurations by a single string pull. Target structures are decomposed into rigid, spatially varied quad tiles optimized to approximate a user-provided surface, forming a flat mechanical linkage. A two-step algorithm is then applied to determine a physically realizable string path that controls only a subset of tiles, enabling smooth actuation from flat to assembled configuration. First, the minimal subset of tiles required for string control is computed by considering both the structure’s geometry and inter-tile interactions. Second, a valid string path is identified through these tiles that minimizes friction, thereby transforming the flat linkage into the target 3D form upon tightening a single string. The resulting designs can be manufactured in flat form using computational fabrication techniques: such as 3D printing, CNC milling, or molding, thereby simplifying both production and transportation. Validation is provided through a series of physical prototypes and application case studies, ranging from medical devices and space shelters to large-scale architectural installations.

Theoretical Foundations for Learning in Games and Dynamic Environments

2025-11-26T03:04:04Z

Theoretical Foundations for Learning in Games and Dynamic Environments Golowich, Noah Decision-making problems lie at the heart of numerous aspects of human and algorithmic behavior across our society, ranging from healthcare systems to financial systems to interactions with the physical world. A central challenge that arises across many decision-making problems is the presence of multiple agents, often with competing incentives. To understand how agents will act in such situations, it is often productive to compute equilibria, which have the property that no agent can deviate from them and improve their utility. An additional challenge is that decisions made by agents often change the state of the environment, which is modeled as dynamic. Thus, we need efficient algorithms for learning good policies, which tell the agent what to do as a function of the environment’s state. Extensive work spanning multiple domains such as economics, computer science, and statistics has been developed to model these decision-making problems. This has led to many celebrated results, which include, for instance, a considerable body of work studying the computational properties of Nash equilibria in normal-form games, and a long line of papers on reinforcement learning. However, many of these classical works suffer from a few shortcomings: first, they often do not account for the enormous state or action spaces available to agents in realistic decision-making settings, and second, many of them do not derive computationally efficient algorithms for the desired solution concepts. These shortcomings are brought to the forefront by the remarkable recent progress in artificial intelligence, which holds promise for solving decision-making problems with enormous state or action spaces but which is often bottlenecked by computation. The objective of this thesis is to develop theoretical foundations for the computational aspects of such decision-making problems: e.g., How do we efficiently compute equilibria in large games?, and: How can we efficiently learn near-optimal policies in complex environments? Some highlights of our results are listed below—first, we study problems in which there are multiple agents and the goal is to compute some notion of equilibrium: • We show the first near-optimal rate of convergence to equilibrium for a no-regret learning algorithm in normal-form games, resolving a decade-long line of work which had aimed to establish increasingly better rates. • We establish the first algorithm with sublinear swap regret against arbitrary adversaries enjoying only polylogarithmic dependence on the number of actions, resolving a question of Blum and Mansour from 2007. • As a corollary of the preceding result, we obtain the first polynomial-time algorithm for approximating a correlated equilibrium in extensive-form games (to constant approximation error), addressing a question of von Stengel & Forges from 2008. Additionally we obtain near-optimal bounds on the communication and query complexity of approximating correlated equilibria in normal-form games (to constant approximation error), addressing several open problems in the literature. • We give the first algorithm for the sequential calibration problem with calibration error beating that of the seminal work of Foster & Vohra from 1998. Moving on to decision-making problems where the environment is modeled as dynamic (typically studied in the framework of reinforcement learning (RL)), our results include the following: • We give the first end-to-end computationally efficient algorithms for learning a nearoptimal policy in many fundamental reinforcement learning problems, such as those of (constant-action) Linear Bellman Complete MDPs and sparse linear MDPs. • We give the first quasi-polynomial time algorithm for finding a near-optimal policy in a general and well-motivated class of partially observable RL environments, and show that our bound is tight. • We prove some (perhaps surprising) hardness results that arise in multi-agent RL problems. For instance, we show that it is computationally hard to implement noregret learning algorithms in multi-agent RL environments even when the agents can coordinate on their choice of algorithm, which creates a stark contrast with simpler multi-agent learning settings (e.g., in normal-form games) where no-regret learning has formed the bedrock for a wide array of developments over the last several decades. • Nevertheless, we show that by adjusting the type of equilibrium appropriately, we can circumvent the above hardness results and derive computationally efficient decentralized algorithms for computing equilibria in multi-agent RL environments. Many of the above results have inspired follow-up work which includes applications of our results to various problems in game theory, reinforcement learning, online learning, and related domains, as well as the formulation of new problems which are inspired by the above results.

Characterizing human vision through large-scale brain imaging and computational models

2025-11-26T03:04:19Z

Characterizing human vision through large-scale brain imaging and computational models Lahner, Benjamin Efforts to understand the neural underpinnings of human visual processing require sufficient experimental data and robust models. This thesis significantly contributes to both these fronts while simultaneously elucidating some of the most intriguing aspects of the human visual system. In the first chapter, I use a combination of classical machine learning, artificial neural networks, and a joint MEG/fMRI neuroimaging dataset to reveal that the human visual system extensively processes highly memorable images in regions distributed throughout visual cortex late in time. In the second chapter, I present the BOLD Moments Dataset, a large-scale fMRI dataset using short video stimuli to extend computational models of visual processing into the video domain to better understand how humans process visual content unfolding over time. The last chapter introduces a fMRI dataset aggregation framework titled MOSAIC to achieve the scale and stimulus diversity needed for training modern neural networks directly on brain responses. This body of work exemplifies how large-scale experimental data and artificial neural networks can contribute towards a robust and generalizable understanding of human visual processing.

Wireless Systems for a Sustainable Future: From Battery-Free Subsea IoT to THz-Based Agriculture Monitoring

2025-11-26T03:03:54Z

Wireless Systems for a Sustainable Future: From Battery-Free Subsea IoT to THz-Based Agriculture Monitoring Afzal, Sayed Saad This thesis describes how wireless sensing can drive significant advancements in climate and sustainability. Specifically, it shows how we can leverage diverse signals—acoustics, ultrasound, THz, and optics— in unconventional ways to unlock new capabilities in underwater climate monitoring, food safety, and disaster response. The thesis introduces two novel technologies. The first technology enables long-term, ultra-low power ocean sensor networks for use in climate modeling, marine monitoring, and sustainable aquaculture. Unlike existing IoT technologies – like Bluetooth, WiFi, and GPS – which cannot work underwater, we design and implement an ultra-low power subsea backscatter communication system, enabling battery-free underwater imaging, sensing and localization. Second, the thesis describes a new technology that can support sustainability in agriculture through real-time food quality assessment that reduces food waste. In contrast to existing food quality technologies that require direct contact with produce, we introduce a new wireless system for accurate, non-invasive sensing using sub-THz signals. We describe the design, implementation, and evaluation of multiple systems that leverage these technologies to monitor the ocean and food waste: First, we present a ultra-wideband metamaterial sensor design that facilitates scalable, and long-range battery-free underwater communication. Next, we describe a system that can push the throughput of this technology using higher order modulation. Beyond building sensor networks, we demonstrate their real-world potential through two systems: one for underwater localization that uses rich spatio-temporal-spectral features for accurate positioning, and another for battery-free imaging that fuses acoustic and optical signals to capture color images in the dark. Finally, we present a novel solution for accurate fruit ripeness sensing using sub-terahertz wireless signals. These systems unlock new IoT applications in climate modeling, aquaculture, robotics, and agriculture.

Automatic Integration and Differentiation of Probabilistic Programs

2025-11-26T03:04:14Z

Automatic Integration and Differentiation of Probabilistic Programs Lew, Alex K. This thesis addresses the challenge of automating fundamental operations from probability theory and calculus on probability distributions defined by higher-order probabilistic programs. It does this by developing a suite of composable program transformations for an expressive core calculus for probabilistic programming: • Integration: Compiling a probabilistic program into a deterministic representation of its expectation operator, handling potentially intractable integrals symbolically. • Unbiased estimation: Transforming programs involving intractable operations (like integration) into runnable probabilistic programs that yield provably unbiased estimates of the original value, with flexible levers for users to navigate cost-variance trade-offs. • Radon-Nikodym differentiation: Compiling probabilistic programs into implementations of a novel interface for the unbiased estimation of density ratios, of the sort that arise in Monte Carlo and variational inference. • Differentiation: Extending automatic differentiation (AD) to compose with the above transformations, enabling the optimization of expected values and density ratios of probabilistic programs. These transformations operate on an expressive higher-order probabilistic programming language and are proven correct using denotational semantics and logical relations. The resulting framework enables the sound and automated implementation of a wide range of algorithms for probabilistic inference and learning. To demonstrate the practical value of these techniques, we use them to implement three systems for scalable probabilistic inference in different domains: (1) extensions to the Gen probabilistic programming system that accelerate and automate a broad range of Monte Carlo and variational inference algorithms, (2) the PClean system for automated Bayesian reasoning about relational data, and (3) the GenLM system for controllable generation from language models. We find that our techniques enable these systems to scale to a variety of complex, real-world problems, and to achieve state-of-the-art performance on a range of benchmarks.

Minimizer-space computation

2025-11-26T03:03:41Z

Minimizer-space computation Ekim, Barış C. As the volume of DNA sequencing data increases, the need for algorithmic advances to efficiently handle the data arises. One such concept is minimizers, which are genomic substrings that allow for reduced representations of larger DNA sequences. In this thesis, we introduce minimizer-space computation as a new algorithmic paradigm for DNA sequence analysis. Instead of DNA nucleotides, we consider minimizers as the letters of an extended alphabet in which algorithms operate. We present several techniques on how to efficiently construct these extended alphabets, demonstrate how to develop approaches that use these alphabets and consequently use only a fraction of sequence data, and show how fundamental biological tasks, such as genome assembly and read mapping, can be significantly accelerated over state-of-the-art methods.

Performant and Resilient Service Composition for Modern Cloud Applications

2025-11-26T03:04:07Z

Performant and Resilient Service Composition for Modern Cloud Applications Li, Tianyu Modern cloud applications are often distributed systems composed from vendor-provided building blocks (e.g., object storage services, container orchestration services). Consequently, distributed fault-tolerance is a central concern for application correctness. Although each building block may offer individual fault-tolerance, the end-to-end application is still susceptible to failures, because the composition logic that orchestrates them may still fail. This thesis explores resilient composition, a systematic way to assemble fault-tolerant components into resilient end-to-end distributed applications. We begin by presenting the fail-restart system model, which captures the unique fault-tolerance challenges that arise when composing services. Based on this model, we define Composable Resilient Steps (CReSt), an atomic programming abstraction that guarantees fault-tolerance across the assembled application. We then detail efficient methods for implementing CReSt using a range of database techniques, and a novel distributed protocol that allow optimistic, speculative execution ahead of slower fault-tolerance safeguards. Together, these pieces allow developers to assemble fault-tolerant distributed systems that are correct by construction and often more performant than existing solutions.

Succinct Cryptography via Propositional Proofs

2025-11-26T03:03:29Z

Succinct Cryptography via Propositional Proofs Mathialagan, Surya The goal in modern cryptography is to obtain security while minimizing the use of computational resources. In recent years, we have been incredibly successful in our pursuit for efficiency, even for cryptographic tasks that were thought to be “science fiction”. For example, we have constructions of fully homomorphic encryption and private information retrieval from standard, cryptographic assumptions which achieve the ideal levels of succinctness. However, there are still some tasks in cryptography where achieving the “ideal” efficiency from standard assumptions has evaded us. In this thesis, we study the problem of achieving succinctness in two such settings: • Can we construct succinct indistinguishability obfuscation (IO) for Turing machines? In particular, can we construct an obfuscated program whose size is independent of the input length? • Can we construct succinct non-interactive arguments (SNARGs) for all of NP? While the problems seem unrelated at first glance, the root difficulty seems to stem from a similar place: both primitives have non-falsifiable security definitions. In fact, this type of barrier exists for many other cryptographic primitives, including witness encryption. This leads to a central question which we refer to as the “non-falsifiability barrier”: how can we construct non-falsifiable primitives from falsifiable assumptions? In this thesis, we show how to leverage propositional proofs to overcome the non-falsifiability barrier, and make substantial progress in the goal of achieving succinctness in both settings. Our main result is universal construction of both SNARGs and succinct IO for Turing machines from standard assumptions using propositional proofs. We then show several applications, including rate-1 IO for many programs, the first succinct secret sharin schemes for monotone circuits, and many more. Our results establish propositional proofs as a foundational tool for achieving succinctness across a broad range of cryptographic settings.

Statistical and Algorithmic Thresholds in Spin Glasses

2025-11-26T03:03:59Z

Statistical and Algorithmic Thresholds in Spin Glasses Huang, Brice This thesis studies spin glasses, disordered complex systems originating in statistical physics. Such systems model optimization, sampling, and inference problems from probability and statistics, which are of fundamental importance to modern data science. In particular, spin glasses provide natural examples of random, high-dimensional, and often highly non-convex cost or log-likelihood functions, making them an excellent testing ground for such questions. Part I of this thesis studies statistical properties of these models. Chapter 2 identifies the storage capacity of the Ising perceptron, a simple model of a neural network, subject to a numerical condition. This gives a conditional proof of a 1989 conjecture of Krauth and M´ezard. Chapter 3 gives a new proof of the celebrated Parisi formula for the free energy of the spherical mean-field spin glass, which was first proved by Talagrand and in more generality by Panchenko. Our proof takes a simpler modular approach, drawing on recent advances in spin glass free energy landscapes due to Subag. Chapter 4 characterizes the topology trivialization phase transition of multi-species spherical spin glasses and shows that lowtemperature Langevin dynamics finds the ground state in the topologically trivial regime; the latter result is new even in the single-species setting. Part II of this thesis concerns algorithms for optimization and sampling problems on spin glasses. Chapter 5 studies the problem of optimizing the Hamiltonian of a multi-species spherical spin glass. Our main result exactly characterizes the maximum value attainable by a class of algorithms that are suitably Lipschitz in the disorder. This class includes gradient-based algorithms and Langevin dynamics on constant time scales, and in particular includes the best algorithm known for this problem. This chapter is part of a series of works where we establish exact algorithmic thresholds using the branching overlap gap property (OGP), a landscape property introduced in our earlier work (which appears in our S.M. thesis). In this chapter, we develop a more robust way to establish the branching OGP that does not require Guerra’s interpolation; this allows our method to be applied to models well beyond the (single-species) mean-field spin glass we previously considered. Chapters 6 and 7 study sampling from the Gibbs measure of a spherical mean-field spin glass. Chapter 6 develops a sampling algorithm based on simulating Eldan’s stochastic localization scheme, while Chapter 7 analyzes simulated annealing of Langevin dynamics. We prove both algorithms succeed for inverse temperatures up to a stochastic localization threshold. Chapter 6 gives the first stochastic localization-based sampler with a guarantee of vanishing total variation error, improving on earlier algorithms with vanishing Wasserstein error. Chapter 7 provides the first provable guarantees for a Markov chain in this model beyond the uniqueness threshold, where mixing from worst-case initialization is provably slow.

A Cavity-Coupled Rydberg Atom Array for Quantum Science and Quantum Computing

2025-11-26T03:03:16Z

A Cavity-Coupled Rydberg Atom Array for Quantum Science and Quantum Computing Hu, Beili Neutral atom arrays have rapidly emerged as a leading platform for quantum computing, boasting scalable, configurable arrays of single atoms trapped in optical tweezers, fast, high-fidelity entangling gates through Rydberg interactions, and programmable, parallelized control of qubit operations. Coupling an atom array to an optical cavity opens a new frontier. Leveraging enhanced light-atom interactions in cavity quantum electrodynamics, cavity- coupled atom arrays acquire capabilities that can further expand the neutral atom toolbox, including cavity-enhanced atom readouts, atom-photon entanglement, and photon-mediated interactions between distant atoms. This thesis presents a quantum hardware platform that integrates an array of neutral atoms with a high-finesse optical cavity. After describing the design and development of the experimental apparatus, I demonstrate high-fidelity atom state readout through the cavity, achieving improved speed and atom survival compared to conventional free-space imaging methods. I then introduce a new technique for selectively controlling atom-cavity coupling on arbitrary subsets of the array, using local AC Stark shifts on the excited states of the atoms. Building on these tools, I demonstrate fast, non-destructive cavity-based readout of atom arrays, a crucial bottleneck of atom array platforms. I also showcase real-time measurement and feedback capabilities with a demonstration of classical error correction, using a register of atomic bits. Finally, I describe progress toward implementing single- and two-qubit gates within the cavity-coupled system. By combining coherent control, tunable interactions, and high-fidelity, non-destructive readout integrated and real-time feedback, the cavity-coupled Rydberg atom array offers a promising path toward fault-tolerant quantum computing.

Estimation, Prediction and Counterfactual Inference with Dependent Observations

2025-11-26T03:03:46Z

Estimation, Prediction and Counterfactual Inference with Dependent Observations Kandiros, Anthimos Vardis The success of modern data science is largely driven by access to large-scale, high dimensional data. Much of classical machine learning has been developed under the assumption that this data is generated independently from some distribution. However, this assumption is often violated when data exhibit complex dependencies across a spatial or temporal domain, or due to social interactions. In this thesis, our goal is to design and analyze methods that address these dependencies for performing three fundamental estimation tasks: unsupervised learning, supervised learning and counterfactual inference. In supervised learning, we observe a sequence of unlabeled examples and our goal is to infer some structural property from the distribution they came from. The presence of dependencies could severely complicate this question. Our results in this direction encompass both fully observable as well as latent variable models. For fully observable models, we use the celebrated Ising model to describe the dependencies. Assuming we have access to a single sample from some Ising model, which captures a variety of real-world scenarios, we design and analyze polynomial time algorithms for recovering the matrix corresponding to the network structure of the model. We then leverage these techniques to obtain improved guarantees for estimating Ising models in Total Variation (TV) distance from multiple samples. For latent variable models, we focus on the case where the structure is a tree and we get samples from the leaves, which is a common scenario in phylogenetics. Assuming the model is Gaussian, we analyze the behavior of the Expectation-Maximization (EM) algorithm, a popular heuristic for latent variable models. We show that for trees with a single latent node, EM converges to the true model and for general tree topologies, the only stationary point in the interior of the domain is the true model. We then shift our focus to discrete models and study latent tree Ising models, for which we provide polynomial time algorithms for learning the distribution of leaves in TV distance. In supervised learning, we observe a sequence of feature-label pairs and our task is to learn the predictive relationship between the features and the labels. Here, this relationship could be confounded by the presence of dependencies among labels. We formulate this question as a regression problem, where the labels of the units follow the joint distribution of an Ising model with an unknown strength parameter and external fields that are determined by the regression function. We characterize the minimax optimal rate of estimation for the various parameters and provide an efficient algorithm that achieves it. Interestingly, it might not be possible to estimate all the parameters in some cases. In counterfactual inference, we focus on the design of network experiments, where the treatment of a unit could affect the outcome of a neighboring unit in an underlying graph. Our goal is to estimate a general causal effect that can be defined as the average difference in outcomes for a unit under two different interventions. For an arbitrary such effect, we propose an experimental design, called the conflict graph design. For an unbiased estimator of that effect, we prove bounds on its variance that yield the best known rates of estimation for various effects studied in the literature, such as the average direct effect and the total effect, but also provide estimation rates for effects that have received less attention from the perspective of experimental design.

Hardening Trusted Execution Environments Against Microarchitectural Side-Channel Attacks: A Constructive Approach

2025-11-26T03:03:06Z

Hardening Trusted Execution Environments Against Microarchitectural Side-Channel Attacks: A Constructive Approach Dréan, Jules Guillaume Jacques Bénony D Trusted Execution Environments (TEEs) [1–5] promised to enable secure computation even in the presence of privileged adversaries by providing hardware-enforced isolation. However, the discovery of microarchitectural side-channel and transient execution attacks [6–10] has severely undermined these security guarantees. These attacks exploit shared hardware resources and speculative execution to leak sensitive information across security boundaries, effectively bypassing the architectural isolation enforced by TEEs. The widespread impact of these vulnerabilities is evidenced by more than 43 published attacks [11] targeting commercial TEE platforms including Intel SGX, AMD-SEV, and ARM TrustZone. Existing approaches to defend against these attacks face significant limitations. Hardware-based solutions [12–14] often require complex processor modifications with significant hardware overhead. Replacing trusted hardware with cryptographic approaches incurs prohibitive performance overheads [15]. Meanwhile, formal verification methods struggle to scale to realistic code base sizes and often fail to capture subtle microarchitectural behaviors [16–18]. This thesis proposes a constructive approach to TEE security and demonstrates that practical defenses against microarchitectural attacks are achievable through careful system design. Rather than relying only on models and simulations, we focus on constructing systems that are secure by design. Our work is concretely realized through the design, implementation, and evaluation of two novel platforms: First, we present Citadel, a TEE platform that enables secure shared memory while providing precise guarantees against microarchitectural side-channel attacks. Citadel introduces relaxed microarchitectural isolation (RMI), a novel security property that allows programs to share memory while restricting information leakage to that of a non-speculative execution. To achieve RMI, Citadel combines hardware-enforced microarchitectural isolation with two simple mechanisms for controlled speculation: SpecSafe, which prevents speculative shared-memory accesses entirely, and Burst mode, which enables better performance through constrained speculation on small code snippets. Through a fully functional FPGA prototype, we demonstrate that Citadel can run real-world applications including cryptographic libraries and private ML inference with less than 5% overhead while maintaining strong security guarantees. Second, we develop Argos, an “integrity-only” TEE specifically designed for verifiable fully homomorphic encryption, that enables the deployment of FHE schemes in real-world settings where malicious security is required. We show that by carefully constraining the attack surface and employing simple hardware mechanisms, we can achieve complete security against microarchitectural attacks. Argos introduces a simplified transcript-based attestation scheme that only requires one signature per FHE computation, amortizing the cost of relying on a physical TPM to microarchitecturally isolate secrets. Argos can be used to not only enforce circuit-level integrity of FHE schemes but can also be extended to support more complex FHE-based applications that take (potentially poisoned) input from the (malicious) circuit evaluator. Argos is compatible with commodity hardware and only incurs minimal performance overhead with an average of 3% overhead for FHE evaluation and 8% overhead for complex protocols. Through these systems, we show that effective defenses can be built against microarchitectural side channel and transient execution attacks. Our constructive approach yields practical systems that are secure by design while maintaining efficiency and usability. This thesis opens new possibilities for the deployment of trusted hardware by demonstrating concrete paths toward robust microarchitectural security.

Steering Robots with Inference-Time Interactions

2025-11-26T03:03:48Z

Steering Robots with Inference-Time Interactions Wang, Yanwei Imitation learning has driven the development of generalist policies capable of autonomously solving multiple tasks. However, when a pretrained policy makes errors during deployment, there are limited mechanisms for users to correct its behavior. While collecting additional data for finetuning can address such issues, doing so for each downstream use case is inefficient at deployment. My research proposes an alternative: keeping pretrained policies frozen as a fixed skill repertoire while allowing user interactions to guide behavior generation toward user preferences at inference time. By making pretrained policies steerable, users can help correct policy errors when the model struggles to generalize—without needing to finetune the policy. Specifically, I propose (1) inference-time steering, which leverages user interactions to switch between discrete skills, and (2) task and motion imitation, which enables user interactions to edit continuous motions while satisfying task constraints defined by discrete symbolic plans. These frameworks correct misaligned policy predictions without requiring additional training, maximizing the utility of pretrained models while achieving inference-time user objectives.

Advancing Deep Learning Efficiency: From Specialized Co-Design to Automated Generation

2025-11-26T03:03:34Z

Advancing Deep Learning Efficiency: From Specialized Co-Design to Automated Generation Lin, Yujun The explosive growth of artificial intelligence (AI) technologies, particularly large-scale deep learning models such as large language models and diffusion models, has intensified the demand for efficient full-stack inference solutions that effectively balance performance and costs. This work will present a comprehensive exploration into the algorithm-system co-optimization, hardware design specialization and automation for scalable AI deployment. First, we begin with algorithmic optimization for large-scale models, including large language models and diffusion models, developing inference libraries that leverage quantization to boost the performance of generative AIs on existing GPU platforms. Next, we design specialized hardware accelerators for domain-specific applications, specifically point cloud understanding, emphasizing efficiency improvements through the exploitation of data sparsity. Finally, we open up the hardware design space beyond template-based sizing, and progress into the automated learning-based co-design of neural network and hardware architectures, maximizing their synergy with a full-stack joint optimization. We then introduce an automated framework for spatial accelerator generation, transforming high-level mappings into custom hardware designs that support scalable deployment. Together, these contributions advance AI inference efficiency by bridging the gap between advanced computational requirements and hardware capabilities, between theoretical potential and practical solutions, and between design cost and effectiveness.

Learning Theoretic Foundations for Understanding Quantum Systems

2025-11-26T03:03:55Z

Learning Theoretic Foundations for Understanding Quantum Systems Liu, Allen Understanding and harnessing the power of quantum systems has the potential to transform many domains in science and technology. However, before we can achieve these aspirations, we must first build a better understanding of how quantum systems fundamentally behave. In this thesis, we approach this question through the lens of learning theory to develop new paradigms for learning about quantum systems and understanding their structural properties. We deliver several surprising results, upending previous beliefs about even fundamental laws and giving provably efficient algorithms for learning about quantum systems in settings previously conjectured to be intractable. Typically in quantum many-body systems, the particles in the system interact locally with respect to some geometry as described by a local Hamiltonian. Two key questions are first, understanding equilibrium properties of a system with a given Hamiltonian and second, recovering the Hamiltonian from measurements of the properties of the system. For the first, we prove a universal law that there is a sudden death of entanglement, at a critical temperature depending only on the geometry but not on the system size. For the second, we give the first efficient algorithm for recovering the Hamiltonian at any temperature, breaking a conjectured barrier at low temperatures. Beyond systems with local interactions, we also consider learning and testing properties of general quantum states, focusing on the interplay between statistical complexity and near-term quantum device constraints, only allowing for entangled measurements over a limited number of copies of the state. We characterize the optimal rates for learning and testing with single-copy measurements and for multi-copy measurements in many relevant near-term regimes.

Domain Wall Based Magnonics in Iron Garnet

2025-11-26T03:03:43Z

Domain Wall Based Magnonics in Iron Garnet Gross, Miela J. Magnonic devices leverage magnons, quantized spin waves, as the mechanism to process and transfer information. In materials with low Gilbert damping, these spin wave-based systems enable ultra-fast operation while eliminating thermal heating and leakage currents inherent to conventional electron-based microelectronics. To maximize energy efficiency and processing speed, materials like iron garnets, ferrimagnetic insulators with tunable magnetic properties, are essential. Key magnetic parameters, including saturation magnetization, perpendicular magnetic anisotropy, coercivity, and Gilbert damping, can be tailored through elemental substitution or strain engineering in thin films. Furthermore, relativistic domain wall velocities reported in yttrium iron garnet (YIG), bismuth substituted YIG, and thulium iron garnet lay the foundations for high-speed operation. These unique attributes position garnets as ideal materials for the development of magnonic devices that integrate efficiency, speed, and versatility. This thesis presents my research on integrating thin film garnets into a domain wall based magnonic devices. It begins by exploring the magnetic characterization of thin film iron garnets, including the growth process, temperature dependent magnetic behavior, and tunable magnetic anisotropy. Next, we report on magnonics within the garnet, focusing on the interactions between spin waves and domain walls. Finally, we demonstrate a write mechanism for a magnonic device driven by spin wave-induced domain wall motion, providing detailed characterization of the device behavior and performance. These results underscore the potential of iron garnets for magnonic-based device applications and offer insights into the efficiency of write mechanism, paving the way for energy-efficient high-speed spintronic technologies.

Mitigating Inhomogeneity in High-Field MRI Excitations: Arbitrary Waveform Optimization and Multiphoton Parallel Transmission (MP-pTx)

2025-11-26T03:03:22Z

Mitigating Inhomogeneity in High-Field MRI Excitations: Arbitrary Waveform Optimization and Multiphoton Parallel Transmission (MP-pTx) Drago, John M. High-field magnetic resonance imaging (MRI) using a standard volume coil results in a spatially varying flip angle across the body, which renders images difficult to clinically interpret. This arises from the complex interactions of electromagnetic fields from current-carrying elements surrounding the imaging region. Parallel transmission (pTx) mitigates this issue by employing multiple high-power, independently controlled transmit elements for more precise excitation control. However, since the wavelength of the applied radio waves is shortened in tissue, the effect becomes highly dependent on the patient’s anatomy. As a result, optimization must be performed on a patient-by-patient basis, and methods that attempt full control of these independent waveforms are too computationally intensive to execute during the limited examination time. Additionally, the high-field excitations create complex electric field distributions that require control and careful monitoring to avoid excessive tissue power deposition (and ultimately heating), quantified as the specific absorption rate (SAR). To address these challenges, we introduce a method for optimizing patient-specific pulses using a global waveform (Ritz) approach, enabling rapid, in-scanner optimization. While pTx effectively addresses flip angle inhomogeneity, it remains costly and introduces challenges in SAR management. We address the SAR management and cost problems of pTx by introducing and characterizing the MP-pTx method, which leverages the multiphoton phenomenon to improve homogeneity using a standard volume coil supplemented with low-frequency (kilohertz) parallel channels. MP-pTx reduces costs and simplifies SAR management by shifting the parallel irradiation to low-cost, lowSAR shim array channels. These channels supplement an off-resonant excitation from a conventional birdcage coil with an oscillating, z-directed field that satisfies the resonance condition for spin state transitions.

Generative Latent Motion Planning and Reinforcement Learning for Legged Locomotion

2025-11-26T03:03:52Z

Generative Latent Motion Planning and Reinforcement Learning for Legged Locomotion Miller, Adam Joseph In recent years, reinforcement learning has demonstrated its promise as a powerful tool for developing innovative and advanced control systems for legged robots. The method’s robustness, versatility, and generality have made it a prime candidate for future robotic systems deployed in the real world. Through the development of more advanced machine learning algorithms and more reliable and efficient physics simulators, reinforcement learning continues to improve and enable new, dynamic, and agile capabilities. While the results are often impressive and the tools relatively beginner-friendly, there remain impediments to scalable and reliable progress. Poor reward function scaling, challenges balancing exploration versus exploitation, and misalignment from the engineer’s intent are roadblocks to better performance. To get beyond these limitations, new tools and frameworks are necessary. In this work, I present novel methods to address these challenges and extend the capabilities of reinforcement learning on robot hardware. Through the quantification of the distributional sim-to-real gap, simulation model optimization for hardware matching, latent space motion sequence planning, and latent style training, I demonstrate never-before-seen performance on legged hardware.

Learning from Weak Supervision: Theory, Methods, and Applications

2025-11-26T03:03:36Z

Learning from Weak Supervision: Theory, Methods, and Applications Lang, Hunter The growing demand for high-quality labeled data to train machine learning models has driven widespread adoption of weak supervision and synthetic data methods, which use automated models instead of humans for annotation. Large language models (LLMs) have further accelerated this trend because their zero- and few-shot classification performance enables them to serve as effective “synthetic annotators” for various tasks. In practice, the data generated by these weak annotators is imperfect, but it enables the training of strong models. However, theoretical understanding of why training one model on the outputs of another leads to strong performance remains limited, especially when the annotator model exhibits suboptimal performance on the target task. In this thesis, I develop a theoretical framework for learning from weak supervision that captures the key aspects of the problem better than existing approaches in the crowdsourcing and learning-with-noisy-label literature. This framework establishes structural conditions that explain when and why weak supervision can reliably train strong models. Building on these theoretical results, the second part of the thesis introduces methods to improve how models learn from weak supervision and applies these methods to low-labeled-data settings.

Multi-fidelity Optimal Trajectory Generation: Optimal Experiment Design for Robot Learning

2025-11-26T03:03:10Z

Multi-fidelity Optimal Trajectory Generation: Optimal Experiment Design for Robot Learning Ryou, Gilhyun Data-driven methods have significantly advanced robot learning, yet their direct application to real-world robots remains challenging, particularly under extreme conditions. This challenge is especially pronounced for highly maneuverable vehicles like quadrotor aircraft, which often operate in scenarios requiring rapid maneuvering, such as racing, defense systems, or safety-critical obstacle avoidance. In such extreme conditions, real-world constraints like control delays, state estimation errors, and battery voltage fluctuations often compromise trajectory reliability, even when conforming to ideal dynamics. However, the typical data-driven methods are usually developed in simulated environments. Consequently, the transition to real-world dynamics requires extensive fine-tuning, which can be risky, as perfect training in simulations does not guarantee safe transitions to real-world dynamics. This thesis employs methods from optimal experiment design to address these challenges. By quantifying uncertainty and maximizing information gain, the approach aims to safely and efficiently design the real-world experiments required for accurate constraint modeling. In the first chapter, we present a multi-fidelity Bayesian optimization method that searches for time-optimal speed profiles for quadrotor aircraft, effectively balancing numerical simulations with real-world flight experiments. The second chapter extends the optimal experiment design method to a high-dimensional online planning problem through integration with reinforcement learning. The proposed algorithms, trained and validated through real-world flight experiments, significantly outperform baseline methods in trajectory time and computational efficiency. Additionally, these algorithms have been adapted to various planning problems, including fixed-wing aircraft planning, cooperative multi-drone systems, and energy-efficient trajectory generation.

Efficient Systems for Large-Scale Graph Representation Learning

2025-11-26T03:03:39Z

Efficient Systems for Large-Scale Graph Representation Learning Huang, Tianhao Graph representation learning has gained significant traction in critical domains including finance, social networks, and transportation systems due to its successful application to graphstructured data. Graph neural networks (GNNs), which integrate the power of deep learning with graph structures, have emerged as the leading methods in this field, delivering superior performance across diverse graph related tasks. However, training graph neural networks on large-scale datasets encounters scalability challenges on current system architectures. First, the sparse, non-localized structures of real-world graphs lead to inefficiencies in data sampling and movement. This characteristic heavily stresses system input/output (I/O), particularly burdening the peripheral buses during the sampling phase of GNN training. Second, the suboptimal mapping of training procedure to GPU kernels leads to compute inefficiencies, including substantial kernel orchestration overhead and redundant computations. Addressing these challenges requires a comprehensive, full-stack optimization approach that fully leverages hardware capabilities. This thesis presents two complementary works to achieve the goal. The first work, Hanoi, unblocks the data loading bottleneck in out-of-core GNN training by co-designing the sampling algorithms to align with the hierarchical memory organization of commodity hardware. Hanoi drastically reduces I/O traffic to external storage, delivering up to 4.2× speedup over strong baselines with negligible impacts on the model quality. Notably, Hanoi is able to obtain competitive performance close to in-memory training with only a fraction of memory requirements. Building on this foundation, the second work, Joestar, introduces a unified framework for optimized GNN training on GPUs. Joestar adapts the multistage sampling approach from Hanoi to in-memory training which frees CPUs from heavy data loading workloads. Joestar also identifies novel kernel fusion opportunities and formulates better execution schedules by jointly considering the sampling and compute stages. Combined with compiler infrastructure in PyTorch, Joestar achieves state-of-the-art GNN training throughputs for billion-edge graph datasets on a single GPU.

Generalizable Long-Horizon Robotic Manipulation under Uncertainty and Partial Observability

2025-11-26T03:03:25Z

Generalizable Long-Horizon Robotic Manipulation under Uncertainty and Partial Observability Curtis, Aidan A central goal in embodied artificial intelligence is to enable autonomous agents to accomplish complex, long-horizon tasks in novel, partially observable environments. In these scenarios, agents must effectively reason about uncertainty, generalize from limited experiences, and proactively plan actions to acquire missing information. This thesis tackles these core challenges by developing and evaluating novel methods specifically designed for partially observable contexts. The first part of this thesis introduces an enhanced heuristicguided planning technique that increases search efficiency in sparse-reward domains with significant uncertainty. Next, we investigate how symbolic reasoning can be integrated into the decision-making framework, accelerating search through the use of temporal and belief-space abstractions. Next, we propose a method for sequencing low-level reinforcement learning skills alongside information gathering actions, enabling increased task complexity and robustness in real-world tasks. Lastly, we show how large language models may be leveraged for few-shot model learning, allowing agents to rapidly adapt and generalize to new scenarios. The methods presented in this thesis advance the state-of-the-art in embodied AI by enabling robots to better handle uncertainty and incomplete information, ultimately paving the way for more capable, exploratory, and risk-aware autonomous systems.

Graph-based Vector Search Algorithms for Retrieval-Augmented AI Systems

2025-11-26T03:06:31Z

Graph-based Vector Search Algorithms for Retrieval-Augmented AI Systems Zhang, Ziyu The recent advancement of large language models (LLMs) and large multimodal models (LMMs) greatly enhances the capabilities of AI systems such as recommendation systems and coding assistants, making them more practical for real-world deployment. However, these models cannot directly interact with large volumes of data in a knowledge corpus during inference/task time due to inherent architectural limits and cost concerns. Encoding data into vector embeddings and leveraging approximate nearest neighbor search (ANNS) have thus become an important data processing primitive in AI systems following the introduction of retrievel-augmented generation (RAG). However, the complexity of tasks these AI systems aim to solve introduces challenges for existing ANNS algorithms. I developed methods to expand existing ANNS algorithms to address two such challenges: freshness and heterogeneity in the data. Graph-based ANNS algorithms have been proven to have superb cost versus approximation quality trade-off yet follow a simple intuition of best-first search. I focus on adapting graph-based ANNS algorithms to two settings featuring emerging challenges. (1) Data is updated constantly. Existing algorithms are inefficient under deletions and not robust against different orderings in the workload. I propose methods addressing these problems and developed an algorithm supporting updates effectively and efficiently based on Vamana, a state-of-the-art graph-based ANNS algorithm. (2) Data is heterogeneous in format, modality, and how they relate to a query, making the similarity difficult to capture by the canonical ANNS definition. I explore ways to model the similarity between heterogeneous sources and using graph-based ANNS approaches to perform semantic search in this setting. I test this approach under an end-to-end multimodal question-answering system developed in-house.

Adaptive Abstractions for Robust Hierarchical Manipulation Planning

2025-11-26T03:03:38Z

Adaptive Abstractions for Robust Hierarchical Manipulation Planning Noseworthy, Michael S. In this thesis, we address the problem of long-horizon robotic manipulation under partial observability. Tasks such as gearbox assembly or tidying a workstation involve many objects and necessitate a variety of manipulation capabilities. These long-horizon tasks are commonly addressed by hierarchical approaches, which introduce state and action abstractions to make planning tractable. However, our abstractions often rely on imperfect models of the world, which can lead to brittle execution. Furthermore, these abstractions depend on having accurate state information, which is often only noisily sensed, if sensed at all. For example, in the assembly domain, the pose of each part may only be known within a few millimeters, and a box’s mass distribution may be completely unsensed. To deploy robots outside of structured environments like the factory, they will need to be robust to model misspecification and partial observability. The central idea of this thesis is that we can develop adaptive abstractions to improve the robustness of hierarchical planning once the robot is deployed. Adaptive abstractions incorporate observations from the real world that are informative about misspecifications and partial observability, essentially allowing the planner to adapt to its deployment environment. We explore this idea by developing three types of models that enable this adaptivity at different levels of the abstraction hierarchy: plan feasibility models, adaptive samplers, and reactive control policies. In our first contribution, we consider adding adaptivity to a task and motion planning system at the task-planning level. We focus on a setup where the robot has access to a set of parameterized skills, but these skills are derived from imperfect models. To enable robust planning, we propose to autonomously learn skill feasibility models once the robot is deployed through a curious exploration phase. Critically, we propose a novel active learning framework to enable efficient learning without human intervention. We show that the resulting feasibility model leads to robust task performance on multiple downstream tasks in a stacking domain. Our second contribution looks at developing adaptive samplers that can incorporate information about object state that is typically unobserved (e.g., inertial and frictional properties). General-purpose belief representations can handle this partial observability, but online inference is computationally expensive. Instead, we propose to use an offline phase to learn an inference network that directly predicts a distribution over object properties that is consistent with the interaction history. We show that inference networks enable efficient adaptation in a grasping domain with heavy objects. Our final contribution focuses on learning adaptive controllers such that robustness is handled at the lowest level of the abstraction. We consider precise contact-rich manipulation tasks that are sensitive to pose estimation errors. To overcome noisy poses at the control level, explorative contact is necessary, but unintended forces can lead to catastrophic outcomes such as part slippage or damage. We propose to use simulation in an offline phase to train reactive force-aware policies. The policies are trained to overcome pose uncertainty while using force-sensing to adaptively limit excessive forces. The result is robust real-world performance on the multistage assembly of a planetary gearbox system, which includes insertion, gear-meshing, and nut-threading tasks. In summary, adaptive abstractions can be used to increase the robustness of hierarchical manipulation planning, an important step in deploying robots outside of the lab or factory. Throughout the thesis, we validate the proposed approaches on the real robot in stacking and assembly domains.

Methods for Generalization Under Distribution Shift

2025-11-26T03:03:12Z

Methods for Generalization Under Distribution Shift Netanyahu, Aviv Machine learning systems have achieved remarkable performance in tasks where test data closely resembles the training distribution. However, real-world applications often require systems capable of handling more challenging situations -- specifically, adapting to new tasks and extrapolating to data points outside the distribution of the training set. The current paradigm for handling distribution shifts is collecting and training models on large datasets. This work offers two more principled frameworks that enable machine learning models to generalize effectively to out-of-distribution scenarios without sacrificing the power of modern overparameterized models. The first framework converts an out-of-support zero-shot generalization problem into an out-of-combination problem via a transductive reparameterization, which is possible under low-rank style conditions. We explore how this idea can be applied to domains like robotics, where the environment is changing, and materials and molecular design, where predicting properties of materials or molecules outside of known ranges is crucial to driving more efficient materials discovery. The second framework focuses on few-shot task learning, which involves agents learning new tasks from minimal data and applying them to new environments. We formulate the problem of few-shot task learning as Few-Shot Task Learning through Inverse Generative Modeling, which allows us to leverage the power of neural generative models pretrained on a set of base tasks. We adapt a method for efficient concept learning to few-shot task learning based on our formulation and rapidly learn new tasks with only a few examples, enabling task execution from autonomous driving to real-world robotic manipulation tasks in novel settings without the need for extensive retraining.

Scaling 3D Scene Perception via Probabilistic Programming

2025-11-26T03:02:56Z

Scaling 3D Scene Perception via Probabilistic Programming Gothoskar, Nishad Understanding and interpreting the 3D structure of the world is a central challenge in artificial intelligence. Our physical world is 3D, yet our AI systems often “see” that world through pixels and images. In order to build truly intelligent AI systems, we must go beyond pixels and images and build 3D vision systems that can build meaningful and useful 3D representations of the world. This is the problem of 3D scene perception. How do we transform raw visual input into 3D representations of the world? 3D scene perception has numerous applications from robotics to augmented reality. Despite the advances over the last decade, 3D perception remains a major bottleneck in real-world robotics applications. The challenge stems from the immense variability in real-world conditions, e.g. lighting, color, viewpoint, camera properties, object appearance, the incompleteness of visual data due to limited resolution, noise, and occlusions, and the approximations in our models of visual data. Developing more robust and generalizable 3D perception systems would be an important step towards more general-purpose robotics. In this thesis, we explore a probabilistic architecture for 3D perception based on structured generative models and probabilistic programs. We begin with 3DP3, the first iteration of our approach, which infers 3D scene graphs from real-world depth image data. 3DP3 demonstrates that our method could work on real-world benchmarks and correct commonsense errors from deep learning systems. Building on this foundation, we develop Bayes3D, which scaled up these ideas using a GPU-accelerated image likelihood and generative model alongside a parallel coarse-to-fine inference algorithm. Next, we explore two approaches for incorporating RGB image data into generative 3D graphics programs, expanding their applicability. We then introduce DurableVS, which extends inverse-graphics techniques to model scenes involving a robot and multiple cameras, enabling precise control of a robot. Finally, we present Gen3D, which integrates all the key ideas from this thesis into a real-time 3D perception system that uses multi-resolution probabilistic models of 3D matter to enable real-time tracking that is competitive with vision transformers and 3D Gaussian splatting, state-of-the-art methods in computer vision and computer graphics.

Efficient Generative Models for Visual Synthesis

2025-11-26T03:03:27Z

Efficient Generative Models for Visual Synthesis Yin, Tianwei While current visual generative models produce high-quality outputs, they suffer from significant computational costs and latency, limiting their applicability in interactive settings. In this dissertation, we introduce a suite of techniques designed to enhance the efficiency of generative models for image and video synthesis. First, we propose distribution matching distillation, a method that enables the training of one- or few-step visual generators by distilling knowledge from computationally expensive yet highly capable diffusion models. Next, we develop improved distillation techniques that enhance robustness and scalability, culminating in a production-grade few-step image generator. This system is now deployed in widely used software, generating hundreds of millions of images annually. Finally, we extend our approach to video generation by adopting an autoregressive paradigm, significantly reducing latency and enabling fast interactive video generation and world simulation.

Characterization of pGaN-gate power HEMTs

2025-11-26T03:06:35Z

Characterization of pGaN-gate power HEMTs Yu, Yue This thesis presents a comprehensive study of p-GaN gate GaN High Electron Mobility Transistors (HEMTs) with a focus on understanding how fabrication process variations and gate structural designs impact key electrical performance metrics. Five industry-fabricated wafers, each processed with distinct etch depths, contact strategies, and p-GaN surface configurations, were characterized using a combination of DC and pulsed I–V measurements. Full-transistor modules were evaluated alongside specialized test structures to enable both system-level and localized analysis. DC measurements using the Keysight B1505A system revealed that more aggressive gate contact schemes improved ON-resistance and transconductance, but often at the cost of increased gate leakage and reduced threshold control. Pulsed-IV characterization with the Auriga AU4750 system uncovered dynamic Ron degradation behavior and charge trapping effects, especially under high drain bias conditions. Extracted time constants demonstrated process-dependent trends, with wafers retaining more of the p-GaN surface exhibiting slower charge detrapping and more severe transient effects. Specialized test structures provided additional insights into gate lateral conduction, sheet resistance, and contact asymmetry, reinforcing the connection between device layout, processing, and observed variability. These findings highlight critical trade-offs in the design and fabrication of p-GaN gate GaN HEMTs and offer design-aware strategies for optimizing performance and reliability.

Contactless Sleep and Physiological Monitoring Using Artificial Intelligence and Radio Waves

2025-11-26T03:03:04Z

Contactless Sleep and Physiological Monitoring Using Artificial Intelligence and Radio Waves He, Hao Remote monitoring of sleep and physiological signals is critical for tracking human health, managing diseases, and enabling early intervention. However, existing monitoring solutions face two major limitations: (1) they are often unsuitable for vulnerable populations—such as infants and seniors—and (2) most of them raise concerns about measurement accuracy. We propose a novel, contactless approach that addresses both challenges by combining advances in artificial intelligence (AI) and radio-frequency (RF) sensing. Our solution makes monitoring more comfortable, accessible, and affordable, while still delivering clinically meaningful insights. This thesis makes four fundamental contributions: First, we introduce a system that can extract high-fidelity breathing signals from ambient RF reflections, even in complex scenarios where multiple individuals are present, such as couples sharing a bed. Second, we develop an AI-based sleep monitoring framework that generates sleep hypnograms and detects respiratory events entirely without the need for on-body sensors. Third we develop AI models that infer critical biomarkers—such as blood oxygen saturation (SpO₂) and inflammation (C-reactive protein levels)—in a fully passive and non-intrusive manner. Finally, inspired by the success of large language models, we show that physiological signals can be represented and interpreted analogously to language. This insight enables effective translation between modalities (e.g., from respiration to EEG) and unlocks robust representation learning for downstream clinical tasks. Together, these contributions establish a new paradigm for remote sleep and physiological monitoring—one that is contactless, continuous, and passive. We validate our system on real world datasets and demonstrate its potential to fundamentally transform clinical care and home health monitoring.

The legacy of Rudolf Nieuwenhuys in perspective

2025-11-26T03:10:59Z

The legacy of Rudolf Nieuwenhuys in perspective Pignatelli, Michele; Rockland, Kathleen S. Professor Nieuwenhuys is among the great neuroanatomists and a historical figure of the later 20th and early 21st centuries. His legacy is manifold. There is the tangible legacy of the multiple scientific volumes, at once physical and conceptual entities. There is the generational legacy of handed-on scientific and intellectual traditions, and there is the legacy of specific scientific directions. In this brief Commentary, we highlight just two examples of his scientific contributions.

Some Advice on Sustainability: ‘I Would Never Get into a Business I Did Not Really Understand’

2025-11-26T03:10:53Z

Some Advice on Sustainability: ‘I Would Never Get into a Business I Did Not Really Understand’ Wright, Randall S. My father, Chester S. Wright, was a business executive. He was president of two manufacturing companies and a member of the board of directors of four others. As a young boy, I remember him coming home from work in a big, black Chrysler Imperial—a “company car”—fitted out with shining chromium bumpers and gleaming radiator grill. After a wonderful home-cooked dinner my mother always made for my father, my two sisters, and me, he and I would head out in the Imperial to Gray’s Drug Store so he could buy House of Windsor cigars, and we could pick up the latest copies of Popular Mechanics, Popular Science, and Mechanix Illustrated.

The 5-methylcytosine DNA glycosylase ROS1 prevents paternal genome hypermethylation in Arabidopsis endosperm

2025-11-26T03:10:22Z

The 5-methylcytosine DNA glycosylase ROS1 prevents paternal genome hypermethylation in Arabidopsis endosperm Hemenway, Elizabeth A.; Gehring, Mary Background DNA methylation patterning is a consequence of opposing activities of DNA methyltransferases and DNA demethylases. In many plant and animal species, reproduction is a period of significant epigenome lability. In flowering plants, two distinct female gametes, the egg cell and the central cell, are fertilized to produce the embryo and the endosperm of the seed. The endosperm is an unusual tissue, exemplified by triploidy and reduced DNA methylation. In Arabidopsis thaliana, a 5-methylcytosine DNA glycosylase, DME, demethylates regions of the central cell genome, leading to methylation differences between maternally- and paternally-inherited endosperm genomes after fertilization. Expression of DME in the central cell is required for gene imprinting, or parent-of-origin specific gene expression, in endosperm. DME is part of a four member gene family in Arabidopsis that includes ROS1, DML2, and DML3. It is unknown whether any of the other DNA glycosylases are required for endosperm methylation patterning. Results Using whole-genome methylation profiling, we identify ROS1 target regions in the endosperm. We show that ROS1 prevents hypermethylation of paternally-inherited alleles in the endosperm at regions that lack maternal or paternal allele methylation in wild-type endosperm. Additionally, we demonstrate that at many ROS1 target regions the maternal alleles are demethylated by DME. Conclusions ROS1 promotes epigenetic symmetry between parental genomes in the endosperm by preventing CG methylation gain on the paternal genome. We conclude that ROS1 and DME act in a parent-of-origin-specific manner at shared endosperm targets, and consider possible implications for the evolution of imprinting mechanisms.

Players chatter and dice clatter: exploring sonic power relations in posthuman game-based learning ecologies

2025-11-26T03:10:57Z

Players chatter and dice clatter: exploring sonic power relations in posthuman game-based learning ecologies Woods, Peter J; Jones, Karis Responding to both recent interest in sound within qualitative education research and sound studies literature that conceptualizes sound as a posthuman technology, we use this paper to explore the following research questions: How does sound both enact and unveil posthuman learning ecologies? And how can education scholars engage sound within posthuman research? Through a posthuman framework, we position noise as an analytical tool for exploring and unveiling more-than-human relations. We then draw parallels between posthuman qualitative research into sound (via noise) and the ideological foundation of experimental music, a musical tradition deeply invested in working with sound as an agentic actor. Within this alignment, we propose using graphic scores to transcribe sonic data without reinscribing humanist research aims. To illustrate, we provide a micro-analysis of preservice teachers engaged in a role-playing game activity and uncover the ways sound asserts its agency within learning ecologies.

Palonosetron, a 5-HT3 Receptor Antagonist, Induces G1 Cell Cycle Arrest and Autophagy in Gastric Cancer Cells

2025-11-26T03:10:27Z

Palonosetron, a 5-HT3 Receptor Antagonist, Induces G1 Cell Cycle Arrest and Autophagy in Gastric Cancer Cells Yoo, Young Chul; Lin, Lin; Lee, Sihak; Shin, Yeeun Rachel; Oh, Ju Eun; Kim, Na Young Serotonin or 5-hydroxytryptamine (5-HT) has been implicated in promoting cancer cell growth by acting on 5-HT receptors, such as 5-HT1 and 5-HT2 receptors. However, the role of 5-HT3 receptor antagonists in gastric cancer cell lines remains unclear. This study aimed to evaluate the effect of 5-HT3 receptor antagonists (ondansetron, palonosetron, and ramosetron) on cancer cell growth using AGS and MKN-1 cell lines, as well as the xenograft mouse model. All the three antagonists inhibited cell proliferation, migration, and colony formation in AGS cells. Specifically, palonosetron induced G1 cell cycle arrest, autophagy, and phosphorylation of GSK3β, along with increased expression of p27, p53, and LC3B. In vivo studies demonstrated that palonosetron reduced tumor growth and modulated pro-inflammatory cytokines—tumor necrosis factor alpha, interleukin 6, and interleukin 1β. These findings suggest that 5-HT3 receptor antagonists, especially palonosetron, exert anti-tumor effects in gastric cancer through G1 cell cycle regulation and immunomodulation. The results position palonosetron as a promising lead for further preclinical development in gastric cancer.

50 years of nanomechanics: Scale-bridging mechanistic insights through the looking glass

2025-11-26T03:10:47Z

50 years of nanomechanics: Scale-bridging mechanistic insights through the looking glass Han, Seung M.; Gianola, Daniel S.; Portela, Carlos M.; Sebastiani, Marco; Kirchlechner, Christoph Historical and recent advances in the field of nanomechanics, ranging from the early development of nanoindentation to recent advances in artificial intelligence- and machine learning-based characterization and modeling are covered in this article. Early advances were motivated by thin-film mechanics challenges driven by the microelectronics industry. In the ensuing years, different methodologies for probing mechanical properties at length scales relevant to a myriad of applications and materials systems have been developed, coupled with a variety of in situ testing methods that shed insights into new mechanisms. Built upon the knowledge base from nanomechanics, new mechanical metamaterials with otherwise unachievable material properties have been discovered, and new methods in testing and analyzing properties for extreme conditions have been recently reported. This article discusses the journey that the nanomechanics community has gone through over the past 50 years and shares the scale-bridging mechanistic insights through the looking glass.

Sustainable Synthesis of CoFe2O4/Fe2O3 Catalyst for Hydrogen Generation from Sodium Borohydride Hydrolysis

2025-11-26T03:10:55Z

Sustainable Synthesis of CoFe2O4/Fe2O3 Catalyst for Hydrogen Generation from Sodium Borohydride Hydrolysis Teixeira, Lucas Tonetti; Medeiros, Marcos; Liu, Liying; Park, Vinicius Novaes; Valente-Rodriguez, Célio; Letichevsky, Sonia; Fajardo, Humberto Vieira; de Siqueira, Rogério Navarro Correia; Maia da Costa, Marcelo Eduardo Huguenin; Botelho Junior, Amilton Barbosa Hydrogen has been explored as a greener alternative for greenhouse gas emissions reduction. Sodium borohydride (NaBH4) is a favorable hydrogen carrier due to its high hydrogen content, safe handling, and rapid hydrogen release. This work presents a novel synthesis of the catalyst CoFe2O4/Fe2O3 using nanocellulose fibers (TCNF) as reactive templates for metal adsorption and subsequent calcination. The resulting material was tested for H2 production from basic NaBH4 aqueous solutions (10–55 °C). The catalyst’s composition is 74.8 wt% CoFe2O4, 25 wt% Fe2O3, and 0.2 wt% Fe2(SO4)3 with agglomerated spheroidal particles (15–20 nm) and homogeneous Fe and Co distribution. The catalyst produced 1785 mL of H2 in 15 min at 25 °C (50 mg catalyst, 4.0% NaBH4, and 2.5 wt% NaOH), close to the stoichiometric maximum (2086 mL). The maximum H2 generation rate (HGR) reached 3.55 L min−1 gcat−1 at 40 °C. Activation energies were determined using empirical (38.4 ± 5.3 kJ mol−1) and Langmuir–Hinshelwood (L–H) models (42.2 ± 5.8 kJ mol−1), consistent with values for other Co-ferrite catalysts. Kinetic data fitted better to the L–H model, suggesting that boron complex adsorption precedes H2 evolution.

On the potential of microtubules for scalable quantum computation

2025-11-26T03:10:41Z

On the potential of microtubules for scalable quantum computation Mavromatos, Nick E.; Mershin, Andreas; Nanopoulos, Dimitri V. We examine the quantum coherence properties of tubulin heterodimers arranged into the protofilaments of cytoskeletal microtubules. In the physical model proposed by the authors, the microtubule interiors are treated as high-Q quantum electrodynamics (QED) cavities that can support decoherence-resistant entangled states under physiological conditions, with decoherence times of the order of O ( 10 - 6 ) s. We identify strong electric dipole interactions between tubulin dimers and ordered water dipole quanta within the microtuble interior as the mechanism responsible for the extended coherence times. Classical nonlinear (pseudospin) σ -models describing solitonic excitations are reinterpreted as emergent quantum-coherent—or possibly pointer—states, arising from incomplete collapse of dipole-aligned quantum states. These solitons mediate dissipation-free energy transfer along microtubule filaments. We discuss logic-gate-like behaviour facilitated by microtubule-associated proteins, and outline how such structures may enable scalable, ambient-temperature quantum computation, with the fundamental unit of information storage realized as a quDit encoded in the tubulin dipole state. We further describe a process akin to “decision-making” that emerges following an external stimulus, whereby optimal, energy-loss-free signal and information transport pathways are selected across the microtubular network. Finally, we propose experimental approaches—including Rabi-splitting spectroscopy and entangled surface plasmon probes—to validate the use of biomatter as a substrate for scalable quantum computation.

Can Artificial Intelligence Improve the Appropriate Use and Decrease the Misuse of REBOA?

2025-11-26T03:10:54Z

Can Artificial Intelligence Improve the Appropriate Use and Decrease the Misuse of REBOA? Bokenkamp, Mary; Ma, Yu; Dorken-Gallastegi, Ander; Proaño-Zamudio, Jefferson A; Gebran, Anthony; Velmahos, George C; Bertsimas, Dimitris; Kaafarani, Haytham MA Background: The use of resuscitative endovascular balloon occlusion of the aorta (REBOA) for control of noncompressible torso hemorrhage remains controversial. We aimed to utilize a novel and transparent/interpretable artificial intelligence (AI) method called Optimal Policy Trees (OPTs) to improve the appropriate use and decrease the misuse of REBOA in hemodynamically unstable blunt trauma patients. Methods: We trained and then validated OPTs that “prescribe” REBOA in a 50:50 split on all hemorrhagic shock blunt trauma patients in the 2010–2019 ACS-TQIP database based on rates of survival. Hemorrhagic shock was defined as a systolic blood pressure ≤90 on arrival or a transfusion requirement of ≥4 units of blood in the first 4 h of presentation. The expected 24 h mortality rate following OPT prescription was compared to the observed 24 h mortality rate in patients who were or were not treated with REBOA. Results: Out of 4.5 million patients, 100,615 were included, and 803 underwent REBOA. REBOA patients had a higher rate of pelvic fracture, femur fracture, hemothorax, pneumothorax, and thoracic aorta injury (p < 0.001). The 24 h mortality rate for the REBOA vs. non-REBOA group was 47% vs. 21%, respectively (p < 0.001). OPTs resulted in an 18% reduction in 24 h mortality for REBOA and a 0.8% reduction in non-REBOA patients. We specifically divert the misuse of REBOA by recommending against REBOA in cases where it leads to worse outcomes. Conclusions: This proof-of-concept study shows that interpretable AI models can improve mortality in unstable blunt trauma patients by optimizing the use and decreasing the misuse of REBOA. To date, these models have been used to predict outcomes, but their groundbreaking use will be in prescribing interventions and changing outcomes.

Ultra-High Resolution 9.4T Brain MRI Segmentation via a Newly Engineered Multi-Scale Residual Nested U-Net with Gated Attention

2025-11-26T03:10:29Z

Ultra-High Resolution 9.4T Brain MRI Segmentation via a Newly Engineered Multi-Scale Residual Nested U-Net with Gated Attention Kalluvila, Aryan; Patel, Jay B.; Johnson, Jason M. A 9.4T brain MRI is the highest resolution MRI scanner in the public market. It offers submillimeter brain imaging with exceptional anatomical detail, making it one of the most powerful tools for detecting subtle structural changes associated with neurological conditions. Current segmentation models are optimized for lower-field MRI (1.5T–3T), and they struggle to perform well on 9.4T data. In this study, we present the GA-MS-UNet++, the world’s first deep learning-based model specifically designed for 9.4T brain MRI segmentation. Our model integrates multi-scale residual blocks, gated skip connections, and spatial channel attention mechanisms to improve both local and global feature extraction. The model was trained and evaluated on 12 patients in the UltraCortex 9.4T dataset and benchmarked against four leading segmentation models (Attention U-Net, Nested U-Net, VDSR, and R2UNet). The GA-MS-UNet++ achieved a state-of-the-art performance across both evaluation sets. When tested against manual, radiologist-reviewed ground truth masks, the model achieved a Dice score of 0.93. On a separate test set using SynthSeg-generated masks as the ground truth, the Dice score was 0.89. Across both evaluations, the model achieved an overall accuracy of 97.29%, precision of 90.02%, and recall of 94.00%. Statistical validation using the Wilcoxon signed-rank test (p < 1 × 10−5) and Kruskal–Wallis test (H = 26,281.98, p < 1 × 10−5) confirmed the significance of these results. Qualitative comparisons also showed a near-exact alignment with ground truth masks, particularly in areas such as the ventricles and gray–white matter interfaces. Volumetric validation further demonstrated a high correlation (R2 = 0.90) between the predicted and ground truth brain volumes. Despite the limited annotated data, the GA-MS-UNet++ maintained a strong performance and has the potential for clinical use. This algorithm represents the first publicly available segmentation model for 9.4T imaging, providing a powerful tool for high-resolution brain segmentation and driving progress in automated neuroimaging analysis.

De novo design of a two-step approach targeting Claudin-6 for enhanced drug delivery to solid tumors

2025-11-26T03:10:50Z

De novo design of a two-step approach targeting Claudin-6 for enhanced drug delivery to solid tumors Yan, Jiayao; Zhong, Liqing; Chen, Xiaotong; Li, Lin; Liu, Fangcen; Lei, Lei; An, Mengchao; Wei, Xiao; Wang, Ying; Chen, Tianran; Guo, Jingyi; Shao, Jie; Yu, Xiaoxiao; Zhao, Yingjie; Li, Rutian; Liu, Qin Background Although antibody-conjugated drugs have achieved success in clinical practice for cancer treatment, challenges remain in developing a highly efficient drug delivery system with specific accumulation in tumors and reduction in side effects. With improved pharmacokinetics, strong covalent bonding and quick binding reactions, a pre-targeting approach via molecular pairs represents an attractive platform for two-step delivery system construction. Methods Bioinformatics and immunohistochemistry assays were performed to assess Claudin-6 (CLDN6) as a highly specific tumor target in solid tumors. A phage-displayed library was used to screen and optimize anti-CLDN6 designed ankyrin repeat proteins (DARPins), which were incorporated into a two-step delivery system based on SpyTag/SpyCatcher. Fluorescent staining, flow cytometry and near-infrared imaging were performed to assess the tumor-targeting ability and biodistribution of this delivery system. The cytotoxic drug, Monomethyl auristatin E (MMAE), was conjugated with the delivery system to evaluate its anti-tumor efficacy and safety profile. Results Anti-CLDN6 DARPins exhibited specific binding to CLDN6+ cancer cells with high affinity instead of negative cells in vitro, ex vivo and in vivo. The DARPins-based two-step delivery system improved background clearance with a high signal-to-noise ratio, enhancing the specific accumulation of payloads in tumors. The cytotoxic drug delivered via the two-step system appeared superior to the one-step approach in IC50, biodistribution, and tumor growth inhibition. Conclusions Our study presented the de novo design of a two-step drug delivery system targeting Claudin-6 with enhanced anti-tumor efficacy and improved biosafety. These findings highlighted the potential of this approach to enhance the efficacy of tumor-targeting therapies and reduce adverse effects, paving the way for more effective cancer treatments.

How Does AI Transform Cyber Risk Management?

2025-11-26T03:10:51Z

How Does AI Transform Cyber Risk Management? Zeijlemaker, Sander; Lemiesa, Yaphet K; Schröer, Saskia Laura; Abhishta, Abhishta; Siegel, Michael Digital transformation embeds smart cities, e-health, and Industry 4.0 into critical infrastructures, thereby increasing reliance on digital systems and exposure to cyber threats and boosting complexity and dependency. Research involving over 200 executives reveals that under rising complexity, only 15% of cyber risk investments are effective, leaving most organizations misaligned or vulnerable. In this context, the role of artificial intelligence (AI) in cybersecurity requires systemic scrutiny. This study analyzes how AI reshapes systemic structures in cyber risk management through a multi-method approach: literature review, expert workshops with practitioners and policymakers, and a structured kill chain analysis of the Colonial Pipeline attack. The findings reveal three new feedback loops: (1) deceptive defense structures that misdirect adversaries while protecting assets, (2) two-step success-to-success attacks that disable defenses before targeting infrastructure, and (3) autonomous proliferation when AI applications go rogue. These dynamics shift cyber risk from linear patterns to adaptive, compounding interactions. The principal conclusion is that AI both amplifies and mitigates systemic risk. The core recommendation is to institutionalize deception in security standards and address drifting AI-powered systems. Deliverables include validated systemic structures, policy options, and a foundation for creating future simulation models to support strategic cyber risk management investment.

Flex-Route Transit for Smart Cities: A Reinforcement Learning Approach to Balance Ridership and Performance

2025-11-26T03:10:31Z

Flex-Route Transit for Smart Cities: A Reinforcement Learning Approach to Balance Ridership and Performance Rodriguez, Joseph; Koutsopoulos, Haris N.; Zhao, Jinhua A major challenge for modern transit systems relying on traditional fixed-route designs is providing broad accessibility to users. Flex-route transit can enhance accessibility in low-density areas, since it combines the directness of fixed-route transit with the coverage of on-demand mobility. Although deviating for optional pickups can increase ridership and transit accessibility, it also deteriorates the service performance for fixed-route riders. To balance this inherent trade-off, this paper proposes a reinforcement learning approach for deviation decisions. The proposed model is used in a case study of a proposed flex-route service in the city of Boston. The performance on competing objectives is evaluated for reward configurations that adapt to peak and off-peak scenarios. The analysis shows a significant improvement of our method compared to a heuristic derived from industry practice as a baseline. To evaluate robustness, we assess performance across scenarios with varying demand compositions (fixed and requested riders). The results show that the method achieves greater improvements than the baseline in scenarios with increased request ridership, i.e., where decision-making is more complex. Our approach improves service performance under dynamic demand conditions and varying priorities, offering a valuable tool for smart cities to operate flex-route services.

Search for a cH signal in the associated production of at least one charm quark with a Higgs boson in the diphoton decay channel in pp collisions at $$\sqrt{s}=13$$ TeV

2025-11-26T03:10:44Z

Search for a cH signal in the associated production of at least one charm quark with a Higgs boson in the diphoton decay channel in pp collisions at $$\sqrt{s}=13$$ TeV Chekhovsky, V.; Hayrapetyan, A.; Makarenko, V.; Tumasyan, A.; Adam, W.; Andrejkovic, J. W.; Benato, L.; Bergauer, T.; Chatterjee, S.; Damanakis, K.; Dragicevic, M.; Hussain, P. S.; Jeitler, M.; Krammer, N.; Li, A.; Liko, D.; Mikulec, I. This paper presents the first search for a cH signal sensitive to the coupling of the charm quark (c) to the Higgs boson (H) in the associated production of at least one charm quark with a Higgs boson decaying to two photons. The results are based on a data set of proton-proton collisions at a center-of-mass energy of 13 TeV collected with the CMS experiment at the LHC, corresponding to an integrated luminosity of 138 fb−1. Assuming the standard model (SM) rates for all other Higgs boson production processes, the observed (expected) upper limit at 95% confidence level on the cH signal strength is 243 (355) times the SM prediction. Under the same assumption, the observed (expected) allowed interval on the Higgs boson to charm quark coupling modifier, κc, is |κc| < 38.1 (|κc| < 72.5) at 95% confidence level.

A Bunch of Gaps: Factors Behind Service Reliability in Chicago’s High-Frequency Transit Network

2025-11-26T03:10:32Z

A Bunch of Gaps: Factors Behind Service Reliability in Chicago’s High-Frequency Transit Network Rodriguez, Joseph; Koutsopoulos, Haris N.; Zhao, Jinhua Frequent transit services in urban areas have the potential to increase their accessibility to transit-dependent riders and reduce congestion by attracting new ridership through a modal shift. However, bus services operating in mixed traffic face operational challenges that reduce reliability and hinder their attractiveness. The sources of unreliability can range from local-level conditions, like the road infrastructure, to higher-level decisions, like the service plan. For the effective planning of improvement strategies, both scales of analysis must be considered. This paper uses a novel modeling framework to understand reliability by analyzing the route and segment factors separately. The Chicago Transit Authority (CTA) bus network is used as a case study for the analysis. The data reflect the operational, demand, and urban conditions of 50 high-frequency bus routes. At the route level, we use the coefficient of headway variation as the dependent variable and diverse route characteristics as explanatory variables. The results indicate that the most significant contributors to the variability of headways are variability in schedules and dispatching at terminals. It is also found that driver experience impacts reliability and that east–west routes are more unreliable than north–south routes. At the segment level, we use data from trips involved in bunching and gaps. As the dependent variable, a novel measure is formulated to capture how quickly bunching or gaps are formed. The bunching and gap events are treated as separate regression models. Findings suggest that link and dwell time variability are the most significant contributors to gap and bunching formation. In terms of infrastructure, bus lane segments reduce gap formations, and left turns increase bunching and gap formations. The insights presented can inform improvements in service and transit infrastructure planning to improve transit level of service (LOS) and support the future of sustainable, smart cities.

Oil Transport Simulation and Oil Consumption Prediction with a Physics-Based and Data-Driven Digital Twin Model for Internal Combustion Engines

2025-11-26T03:10:58Z

Oil Transport Simulation and Oil Consumption Prediction with a Physics-Based and Data-Driven Digital Twin Model for Internal Combustion Engines Zhong, Xinlin; Tian, Tian Lubrication oil consumption (LOC) is one of the major sources of emissions from internal combustion (IC) engines; yet, analyzing and predicting it through modeling is challenging due to its multi-physics nature, which spans different time and length scales. In this work, a digital twin model is developed to simulate oil transport in the piston ring pack of IC engines and predict the resulting oil consumption with all major physical mechanisms considered. Three main contributors to LOC, namely, top ring up-scraping, oil vaporization on the liner, and reverse gas flows through the top ring gap, are included in the model. It was found that their behaviors are heavily dependent on the arrangement of the piston ring gaps. Therefore, with the ring rotation behavior still not resolved, the current model can predict the LOC range of a given engine profile. Results show that the predicted range can well encapsulate the experimentally measured LOC value.

LSM and CPT

2025-11-26T03:10:45Z

LSM and CPT Seiberg, Nathan; Shao, Shu-Heng; Zhang, Wucheng We study a number of 1+1d lattice models with anti-unitary symmetries that simultaneously reflect space and reverse time. Some of these symmetries are anomalous, leading to Lieb-Schultz-Mattis-type constraints, thus excluding a trivially gapped phase. Examples include a mod 8 anomaly in the Majorana chain and various mod 2 anomalies in the spin chain. In some cases, there is an exact, non-anomalous lattice symmetry that flows in the continuum to CPT. In some other cases, the CPT symmetry of the continuum theory is emergent or absent. Depending on the model, the anomaly of the lattice model is matched in the continuum in different ways. In particular, it can be mapped to an emergent anomaly of an emanant symmetry.

Crystallization of Glauber's salt

2025-11-25T06:32:25Z

Crystallization of Glauber's salt Coberly, C. Wheeler. Thesis: M.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1936; Includes bibliographical references (leaf 39).

Investigation of torquemeters for high speed shafts

2025-11-25T06:33:34Z

Investigation of torquemeters for high speed shafts Saluja, Narinder S. (Narinder Singh) Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1959; Includes bibliographical references (leaves 64-67).

The resonant-frequency shift of a microwave cavity caused by the high-density plasma in semiconductors, as a function of magnetic field

2025-11-25T03:04:01Z

The resonant-frequency shift of a microwave cavity caused by the high-density plasma in semiconductors, as a function of magnetic field Weber, Robert. Thesis: Sc. D., Massachusetts Institute of Technology, Department of Physics, 1959; Includes bibliographical references (leaves 46-47).

Analysis of angular scintillation of radar echoes

2025-11-25T06:32:44Z

Analysis of angular scintillation of radar echoes Graham, James William. Thesis: M.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1952

Rapid transit use of existing rail lines

2025-11-25T06:33:31Z

Rapid transit use of existing rail lines Kenyon, Michael D. Thesis: B.S., Massachusetts Institute of Technology, Department of Civil Engineering, 1958; Includes bibliographical references (leaf 25).

Mass transfer from rotating cylinders

2025-11-25T06:33:29Z

Mass transfer from rotating cylinders Cotter, John.; Schmidt, Guy L. Thesis: B.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1956; Bibliography: leaf 38.

An observation about the Chicago Council and its policies

2025-11-25T06:33:26Z

An observation about the Chicago Council and its policies Naber, Fred P. Thesis: B.S., Massachusetts Institute of Technology, Department of Business and Engineering Administration, 1948

The design and construction of an ultra-high vacuum field-ion microscope.

2025-11-25T06:32:36Z

The design and construction of an ultra-high vacuum field-ion microscope. Olson, Gregory Bruce. Thesis: M.S., Massachusetts Institute of Technology, Department of Metallurgy and Materials Science, 1970; Bibliography: leaf 35.

South End Center for the Arts.

2025-11-25T06:33:23Z

South End Center for the Arts. Dunbar, Gary Arthur. Thesis: B. Arch., Massachusetts Institute of Technology, Department of Architecture, 1965; "Special requirements for group A occupancy: theatres" leaves [34-42] inserted. "Special requirements for group C occupancy: schools" leaves [50-54] inserted.; Bibliography: leaf 20.

Analysis of braced excavations.

2025-11-25T03:04:06Z

Analysis of braced excavations. Wong, Ing Hieng. Thesis: Sc. D., Massachusetts Institute of Technology, Department of Civil Engineering, 1971; Three leaves on transparent sheets. Vita.; Bibliography: leaves 95-99.

Shipleasing as a prospective method of l/t financing for international shipowners.

2025-11-25T06:32:28Z

Shipleasing as a prospective method of l/t financing for international shipowners. Angelicoussis, John Anthony. Thesis: M.S., Massachusetts Institute of Technology, Sloan School of Management, 1974; Includes bibliographical references.

Modelling rail freight management.

2025-11-25T03:03:57Z

Modelling rail freight management. Assad, A. (Arjang) Thesis: Ph. D., Massachusetts Institute of Technology, Sloan School of Management, 1978; Vita.; Bibliography: leaves 277-292.

The effects of rapid thermal annealing on gallium arsenide grown by MOCVD on silicon substrates

2025-11-25T06:32:40Z

The effects of rapid thermal annealing on gallium arsenide grown by MOCVD on silicon substrates Lehman, LeNore Louise. Thesis: M.S., Massachusetts Institute of Technology, Department of Materials Science and Engineering, 1988; Includes bibliographical references.

Acoustic-phonetic and lexical constraints in word recognition: lexical access using partial information

2025-11-25T06:32:20Z

Acoustic-phonetic and lexical constraints in word recognition: lexical access using partial information Huttenlocher, Daniel P. Thesis: M.S., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 1984; Bibliography: leaves 73-77.

Metabolism in vivo of 1, 3-butanediol in the rat

2025-11-25T03:03:27Z

Metabolism in vivo of 1, 3-butanediol in the rat Nahapetian, Aratoonnaz, author. The metabolism of 1, 3-butanediol (BD) was investigated in vitamin B 12 -deficient and normal rats and in liver slice and diaphragm systems. Body weight gain and feed efficiency were determined in rats fed ad libitum for five weeks on a basal 5% BD or 5% sodium propionate diet with and without vitamin B12. The rats were train-fed for ten months on the same diets. The presence of sodium prop i onate in vitamin B12-deficient basal diets resulted in reduced food intake while BD had the opposite effect. As a result, vitamin B12-deficient rats fed a 5% sodium propionate diet grew less than those fed a 5% BD diet. The metabolism in vivo of BD labeled in carbon-1 (BD-l-cl4) and carbon-4 (BD-4-cl4) were compared to the metabolism of propionate-l-cl4 (PRP-l-cl4) in vitamin B12-deficient and normal rats. Vitamin B12 deficiency reduced the oxidation of sodium propionate but not that of BD, and had no effect on glycogen labeling from BD-l-cl4 and BD-4-cl4. For PRP-l-cl4 however, vitamin B12 deficiency resulted in not only no incorporation of label but liver glycogen levels were very small. On the other hand , when vitamin B12 was present in the diet, the labeling of glycogen from propionate was higher than that from either of the BD-labeled test compounds. Methylmalonic aciduria and urinary loss of ingested activity was higher in vitamin B12-deficient rats fed PRP-l-cl4 than in those fed l abe l ed BD. Nearly all of the urinary activity of vitamin B 1 2-deficient rats fed PRP-l-cl4 was in the form of me t hy l malonic acid (MMA), while little, if any, of the activity was found in the MMA fraction of urine of vita m in B12-deficient rats fed labeled BD. The metabolism in vivo of BD-c14 and BD-3-c14 was investigated in normal rats. About eighty percent of BD was oxidized to carbon dioxide within 32 hours. Its oxidation in the first eight hours was higher when BD was administered intraperitoneally than when it was fed by stomach tube. The loss of ingested activity in the urine expressed as a percentage of total intake and 1,3-BD was higher at the higher doses of BD. However, the activity in urinary BD could not account for all the activity in the urine. A considerable amount of ketone bodies was detected in urine of rats after feeding BD while no detectable ketone bodies were found in the urine of control rats. In addition, relative specific activities of urinary BD and S -hydroxybutyrate were 0.91 and 0.50 respectively. Polarimetry of both purified urinary BD and S -hydroxybutyrate showed that the percentages of (+)- and (-)-isomers of both compounds were 40 and 60% respectively. The metabolism in vitro of BD-3-c14 and DL-S - hydroxybutyrate-4-cl4 were investigated in systems which contained liver slices alone, diaphragm alone or both liver slices plus diaphragm. The oxidation rate of S -hydroxybut y rate was lower in liver slices than in either the diaphragm or the liver slices plus diaphragm systems. Moreover, the rate of oxidation of S -hydrox y- butyrate was highest in the system which included both liver slices and diaphragm. On the other hand, the oxidation rate of BD was lower in the system which had only diaphragm than in the other two systems. However, the rate of BD oxidation was highest in the system - wh ich included both liver slices and diaphragm. Presence of BD gave rise to increased D-(-)- S -hydroxybutyrate and acetoacetate in systems which contained liver slices or liver slices plus diaphragm. In addition, the production rate of D-(-)- S - hydroxybutyric acid was higher than that of acetoacetate in the pre sence of BD, while the opposite was true in its absence. Finally, all the radioactivity in the control incubation media was accounted for by BD-3-cl4, while about 1.5 and 98.5 percent of incubation media activity were recovered in S -hydroxybutyrate and BD peaks, respectively, in incub at ion systems containing liver. The results of this study indicate that 1,3-BD and sodium propionate do not share a common metabolic pathway in the rat. The data suggest, however, that nahapetian-4-1, 3-BD is most probably oxidized to S-hydroxybutyric acid using a "1,3-butanediol dehydrogenase" that is higher in activity in the liver than in the diaphragm. M oreover the (+)-isomer of BD is oxidized at a faster rate than the (-)-isomer, suggesting that the two isomers are oxidized by two different pathways. Thesis: Sc. D., Massachusetts Institute of Technology, Department of Nutrition and Food Science, 1971; Thesis supervised by Sanford A. Miller Vita: page 196; Includes bibliographical references (pages 182-196)

An investigation of the merits of a four-element coplanar vacuum tube when used as a modulator at carrier telephone frequencies

2025-11-25T06:32:33Z

An investigation of the merits of a four-element coplanar vacuum tube when used as a modulator at carrier telephone frequencies Perkins, Edwin H. Thesis: M.S., Massachusetts Institute of Technology, Department of Electrical Engineering, 1930; Includes bibliographical references (leaf 115).

Micro-analysis of grinding machine cuttings

2025-11-25T06:33:21Z

Micro-analysis of grinding machine cuttings Zurlo, J. V.; Terkelsen, E. A. Thesis: B.S., Massachusetts Institute of Technology, Department of Mechanical Engineering, 1922

Tests upon bamboo as a concrete reinforcement and a consideration of its application in construction

2025-11-25T06:33:09Z

Tests upon bamboo as a concrete reinforcement and a consideration of its application in construction Young, Joe W.; Guo, Dianbang. Thesis: B.S., Massachusetts Institute of Technology, Department of Civil Engineering, 1924

A precision method for the determination of dew points of complex gaseous systems

2025-11-25T06:32:38Z

A precision method for the determination of dew points of complex gaseous systems Cox, John Tatum. Thesis: M.S., Massachusetts Institute of Technology, Department of Chemical Engineering, 1936; Includes bibliographical references (leaf 43).

A biomimetic chip to assess subcutaneous bioavailability of monoclonal antibodies in humans

2025-11-25T06:37:31Z

A biomimetic chip to assess subcutaneous bioavailability of monoclonal antibodies in humans Chandran Suja, Vineeth; Qi, Qin M; Halloran, Kevin; Zhang, Jifeng; Shaha, Suyog; Prakash, Supriya; Kumbhojkar, Ninad; Deslandes, Antoine; Huille, Sylvain; Gokarn, Yatin R; Mitragotri, Samir Subcutaneous (subQ) injection is a common route for delivering biotherapeutics, wherein pharmacokinetics is largely influenced by drug transport in a complex subQ tissue microenvironment. The selection of good drug candidates with beneficial pharmacokinetics for subQ injections is currently limited by a lack of reliable testing models. To address this limitation, we report here a Subcutaneous Co-Culture Tissue-on-a-chip for Injection Simulation (SubCuTIS). SubCuTIS possesses a 3D coculture tissue architecture, and it allows facile quantitative determination of relevant scale independent drug transport rate constants. SubCuTIS captures key in vivo physiological characteristics of the subQ tissues, and it differentiates the transport behavior of various chemically distinct molecules. We supplemented the transport measurements with theoretical modeling, which identified subtle differences in the local absorption rate constants of seven clinically available mAbs. Accounting for first-order proteolytic catabolism, we established a mathematical framework to assess clinical bioavailability using the local absorption rate constants obtained from SubCuTIS. Taken together, the technology described here broadens the applicability of organs-on-chips as a standardized and easy-to-use device for quantitative analysis of subQ drug transport.

Nanoparticle-induced lipid membrane deformation influences the design of biomedicine

2025-11-25T06:37:41Z

Nanoparticle-induced lipid membrane deformation influences the design of biomedicine Pincus, Isaac; Qi, Qin M Controlling the physicochemical properties of nanoparticles is important for their performance as drug carriers, pharmaceuticals, or imaging contrast agents in nanomedicine. Predictive models can accelerate experimental designs at reduced time and costs compared to a brute-force approach conventionally used. However, physical principles underlying particle-cell interactions are still poorly understood due to their large size contrast, hindering the model development. In this work, we describe a model that examines the interaction between multiple particles and the membrane of a mammalian cell or an artificial vesicle, thus influencing the outcomes of surface adsorption, detachment or uptake of particles. Compared to existing biophysical models on particle-membrane interactions accounting for membrane adhesion, stretching and bending energies, we make several important updates that are essential to reaching quantitative agreement with existing experimental data. Particle-induced membrane tension changes are crucial to the membrane deformation even at very low surface concentrations (0.1%); we explain this surprising finding using a new length scale previously neglected. Furthermore, a multi-step and non-equilibrium endocytosis mechanism is proposed in the absence of specific receptor-ligand interactions, inspired by recent experimental evidence on the dynamic regulation of membrane tension through the active transport of lipid molecules. We demonstrate the predictive power of our model in generating the adsorption isotherms and shear-induced particle detachment from cell surfaces and the size-dependent rate of particle uptake. Our research provides a framework to design tailor-made nanoparticles with controllable interaction outcomes with various cell types based on a quantitative and fundamental understanding.

Effective use of biosensors for high-throughput library screening for metabolite production

2025-11-25T06:37:39Z

Effective use of biosensors for high-throughput library screening for metabolite production Kaczmarek, Jennifer A; Prather, Kristala LJ The development of fast and affordable microbial production from recombinant pathways is a challenging endeavor, with targeted improvements difficult to predict due to the complex nature of living systems. To address the limitations in biosynthetic pathways, much work has been done to generate large libraries of various genetic parts (promoters, RBSs, enzymes, etc.) to discover library members that bring about significantly improved levels of metabolite production. To evaluate these large libraries, high throughput approaches are necessary, such as those that rely on biosensors. There are various modes of operation to apply biosensors to library screens that are available at different scales of throughput. The effectiveness of each biosensor-based method is dependent on the pathway or strain to which it is applied, and all approaches have strengths and weaknesses to be carefully considered for any high throughput library screen. In this review, we discuss the various approaches used in biosensor screening for improved metabolite production, focusing on transcription factor-based biosensors.

Lincoln Laboratory and MIT Haystack Observatory partner to unveil hidden parts of the galaxy

2025-11-25T06:39:00Z

Lincoln Laboratory and MIT Haystack Observatory partner to unveil hidden parts of the galaxy Parde, Nathan They propose building a telescope made of thousands of tiny, identical satellites that will work together to reveal low-frequency radio waves in space.

A method for correcting the substructure of multiprong jets using the Lund jet plane

2025-11-25T06:37:07Z

A method for correcting the substructure of multiprong jets using the Lund jet plane Hayrapetyan, A.; Makarenko, V.; Tumasyan, A.; Adam, W.; Andrejkovic, J. W.; Benato, L.; Bergauer, T.; Damanakis, K.; Dragicevic, M.; Giordano, C.; Hussain, P. S.; Jeitler, M.; Krammer, N.; Li, A.; Liko, D.; Mikulec, I.; Schieck, J.; Schöfbeck, R. Many analyses at the CERN LHC exploit the substructure of jets to identify heavy resonances produced with high momenta that decay into multiple quarks and/or gluons. This paper presents a new technique for correcting the substructure of simulated large-radius jets from multiprong decays. The technique is based on reclustering the jet constituents into several subjets such that each subjet represents a single prong, and separately correcting the radiation pattern in the Lund jet plane of each subjet using a correction derived from data. The data presented here correspond to an integrated luminosity of 138 fb−1 collected by the CMS experiment between 2016–2018 at a center-of-mass energy of 13 TeV. The correction procedure improves the agreement between data and simulation for several different substructure observables of multiprong jets. This technique establishes, for the first time, a robust calibration for the substructure of jets with four or more prongs, enabling future measurements and searches for new phenomena containing these signatures.

Inclusionary and Exclusionary Preferences: A Test of Three Cognitive Mechanisms

2025-11-25T06:37:00Z

Inclusionary and Exclusionary Preferences: A Test of Three Cognitive Mechanisms Landau-Wells, Marika; Lydic, Kirsten O.; Kennedy, Joachim; Mittman, Benjamin G.; Thompson, Todd W.; Gupta, Akhil; Saxe, Rebecca Exclusionary social policies take a significant toll on the mental and physical health of targeted groups. Support for specific exclusionary policies does not always align with general antipathy towards the targeted group, however. Does support for specific exclusionary policies rely on particular thought processes (i.e., cognitive mechanisms)? Does opposition? We investigate these questions through the lens of “bathroom laws” across two studies. In Study 1, we use functional neuroimaging to test three candidate cognitive mechanisms from the literature: (1) threat-related emotions (e.g., fear, disgust) supporting exclusionary preferences; (2) mentalizing (e.g., empathy, perspective-taking) supporting inclusionary preferences; and (3) self-regulation (e.g., aligning one’s behavior with one’s goals) supporting inclusionary preferences. Consistent with the intergroup conflict and prejudice literatures, we find evidence of a motivated self-regulation mechanism in bathroom law opponents. In Study 2, we investigate a possible source of this motivation using text analysis of open-ended policy preference justifications. We find that bathroom law opponents link their policy preference to a small number of specific values, particularly autonomy of action. Taken together, these studies point to a value-driven, motivational account of inclusionary preferences that reconciles puzzling patterns of public opinion, offers new levers for tolerance interventions, and provides some insight into the brain-basis of political behavior.

Accelerated Bayesian Calibration and Uncertainty Quantification of RANS Turbulence Model Parameters for Stratified Atmospheric Boundary Layer Flows

2025-11-25T06:37:03Z

Accelerated Bayesian Calibration and Uncertainty Quantification of RANS Turbulence Model Parameters for Stratified Atmospheric Boundary Layer Flows Shin, Ethan Y.; Howland, Michael F. In operational weather models, the effects of turbulence in the atmospheric boundary layer (ABL) on the resolved flow are modeled using turbulence parameterizations. These parameterizations typically use a predetermined set of model parameters that are tuned to limited data from canonical flows. Using these fixed parameters results in deterministic predictions that neglect uncertainty in the unresolved turbulence processes. In this study, we perform a machine learning-accelerated Bayesian inversion of a single-column model of the ABL. This approach is used to calibrate and quantify uncertainty in model parameters of Reynolds-averaged Navier–Stokes turbulence models. To verify the data-driven uncertainty quantification methodology, we test in an idealized setup in which a prescribed but unobserved set of parameters is learned from noisy approximations of the model output. Following this verification, we learn the parameters and their uncertainties in two different turbulence models conditioned on scale-resolving large-eddy simulation data over a range of ABL stabilities. We show how Bayesian inversion of a numerical model improves flow predictions by investigating the underlying mean momentum budgets. Further, we show that uncertainty quantification based on neutral ABL surface layer data recovers the relationships between parameters that have been predicted using theoretical modeling, but that learning the parameters based on stable ABL data or data from outside the surface layer can lead to different parameter relationships than neutral surface layer theory. Efforts to systematically reduce parameter uncertainty reveal that (1) sampling wind speed up to the ABL height can reduce uncertainty in key model parameters by up to $$84\%$$ , and (2) assimilating fluid flow quantities beyond first-order moment statistics can further reduce uncertainty in ways that baseline wind speed assimilation alone cannot achieve. The parameters learned using Bayesian uncertainty quantification generally yield lower error than standard deterministic parameters in out-of-sample tests and also provide uncertainty intervals on predictions.

Europa Clipper Magnetometer Boom Deployment: A First Look at the Magnetometer Observations of the Spacecraft and the Interplanetary Magnetic Field

2025-11-25T06:37:05Z

Europa Clipper Magnetometer Boom Deployment: A First Look at the Magnetometer Observations of the Spacecraft and the Interplanetary Magnetic Field Cochrane, Corey J.; Joy, Steven P.; Korth, Haje; Biersteker, John B.; Blacksberg, Jordana; Bouchard, Michael; Contreras, Jacob; Dawson, Olivia R.; Khurana, Krishan K.; Murphy, Neil; Palm, Derek; Perley, Mitch O.; Pierce, David R.; Richter, Ingo; Russell, Christopher T. NASA’s Europa Clipper flagship mission is designed to investigate the habitability of Jupiter’s moon Europa. A key instrument aboard the spacecraft is the Europa Clipper Magnetometer (ECM), a suite of fluxgate magnetometer sensors deployed on a boom to minimize spacecraft-induced magnetic interference. The ECM investigation aims to characterize Europa’s induced magnetic field, offering constraints on the salinity, depth, and thickness of its subsurface ocean. This work presents the first in-flight ECM observations acquired during the magnetometer boom deployment and shortly thereafter. We show how these observations provide the requisite evidence needed to validate a successful deployment. We also demonstrate how these observations can be used to calibrate the sensor offsets and to develop new magnetic field models of the spacecraft of varying complexity, thus enabling the robust removal of the instrument’s zero-levels which is critical for achieving the mission’s science objectives. We finally share preliminary calibrated magnetometer observations acquired over a two-month period after deployment, revealing a very active interplanetary magnetic field characteristic of solar maximum.

DOE selects MIT to establish a Center for the Exascale Simulation of Coupled High-Enthalpy Fluid–Solid Interactions

2025-11-25T06:39:03Z

DOE selects MIT to establish a Center for the Exascale Simulation of Coupled High-Enthalpy Fluid–Solid Interactions Hadley, F The research center, sponsored by the DOE’s National Nuclear Security Administration, will advance the simulation of extreme environments, such as those in hypersonic flight and atmospheric reentry.

Deuteron identification via time of flight with LHCb

2025-11-25T06:36:57Z

Deuteron identification via time of flight with LHCb LHCb Collaboration It is shown that the timing capabilities of the LHCb detector operated during the LHC Run 2 can be used to identify light ion particles with momenta of a few GeV/c. This is achieved by estimating the particle time of flight through a newly developed technique. A dedicated reconstruction procedure and a neural-network-based estimator of the particle speed have been developed to enable deuteron identification by suppressing the abundant background from lighter particles. The performance of the identification procedure is demonstrated in a sample of proton-helium collisions at s NN = 110 GeV, where the production of deuteron and triton particles is observed. This novel approach opens the way to study deuteron and antideuteron production for different collision systems at different energy scales, exploiting the rich dataset collected by the LHCb experiment.

Confidently Comparing Estimates with the c-value

2025-11-25T06:37:35Z

Confidently Comparing Estimates with the c-value Trippe, Brian L; Deshpande, Sameer K; Broderick, Tamara Modern statistics provides an ever-expanding toolkit for estimating unknown parameters. Consequently, applied statisticians frequently face a difficult decision: retain a parameter estimate from a familiar method or replace it with an estimate from a newer or more complex one. While it is traditional to compare estimates using risk, such comparisons are rarely conclusive in realistic settings. In response, we propose the “c-value” as a measure of confidence that a new estimate achieves smaller loss than an old estimate on a given dataset. We show that it is unlikely that a large c-value coincides with a larger loss for the new estimate. Therefore, just as a small p-value supports rejecting a null hypothesis, a large c-value supports using a new estimate in place of the old. For a wide class of problems and estimates, we show how to compute a c-value by first constructing a data-dependent high-probability lower bound on the difference in loss. The c-value is frequentist in nature, but we show that it can provide validation of shrinkage estimates derived from Bayesian models in real data applications involving hierarchical models and Gaussian processes. Supplementary materials for this article are available online.

AI Accelerator Announces Award Winners

2025-11-25T06:39:01Z

AI Accelerator Announces Award Winners Accelerator, AI The Department of Air Force (DAF)-MIT AI Accelerator is a unique collaboration designed to advance the field of AI to improve DAF operations and address broader societal needs. In June 2025, the DAF-MIT Artificial Intelligence Accelerator named the recipients of AI Accelerator awards, recognizing scientific excellence, distinguished contributions, and other exceptional accomplishments. The awardees were nominated and selected from members of the AI Accelerator community, including individuals from the DAF, MIT campus, and MIT Lincoln Laboratory.

Future circular collider feasibility study report

2025-11-25T06:37:26Z

Future circular collider feasibility study report Benedikt, M.; Zimmermann, F.; Auchmann, B.; Bartmann, W.; Burnet, J. P.; Carli, C.; Chancé, A.; Craievich, P.; Giovannozzi, M.; Grojean, C.; Gutleber, J.; Hanke, K.; Henriques, A.; Janot, P.; Lourenço, C.; Mangano, M.; Otto, T.; Poole, J.; Rajagopalan, S.; Raubenheimer, T. In response to the 2020 Update of the European Strategy for Particle Physics, the Future Circular Collider (FCC) Feasibility Study was launched as an international collaboration hosted by CERN. This report describes the FCC integrated programme, which consists of two stages: an electron-positron collider (FCC-ee) in the first phase, serving as a high-luminosity Higgs, top, and electroweak factory; followed by a proton-proton collider (FCC-hh) at the energy frontier in the second phase. The FCC-ee is designed to operate at four key centre-of-mass energies: the Z pole, the WW pair production threshold, the ZH production peak, and the top/anti-top production threshold—each delivering the highest possible luminosities to four experiments. Over 15 years of operation, FCC-ee will produce more than 6 trillion Z bosons, 200 million WW pairs, nearly 3 million Higgs bosons, and 2 million top anti-top pairs. Precise energy calibration at the Z pole and WW threshold will be achieved through frequent resonant depolarisation of pilot bunches. The sequence of operation modes between the Z, WW, and ZH substages remains flexible. The FCC-hh will operate at a centre-of-mass energy of approximately 85 TeV—nearly an order of magnitude higher than the LHC—and is designed to deliver 5 to 10 times the integrated luminosity of the upcoming High-Luminosity LHC. Its mass reach for direct discovery extends to several tens of TeV. In addition to proton-proton collisions, the FCC-hh is capable of supporting ion-ion, ion-proton, and lepton-hadron collision modes. This second volume of the Feasibility Study Report presents the complete design of the FCC-ee collider, its operation and staging strategy, the full-energy booster and injector complex, required accelerator technologies, safety concepts, and technical infrastructure. It also includes the design of the FCC-hh hadron collider, development of high-field magnets, hadron injector options, and key technical systems for FCC-hh.

Nitrous Oxide Distributions in the Oxygenated Water Column of the Sargasso Sea

2025-11-25T06:37:36Z

Nitrous Oxide Distributions in the Oxygenated Water Column of the Sargasso Sea Meyer, Annaliese C. S.; Cullen, Jay T.; Grundle, Damian S. This study presents dissolved nitrous oxide (N2O) concentrations in the water column at the Bermuda Atlantic Time-series Study (BATS) station and uses a subset of these measurements to estimate air-to-sea flux for four specific time points between September 2018 and June 2019. N2O concentrations at BATS were in the range of 4.0 nmol L−1–16.9 nmol L−1, with vertical profiles which were the mirror inverse of dissolved oxygen. Regardless of season, N2O concentration maxima were found within the oxygen minimum zone (OMZ). The highest maximum N2O values were observed in November and lowest in October. As the water column at BATS remains consistently at dissolved oxygen concentrations greater than 140 µmol L−1, and therefore aerobic, we assume that the bulk of N2O production occurs through nitrification. A nitrification source is supported by a correlation between excess N2O (ΔN2O) below the mixed layer, apparent oxygen utilization (AOU) and nitrate concentrations. We estimate a pooled average yield of 0.027% to 0.038% N2O from nitrification at BATS. Finally, estimates of air–sea exchange of N2O using regional average monthly wind speeds indicated that this region acts as a weak source or a sink of atmospheric N2O, and varies between months.

Responding to the Climate Impact of Generative AI

2025-11-25T06:39:02Z

Responding to the Climate Impact of Generative AI Zewe, Adam Explosive growth of AI data centers is expected to increase greenhouse gas emissions. Researchers are now seeking solutions to reduce these environmental harms.

A holistic model for understanding the dynamics of outsourcing

2025-11-25T06:37:38Z

A holistic model for understanding the dynamics of outsourcing Uygun, Yilmaz; Gotsadze, Nikoloz; Schupp, Florian; Gzirishvili, Lizi; Tindjou Nana, Brigitte Stephanie Outsourcing is a complex process as many external and internal factors that look convincing in the first place might, however, lead to a failure in the long run. Motivated by this, we wanted to get a holistic understanding of such outsourcing decisions. Thus, we created a comprehensive System Dynamics simulation model including all relevant variables to examine the dynamic nature of outsourcing in a holistic manner and over time that consists of more than 200 interrelated variables. Our results show, amongst others, that higher process specialisation that requires substantial investments by the supplier appears to be favourable for an outsourcing company and shifting a larger quantity to such a supplier achieves better cost savings and thus accounts for a better overall outsourcing result. On an operational level, we identified an innovation trap, a bargaining power shift, a plagiarism trap, and a knowledge trap. Based on that, we give specific managerial recommendations to tackles these aspects. We conclude that, amongst others, it is important for innovative companies with rather complex processes and parts to carefully plan which and how many employees to release so as not to lose the knowledge on those outsourced processes and parts.

Lincoln Lab Unveils the Most Powerful AI Supercomputer at any US University

2025-11-25T06:38:54Z

Lincoln Lab Unveils the Most Powerful AI Supercomputer at any US University Foy, Kylie Optimized for generative AI, TX-GAIN is driving innovation in biodefense, materials discovery, cybersecurity, and other areas of research and development.

Mediation and ANCOVA Models to Study the Influence of Solvent Retting Traits and Plant Physique on Bast Fiber Yield and Retting Time

2025-11-25T06:37:33Z

Mediation and ANCOVA Models to Study the Influence of Solvent Retting Traits and Plant Physique on Bast Fiber Yield and Retting Time Shuvo, Ikra Iftekhar; Hoque, Md. Saiful; Khandakar, Lovely K. M. The study aims in applying two statistical tools to analyze the retting beha-vior of plant stems for extracting bast fibers for industrial applications. Atfirst, a mediation model is employed to investigate the first hypothesis of thiswork that involves studying the color response of the retted solvent asa function of retting time on the responsible variable, fiber yield (%).Statistically, there is a significant indirect effect of retting time on fiberyield (%) through retting trait (β = −0.0142, 95% C.I. [−0.0274, −0.0011]) –a statistical inference bolstered by the Sobel test result, confirming themediation effect (p-value = 0.0329 < 0.05; z-score = −2.1334; bootstrappingof 5000 resamples). Next, the second hypothesis of the current work involvesanalyzing the impact of stem form-factors on their retting time using thestatistical tool, ANCOVA. The partial- η2 indicates that cultivar treatmentaccounts for 30% variance of the retting time while controlling for the effectsof two covariates – diameter and length of the stems, in this case. Bycontrolling the Type-I error, Bonferroni and similar post-hoc tests also con-firm the statistical significance of cultivar categories pertaining to their meanretting time. Future work could focus on these underlying hypotheses andstudy the impact of microorganisms, environmental factors, and cultivartreatment variables on the retting time to optimize the overall fiber yieldand production process.

Improving Autistic Experiences in the Workplace: Key Factors and Actionable Steps

2025-11-25T06:36:41Z

Improving Autistic Experiences in the Workplace: Key Factors and Actionable Steps Nishith, Shruti; O’Brien, Amanda M.; Li, Cindy; Bungert, Lindsay; Oddis, Kyle; Riddle, Joseph; Gabrieli, John D. E. Autistic adults have higher rates of unemployment and underemployment than non-autistic adults with and without disabilities. While previous work has highlighted factors specific to individuals and/or job sectors that serve as barriers or facilitators to autistic employment, the question of how to modify the workplace to best support autistic people remains under-researched. The present study utilized an ecological framework to investigate what workplace factors can be modified to improve autistic experiences and how these modifications may be enacted across different levels of workplace ecosystem to promote autistic success. Autistic participants (N = 85) across employment sectors provided quantitative ratings and written descriptions of positive and negative factors related to their workplace experiences. Quantitative and qualitative analyses were used to examine which factors and overarching principles most impact employment. Actionable strategies to modify these factors were derived from participant responses and validated by autistic collaborators and neuroinclusion experts. On average, participants rated task training as having the most positive, and mental health as having the most negative, impact on their employment. Participants described four themes (acceptance, communication, autonomy, accommodations) that can be embedded in the work environment to improve experiences. Steps to improve autistic employment that can be enacted by stakeholders across levels of the workplace experiences are provided. Autistic adults face multifaceted barriers to employment across levels of the workplace. Modifying the workplace itself, across multiple levels and stakeholders, may serve to improve autistic employment outcomes.

The Connectivity of Friends-and-Strangers Graphs on Complete Multipartite Graphs

2025-11-25T06:36:39Z

The Connectivity of Friends-and-Strangers Graphs on Complete Multipartite Graphs Zhu, Honglin For simple graphs X and Y on n vertices, the friends-and-strangers graph FS ( X , Y ) is the graph whose vertex set consists of all bijections σ : V ( X ) → V ( Y ) , where two bijections σ and σ ′ are adjacent if and only if they agree on all but two adjacent vertices a , b ∈ V ( X ) such that σ ( a ) , σ ( b ) ∈ V ( Y ) are adjacent in Y. Resolving a conjecture of Wang, Lu, and Chen, we completely characterize the connectedness of FS ( X , Y ) when Y is a complete bipartite graph. We further extend this result to when Y is a complete multipartite graph. We also determine when FS ( X , Y ) has exactly two connected components where X is bipartite and Y is a complete bipartite graph.

Reducing Aerodynamic Interference Through Layout Optimization of Symmetrically Cambered Wingsails: A Comparative Study of In-Line and Parallel Configurations

2025-11-25T06:37:30Z

Reducing Aerodynamic Interference Through Layout Optimization of Symmetrically Cambered Wingsails: A Comparative Study of In-Line and Parallel Configurations van Reen, Stephan; Lin, Jianfeng; Niu, Jiqiang; Sharpe, Peter; Li, Xiaodong; Yao, Hua-Dong Rigid wingsails are increasingly adopted for wind-assisted ship propulsion, with Symmetrically Cambered (SC) profiles identified as highly efficient for thrust generation. This study investigates installation layouts for multiple SC wingsails, focusing on aerodynamic interference that limits their performance. A fast 2D potential-flow panel method is employed and benchmarked against wind tunnel and 3D IDDES data. Two representative layouts are analyzed: triple-in-line (TL) and quad-in-parallel (QP). Layout optimization is performed using a genetic algorithm with distances between sails as design variables, constrained by the total installation span, at apparent wind angles (AWAs) of 60◦ , 90◦ , and 120◦ . Results show that thrust generation decreases progressively from upstream to downstream sails due to interference effects, with penalties of about 4–6% in the TL and up to 28% in the QP layout. The optimization improves performance only for the TL layout at 60◦ , while the QP layout shows negligible gains. Analysis of pressure distributions confirms that downstream sails suffer from reduced suction on the leading edge caused by upstream wakes. Overall, the TL layout demonstrates significantly higher aerodynamic reliability than the QP layout. These findings provide new insights into multi-sail configurations and highlight the importance of layout optimization in maximizing thrust efficiency.

A rapid experimental workflow for studying melt track scaling in laser powder bed fusion using high-precision metal template substrates

2025-11-25T06:36:30Z

A rapid experimental workflow for studying melt track scaling in laser powder bed fusion using high-precision metal template substrates Weissbach, Reimar; Penny, Ryan W.; Hart, A. J. Development and qualification of process parameters in laser powder bed fusion (LPBF) involves many variables. At the outset of development, whether transferring known parameters to a new machine, or exploring a new material, single-track and single-layer experiments are a convenient means of down-selecting key variables and exploring parameter scaling behavior. We present an experimental workflow for single-layer LPBF experiments using high-precision metal template substrates, overcoming challenges with precision single-layer alignment in LPBF systems and enabling efficient processing and cross-sectional analysis. Templates are fabricated using chemical etching and machining, and are characterized using optical profilometry and X-ray transmission imaging of powder layers. Using the etched templates, a single-track parameter study is performed in SS316 including three powder layer thicknesses, and spanning common laser melting modes (lack-of-fusion, conduction, and keyhole mode). Analysis of melt track geometries using automated image processing allows a scaling law to be applied to define the process window, quantifying the amount of material added with increasing powder layer thickness. Single-track results are verified with raster scanning experiments, showing the potential to transfer single-track results to full LPBF builds.

Augmented intelligence should be good for medicine, if medicine is to remain good for us

2025-11-25T06:36:35Z

Augmented intelligence should be good for medicine, if medicine is to remain good for us Idan, Daphna; Celi, Leo A.; Einav, Sharon; Frenkel, Amit Throughout history, the medical community has failed to address health disparities. Augmented intelligence (AI) is poised to cement these structural inequities permanently. The need to establish a triage process that ensures fair and equitable access to medical care, and to consider all patient populations equally researchable, should not overshadow the need to learn how best to exploit AI for furthering medical fairness and equity despite resource limitations. Open discussion of the shortcomings of medical AI, approaching medical AI development, testing, and implementation from a critical ethical perspective, constant testing and analysis of AI outputs, and human oversight in the loop constitute only the first part of ensuring augmented intelligence tools are equitably robust and free of bias.

Design Principles and Impact of a Learning Analytics Dashboard: Evidence from a Randomized MOOC Experiment

2025-11-25T06:36:51Z

Design Principles and Impact of a Learning Analytics Dashboard: Evidence from a Randomized MOOC Experiment Borrella, Inma; Ponce-Cueto, Eva Learning Analytics Dashboards (LADs) are increasingly deployed to support self-regulated learning on online courses. Yet many existing dashboards lack strong theoretical grounding, contextual alignment, or actionable feedback, and some designs have been shown to inadvertently discourage learners through excessive social comparison or high inference costs. In this study, we designed and evaluated a LAD grounded in the COPES model of self-regulated learning and tailored to a credit-bearing Massive Open Online Course (MOOC) using a data-driven approach. We conducted a randomized controlled trial with 8745 learners, comparing a control group, a dashboard without feedback, and a dashboard with ARCS-framed actionable feedback. The results showed that the dashboard with feedback significantly increased learners’ likelihood of verification (i.e., paying for the certification track), with mixed effects on engagement and no measurable impact on final grades. These findings suggest that dashboards are not uniformly beneficial: while feedback-supported LADs can enhance motivation and persistence, dashboards that lack interpretive support may impose cognitive burdens without improving outcomes. This study contributes to the literature on learning analytics by (1) articulating the design principles for theoretically and contextually grounded LADs and (2) providing experimental evidence on their impact in authentic MOOC settings.

The Effect of IL-17A and Combined Mechanical Injury on Meniscal Tissue Integrity In Vitro

2025-11-25T06:37:29Z

The Effect of IL-17A and Combined Mechanical Injury on Meniscal Tissue Integrity In Vitro Ahrens, Greta; Gellhaus, Florian; Weitkamp, Jan-Tobias; Behrendt, Peter; Cossais, François; Rolauffs, Bernd; Grodzinsky, Alan J; Kurz, Bodo Objectives: Meniscal integrity is crucial for knee joint stability and the prevention of osteoarthritis (OA) development. Recent studies suggested that mechanical overload and interleukin (IL)-17A may be important intertwined players in meniscal degeneration, but a direct impact of IL-17A on the meniscus has not been investigated. Therefore, the aim of this study was to analyze the effect of IL-17A on meniscal tissue with and without combined mechanical injury (MI). Methods: Meniscal explant disks (1 mm height, 3 mm diameter) were isolated from bovine menisci (preserving the native tibial superficial zone) and exposed to IL-17A [0–100 ng/mL] and/or MI (single compression, 50% strain, strain rate 1 mm/sec). After three days of incubation in a serum-free medium, the proteoglycan release (sGAG; DMMB assay), mRNA level of matrix-degrading enzymes (qRT-PCR), aggrecan degradation (NITEGE immunostaining), and cell death (histomorphometry of nuclear blebbing/apoptosis and condensed nuclei/unspecified cell death) were determined. Statistics: one- and two-way ANOVA with Tukey’s multiple comparisons or Kruskal– Wallis with post hoc testing. Results: IL-17A increased sGAG release in a dose-dependent significant manner. MI also induced the release of sGAG significantly, but the combination with IL-17A showed the highest levels. Both IL-17A and MI individually affected the mRNA levels for ADAMTS4 and MMP-13 slightly, but the combination of both particularly induced a significant increase in mRNA levels. Signals for the ADAMTS4-related aggrecan neoepitope NITEGE were elevated by IL-17A in superficial areas of the excised tissue and by MI in superficial and deeper areas. The combination of both stimuli intensified this signal further. MI increased the number of cells with condensed nuclei significantly and induced apoptosis in a small proportion of cells. IL-17A had no significant impact on the amount of condensed or apoptotic nuclei. Conclusions: Our findings emphasize an interaction between inflammatory cytokine IL-17A signaling and mechanical stress since IL17A induced matrix degeneration in meniscal tissue, which intensified in combination with a trauma. The latter might create a post-traumatic environment that promotes meniscal degeneration and subsequently osteoarthritis progression.

Molecular hallmarks of excitatory and inhibitory neuronal resilience to Alzheimer’s disease

2025-11-25T06:36:38Z

Molecular hallmarks of excitatory and inhibitory neuronal resilience to Alzheimer’s disease Castanho, Isabel; Naderi Yeganeh, Pourya; Boix, Carles A.; Morgan, Sarah L.; Mathys, Hansruedi; Prokopenko, Dmitry; White, Bartholomew; Soto, Larisa M.; Pegoraro, Giulia Background A significant proportion of individuals maintain cognition despite extensive Alzheimer’s disease (AD) pathology, known as cognitive resilience. Understanding the molecular mechanisms that protect these individuals could reveal therapeutic targets for AD. Methods This study defines molecular and cellular signatures of cognitive resilience by integrating bulk RNA and single-cell transcriptomic data with genetics across multiple brain regions. We analyzed data from the Religious Order Study and the Rush Memory and Aging Project (ROSMAP), including bulk RNA sequencing (n = 631 individuals) and multiregional single-nucleus RNA sequencing (n = 48 individuals). Subjects were categorized into AD, resilient, and control based on β-amyloid and tau pathology, and cognitive status. We identified and prioritized protected cell populations using whole-genome sequencing-derived genetic variants, transcriptomic profiling, and cellular composition. Results Transcriptomics and polygenic risk analysis position resilience as an intermediate AD state. Only GFAP and KLF4 expression distinguished resilience from controls at tissue level, whereas differential expression of genes involved in nucleic acid metabolism and signaling differentiated AD and resilient brains. At the cellular level, resilience was characterized by broad downregulation of LINGO1 expression and reorganization of chaperone pathways, specifically downregulation of Hsp90 and upregulation of Hsp40, Hsp70, and Hsp110 families in excitatory neurons. MEF2C, ATP8B1, and RELN emerged as key markers of resilient neurons. Excitatory neuronal subtypes in the entorhinal cortex (ATP8B+ and MEF2Chigh) exhibited unique resilience signaling through activation of neurotrophin (BDNF-NTRK2, modulated by LINGO1) and angiopoietin (ANGPT2-TEK) pathways. MEF2C+ inhibitory neurons were over-represented in resilient brains, and the expression of genes associated with rare genetic variants revealed vulnerable somatostatin (SST) cortical interneurons that survive in AD resilience. The maintenance of excitatory-inhibitory balance emerges as a key characteristic of resilience. Conclusions We have defined molecular and cellular hallmarks of cognitive resilience, an intermediate state in the AD continuum. Resilience mechanisms include preserved neuronal function, balanced network activity, and activation of neurotrophic survival signaling. Specific excitatory neuronal populations appear to play a central role in mediating cognitive resilience, while a subset of vulnerable interneurons likely provides compensation against AD-associated hyperexcitability. This study offers a framework to leverage natural protective mechanisms to mitigate neurodegeneration and preserve cognition in AD.

Nationwide Trends in Hospitalizations for Sudden Cardiac Arrest Before and During the COVID Outbreak

2025-11-25T06:36:53Z

Nationwide Trends in Hospitalizations for Sudden Cardiac Arrest Before and During the COVID Outbreak Daoudi, Sarah; Furer, Ariel; John, Kevin; Chalhoub, Fadi; Chee, Jennifer; Infeld, Margaret; Elbaz-Greener, Gabby; Homoud, Munther; Udelson, James; Madias, Christopher; Rozen, Guy Background/Objectives: Sudden cardiac arrest (SCA) accounts for ~50% of cardiovascular mortality in the U.S. Cardiovascular complications are common in acute and post-acute COVID-19 infection. We aimed to examine nationwide trends in SCA-related hospitalizations in the United States before and during the COVID-19 outbreak. Methods: Using data from the National Inpatient Sample, we conducted a retrospective analysis of hospitalizations for SCA in the U.S. between 2016 and 2020. Sociodemographic and clinical characteristics and in-hospital mortality were compared between the pre-COVID (2016– 2019) and COVID (2020) eras. Multivariable analysis was performed to identify factors associated with mortality. Results: Among a weighted total of 153,100 SCA hospitalizations between 2016 and 2020, the median age was 65 years, 62.7% were male, and 66.6% were white. There was a trend towards fewer hospitalizations in 2020 compared to prior years (n = 28,585 vs. naverage = 32,129, p = 0.07). In-hospital mortality remained unchanged between the pre-COVID and COVID eras (47.7% vs. 47.3%, p = 0.66). Increased mortality was associated with female sex (OR: 1.21; 95% CI: 1.15–1.28; p < 0.001), non-white race (OR: 1.24; 95% CI: 1.15–1.28; p < 0.001), history of renal failure (OR: 1.08; 95% CI: 1.02–1.15; p = 0.007), and diabetes (OR: 1.32; 95% CI: 1.25–1.39; p < 0.001). In 2020, 1.5% of the study population was diagnosed with COVID-19 infection, which was found to be independently associated with increased in-hospital mortality (OR: 1.57; 95% CI: 1.27–1.95; p < 0.001). Conclusions: In 2020, there was a trend towards a decrease in hospitalizations for SCA, while COVID-19 infection was independently associated with higher in-hospital mortality among patients admitted with SCA.

Biosensor development for single-cell detection of glucuronate

2025-11-25T06:36:58Z

Biosensor development for single-cell detection of glucuronate Nash, Jennifer Kaczmarek; Prather, Kristala LJ Recent work in biosensors has shown promise to enable high throughput searches through large genetic libraries. However, just as physiological limitations and lack of in-depth mechanistic knowledge can prevent us from achieving high titers in microbial systems; similar roadblocks can appear in the application of biosensors. Here, we characterized a previously developed transcription-factor (ExuR) based galacturonate biosensor for its other cognate ligand, glucuronate. Though we saw an ideal response to glucuronate from the biosensor in controlled and ideal experimental circumstances, these results began to deviate from a well-behaved system when we explored the application of the sensor to different MIOX homologs. Through modifications to circuit architecture and culture conditions, we were able to decrease this variation and use these more optimal conditions to apply the biosensor for the separation of two closely related MIOX homologs.

Strategies in engineering sustainable biochemical synthesis through microbial systems

2025-11-22T03:15:33Z

Strategies in engineering sustainable biochemical synthesis through microbial systems Song, Yoseb; Prather, Kristala LJ Growing environmental concerns and the urgency to address climate change have increased demand for the development of sustainable alternatives to fossil-derived fuels and chemicals. Microbial systems, possessing inherent biosynthetic capabilities, present a promising approach for achieving this goal. This review discusses the coupling of systems and synthetic biology to enable the elucidation and manipulation of microbial phenotypes for the production of chemicals that can substitute for petroleum-derived counterparts and contribute to advancing green biotechnology. The integration of artificial intelligence with metabolic engineering to facilitate precise and data-driven design of biosynthetic pathways is also discussed, along with the identification of current limitations and proposition of strategies for optimizing biosystems, thereby propelling the field of chemical biology towards sustainable chemical production.

Report to the President for year ended June 30, 2025, Office of the Vice Chancellor for Undergraduate and Graduate Education

2025-11-22T03:18:43Z

Report to the President for year ended June 30, 2025, Office of the Vice Chancellor for Undergraduate and Graduate Education Darmofal, David This report contains the following sections: Executive Summary, OVC Headquarters, Office of Admissions, Career Advising & Professional Development, Office of Experiential Learning, Edgerton Center, PKG Public Service Center, Undergraduate Research Opportunities Program, Concourse, Experimental Study Group, Terrascope, Office of Graduate Education, International Students Office, Registrar’s Office, Air Force ROTC, Army ROTC, Navy ROTC, Student Financial Services, Teaching + Learning Lab, and Undergraduate Advising Center.

Covert reciprocals: a scope-based analysis of reciprocal alternations

2026-03-08T03:26:44Z

Covert reciprocals: a scope-based analysis of reciprocal alternations Wehbe, Jad This paper argues that the class of predicates that participate in reciprocal alternations, like the seemingly 1-place predicate hug in Jane and Mary hugged, should in fact be analyzed as 2-place predicates with a covert reciprocal in object position. The main challenge for this analysis is that there are truth-conditional differences between covert reciprocals and their overt counterparts. Focusing on a few case studies, this paper will argue that these seemingly lexical differences can be reanalyzed in terms of scope, allowing the differences to be systematically predicted once appropriate scope restrictions on covert reciprocals are established. More specifically, I propose that covert reciprocals are simply reciprocals that have to be bound at the lowest possible scope position. I show that these seemingly 1-place predicates behave just like overt reciprocals, modulo the low-scope requirement, for example giving rise to homogeneity and non-maximality. I therefore conclude that in order to account systematically for these inferences, covert reciprocals (at least the case studies that the paper considers) must be treated as having the same LFs as low-scope overt reciprocals.

Robust resonant anomaly detection with NPLM

2026-03-08T03:26:44Z

Robust resonant anomaly detection with NPLM Grosso, Gaia; Sengupta, Debajyoti; Golling, Tobias; Harris, Philip In this study, we investigate the application of the New Physics Learning Machine (NPLM) algorithm as an alternative to the standard CWoLa method with Boosted Decision Trees (BDTs), particularly for scenarios with rare signal events. NPLM offers an end-to-end approach to anomaly detection and hypothesis testing by utilizing an in-sample evaluation of a binary classifier to estimate a log-density ratio, which can improve detection performance without prior assumptions on the signal model. We examine two approaches: (1) a end-to-end NPLM application in cases with reliable background modelling and (2) an NPLM-based classifier used for signal selection when accurate background modelling is unavailable, with subsequent performance enhancement through a hyper-test on multiple values of the selection threshold. Our findings show that NPLM-based methods outperform BDT-based approaches in detection performance, particularly in low signal injection scenarios, while significantly reducing epistemic variance due to hyperparameter choices. This work highlights the potential of NPLM for robust resonant anomaly detection in particle physics, setting a foundation for future methods that enhance sensitivity and consistency under signal variability.

Tackling the Cardio-Kidney-Metabolic Burden in Cancer

2025-11-22T03:15:06Z

Tackling the Cardio-Kidney-Metabolic Burden in Cancer Nahle, Tarek; Shah, Viraj; Kunhiraman, Harikrishnan H.; Makram, Omar M.; Ahmed, Ola; Yerraguntla, Sandeep; Gopu, Gaurav; Vy, Jenny; Singh, Shivam; Borse, Tanvi; Kalinsky, Kevin; Deswal, Anita; Sadler, Diego; Chitalia, Vipul; Weintraub, Neal L. Purpose of the Review This review aims to examine the clinical relevance of cardio-kidney-metabolic syndrome (CKMS) in oncology, highlighting its role as both a preexisting comorbidity and a consequence of cancer treatment. It aims to integrating CKMS staging into personalized cancer care. Recent Findings CKMS is a progressive syndrome marked by dysfunction across cardiovascular, renal, and metabolic systems. Cancer therapies—particularly hormonal agents, immune checkpoint inhibitors, and chemotherapeutics—can accelerate or reveal underlying CKMS through inflammatory and metabolic pathways. Early risk stratification based on CKMS stage enables more effective monitoring, referral, and therapeutic strategies. A stage-based, multidisciplinary approach tailored to cancer type and comorbidity burden is essential for optimizing outcomes. Summary With rising multimorbidity among cancer patients, recognizing and addressing CKMS is increasingly critical. Routine CKMS assessment in oncology offers a pathway for earlier intervention and potentially altering its course. A comprehensive, individualized care model based on CKS stage is necessary to mitigate CKMS-related complications and deliver high-quality, integrated cancer care.

Initial checkout of the Psyche electric propulsion system

2026-03-08T03:22:59Z

Initial checkout of the Psyche electric propulsion system Snyder, John S.; Kelly, Charles L.; Garner, Charles; Bradley, Nicholas; Johnson, Ian; Corey, Ron; Ream, Jodie B.; Weiss, Benjamin P. NASA’s Psyche spacecraft launched on October 13, 2023, and soon afterward the mission operations team began spacecraft initial checkout activities. For the electric propulsion system, the feed system and thruster gimbals were first prepared and then the rest of the subsystem completed an initial operations test during thruster bakeout. Thrust for each thruster was measured across the full range of operating powers and was in good agreement with pre-flight expectations. A weeklong test of the spacecraft and mission operations plan during thrusting activities was successful, but a thruster burn-in phenomenon was observed during full power operation that was longer than expected based on previous flight history. Data accumulated during the initial checkout activities shows that this burn-in behavior is different for each thruster and suggests that it is a result of the thruster discharge transitioning between two different plasma modes that can be mitigated by reducing discharge power and by adjusting the thruster magnet current. At the conclusion of the checkout activities, the subsystem had accumulated 357 h of thrusting operations while consuming 18.5 kg of propellant and was fully ready to begin the cruise phase of the mission.

Derandomizing Logspace With a Small Shared Hard Drive

2026-03-08T03:26:39Z

Derandomizing Logspace With a Small Shared Hard Drive Pyne, Edward We obtain new catalytic algorithms for space-bounded derandomization. In the catalytic computation model introduced by (Buhrman, Cleve, Koucký, Loff, and Speelman STOC 2013), we are given a small worktape, and a larger catalytic tape that has an arbitrary initial configuration. We may edit this tape, but it must be exactly restored to its initial configuration at the completion of the computation. We prove that B P S P A C E [ S ] ⊆ C S P A C E [ S , S 2 ] where B P S P A C E [ S ] corresponds to randomized space S computation, and C S P A C E [ S , C ] corresponds to catalytic algorithms that use O(S) bits of workspace and O(C) bits of catalytic space. Previously, only B P S P A C E [ S ] ⊆ C S P A C E [ S , 2 O ( S ) ] was known. In fact, we prove a general tradeoff, that for every α ∈ [ 1 , 1.5 ] , B P S P A C E [ S ] ⊆ C S P A C E [ S α , S 3 - α ] . We do not use the algebraic techniques of prior work on catalytic computation. Instead, we develop an algorithm that branches based on if the catalytic tape is conditionally random, and instantiate this primitive in a recursive framework. Our result gives an alternate proof of the best known time-space tradeoff for B P S P A C E [ S ] , due to (Cai, Chakaravarthy, and van Melkebeek, Theory Comput. Sys. 2006). As a final application, we extend our results to solve search problems in C S P A C E [ S , S 2 ] . As far as we are aware, this constitutes the first study of search problems in the catalytic computing model.

Report to the President for year ended June 30, 2004, Center for Real Estate

2025-11-22T03:17:40Z

Report to the President for year ended June 30, 2004, Center for Real Estate Geltner, David M This report contains the following sections: Strategic Planning Initiative; Education; Research Activities; Professional Education and Industry Interface; Membership; Alumni Outreach; Administration

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Minority Graduate Students by Course and Year

2025-11-22T03:17:13Z

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Minority Graduate Students by Course and Year MIT Registrar's Office

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Minority Undergraduates by Course and Year

2025-11-22T03:16:55Z

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Minority Undergraduates by Course and Year MIT Registrar's Office

Report to the President for year ended June 30, 2004, Vice President for Human Resources and Equal Opportunity Officer

2025-11-22T03:17:47Z

Report to the President for year ended June 30, 2004, Vice President for Human Resources and Equal Opportunity Officer Avakian, Laura; Lima, Philip; Foley, Shawn; Roberts, Barbara; Weiss, Ellen; Jablon, Barbara; Paulding, Claire; Culver, Kande; Pierce, Marianna; Gray, Margaret Ann; Williams, Wendy; Jacobs, Annette; Friscino, Deborah; Murray, Mary; O’Keefe, Eileen; Joyce, Shelagh; Wattendorf, Maryann; Simpson, Rae; Simons, Kathy This report contains the following sections: Highlights of AY2004; Labor and Employee Relations Issues; Staff Diversity, Affirmative Action, and Equal Opportunity Management Team; Benefits Services; Disabilities Services Office; Retirement Programs Office; Compensation; Human Resources Information Systems; MIT Rewards and Recognition; Labor and Employee Relations; Organization and Employee Development; Organization Development Services; Center for Career Planning at MIT; Professional Development Programs; Staffing Services; MIT Medical; Center for Work, Family & Personal Life

Report to the President for year ended June 30, 2004, Operations: Environmental Programs Office and Environment, Health, and Safety Office

2025-11-22T03:18:23Z

Report to the President for year ended June 30, 2004, Operations: Environmental Programs Office and Environment, Health, and Safety Office Keith, Jamie Lewis; Van Schalkwyk, William; DiBerardinis, Lou This report contains the following sections: Highlights; Positive EHS Initiatives and Collaborations; Communications, Outreach, and Awareness; Security and Emergency Preparedness Programs; Regulatory Interactions

Report to the President for year ended June 30, 2004, Appendix C: Personnel Changes

2025-11-22T03:17:46Z

Report to the President for year ended June 30, 2004, Appendix C: Personnel Changes MIT Registrar's Office

Report to the President for year ended June 30, 2004, Association of Alumni and Alumnae of MIT

2025-11-22T03:16:50Z

Report to the President for year ended June 30, 2004, Association of Alumni and Alumnae of MIT Garvin HM, Elizabeth A. This report contains the following sections: By the Numbers; Organizational Change; Special Initiatives; The Alumni Fund; Fund Staff Restructuring; Responsibilities and Goals; Alumni Fund Strategic Initiatives; Fund Results; MIT Capital Campaign; Alumni Activities; Alumni Clubs and Regional Programs; Affinity Groups; Tech Reunions and Class Programs; Enterprise Forum; MIT Parents Association; Student and Young Alumni Program; Alumni Education; Online Services; Alumni Career Services; Travel Program; Volunteers, Leadership, and Governance; Association Board of Directors; National Selection Committee; National Boards and Directors; Alumni Leadership Conference; Association Volunteer Awards; Communications Department; Web and Electronic Communications; Operations and Information Systems; Web-Based Systems Achievements; Office of Records; Process Improvements; Personnel and Operations Update; Association Staff; Renovations;

Report to the President for year ended June 30, 2004, Operations: Senior Counsel’s Office

2025-11-22T03:18:03Z

Report to the President for year ended June 30, 2004, Operations: Senior Counsel’s Office Keith, Jamie Lewis This report contains the following sections: Highlights; Serving Education and the Nation through Diversity; Service in the Post-9/11 Environment; Supporting Human Resources Management; Supporting Research and Other Compliance Initiatives; Risk Management and Litigation;

Report to the President for year ended June 30, 2004, Student Support Services

2025-11-22T03:17:53Z

Report to the President for year ended June 30, 2004, Student Support Services Simonis, Jackie; Henderson, Arnold, Jr.; Randolph, Robert M. This report contains the following sections: Counseling and Support Services; Religious Life at MIT

Report to the President for year ended June 30, 2004, Student Life Programs

2025-11-22T03:18:10Z

Report to the President for year ended June 30, 2004, Student Life Programs Baker, Barbara A. This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Summary of Staffing Changes

Report to the President for year ended June 30, 2004, Housing

2025-11-22T03:18:10Z

Report to the President for year ended June 30, 2004, Housing Nilsson, Karen A. This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Housing Staffing Changes

Report to the President for year ended June 30, 2004, Administrative Services

2025-11-22T03:18:22Z

Report to the President for year ended June 30, 2004, Administrative Services Capone, Laura; Salamone, Frank This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Summary of Staffing Changes

Report to the President for year ended June 30, 2004, Dean for Student Life

2025-11-22T03:17:54Z

Report to the President for year ended June 30, 2004, Dean for Student Life Benedict, Larry G. This report contains the following sections: Accomplishments;

Report to the President for year ended June 30, 2004, Edgerton Center

2025-11-22T03:18:20Z

Report to the President for year ended June 30, 2004, Edgerton Center Vandiver, J. Kim This report contains the following sections: Service Learning; Curricular Initiative for Development Design; IDEAS Competition; Ongoing Programs; Staff Changes

Report to the President for year ended June 30, 2004, Graduate Student Council

2025-11-22T03:16:53Z

Report to the President for year ended June 30, 2004, Graduate Student Council Singh, Barun; Hernandez, Hector; Wong, Lucy; Villacorta, Virgilio This report contains the following sections: Graduate Student Income and Expenses; Housing, Safety, and Transportation; New Groups and Projects; Other Initiatives and Programs; Collaborations; Historic Events and Activities;

Report to the President for year ended June 30, 2004, Chancellor

2025-11-22T03:17:25Z

Report to the President for year ended June 30, 2004, Chancellor Clay, Phillip This report contains the following sections: Highlights; Other Areas; The Cambridge–MIT Institute; The MIT–Ford Alliance; Faculty Vote on Reserve Officers’ Training Corps;

Report to the President for year ended June 30, 2004, George R. Wallace, Jr., Astrophysical Observatory

2025-11-22T03:17:15Z

Report to the President for year ended June 30, 2004, George R. Wallace, Jr., Astrophysical Observatory Elliot, James L. This report contains the following sections: Facilities; Research and Student Work;

Report to the President for year ended June 30, 2004, Department of Chemistry

2025-11-22T03:18:00Z

Report to the President for year ended June 30, 2004, Department of Chemistry Lippard, Stephen J. This report contains the following sections: Major Faculty Awards and Honors; Infrastructure Developments; Education; Graduate Student Awards and Honors; Named Lectureships; Selected Research Highlights;

Report to the President for year ended June 30, 2004, Dean, Department of Biology

2025-11-22T03:16:56Z

Report to the President for year ended June 30, 2004, Dean, Department of Biology Kaiser, Chris A. This report contains the following sections: Educational Activities; Student Awards; Biology Department Awards; Degrees; Research; Personnel; Faculty Awards;

Report to the President for year ended June 30, 2004, Dean, School of Science

2025-11-22T03:17:59Z

Report to the President for year ended June 30, 2004, Dean, School of Science Silbey, Robert J. This report contains the following sections: New Initiatives; Building and Strengthening a Diverse Community; Faculty Awards; Staff Awards; Academic Program Statistics; Fundraising; Research Volume

Report to the President for year ended June 30, 2004, Program in Women's Studies

2025-11-22T03:18:14Z

Report to the President for year ended June 30, 2004, Program in Women's Studies Wood, Elizabeth This report contains the following sections: Program Administration, Curriculum and Faculty Development, Programming Highlights 2003–2004, Research, Publications, and Service; Affirmative Action Goals and Successes; Future Plans

Report to the President for year ended June 30, 2004, Literature

2025-11-22T03:17:16Z

Report to the President for year ended June 30, 2004, Literature Donaldson, Peter S. This report contains the following sections: Highlights of the Year; Academic Program and Student Enrollment; Research and Publication; Conferences and Invited Addresses; Electronic Projects and Sponsored Research; Service and Committees;

Report to the President for year ended June 30, 2004, History

2025-11-22T03:17:26Z

Report to the President for year ended June 30, 2004, History Ritvo, Harriet

Report to the President for year ended June 30, 2004, Foreign Languages and Literatures

2025-11-22T03:17:19Z

Report to the President for year ended June 30, 2004, Foreign Languages and Literatures Garrels, Elizabeth J. This report contains the following sections: Highlights of the Year; Research and Publications; Conferences and Presentations; MIT Service and Enrollments;

Report to the President for year ended June 30, 2004, Terrascope

2025-11-22T03:17:58Z

Report to the President for year ended June 30, 2004, Terrascope Hodges, Kip; Chisholm, Penny This report contains the following sections: Program Highlights; New Developments;

Report to the President for year ended June 30, 2004, Program in Polymer Science and Technology

2025-11-22T03:17:02Z

Report to the President for year ended June 30, 2004, Program in Polymer Science and Technology McKinley, Gareth H.

Report to the President for year ended June 30, 2004, Lemelson–MIT Program

2025-11-22T03:17:41Z

Report to the President for year ended June 30, 2004, Lemelson–MIT Program Finn, Kristin This report contains the following sections: The Invention Study: Workshops and Assembly; Annual Invention Awards; Outreach Activities and Events;

Report to the President for year ended June 30, 2004, Deshpande Center for Technological Innovation

2025-11-22T03:17:24Z

Report to the President for year ended June 30, 2004, Deshpande Center for Technological Innovation Holly, Krisztina This report contains the following sections: Highlights; Deshpande Grant Awards; Innovation Grants; Catalyst Program; Deshpande Center Events; Administrative Changes

Report to the President for year ended June 30, 2004, Department of Nuclear Engineering

2025-11-22T03:17:20Z

Report to the President for year ended June 30, 2004, Department of Nuclear Engineering Hutchinson, Ian H. This report contains the following sections: Undergraduate Program; Graduate Program; Faculty Awards, Honors, and Activities; Research; Student Awards and Activities;

Report to the President for year ended June 30, 2004, Department of Chemical Engineering

2025-11-22T03:17:12Z

Report to the President for year ended June 30, 2004, Department of Chemical Engineering Armstrong, Robert C.; Rutledge, Gregory C. This report contains the following sections: Undergraduate Education; Graduate Education; Faculty Notes; Research Highlights; Annual Lectures, Seminars, and Symposium; Departmental Awards;

Report to the President for year ended June 30, 2004, Department of Aeronautics and Astronautics

2025-11-22T03:18:19Z

Report to the President for year ended June 30, 2004, Department of Aeronautics and Astronautics Harris, Wesley This report contains the following sections: Undergraduate Awards; Faculty Awards; Staff Awards; Communication for Challenging Environments; Complex Systems Research Lab; Embedded Systems Laboratory; System-on-Chip Design Approaches; Verification and Validation; Operating System Design; Man-Vehicle Laboratory; Massachusetts Space Grant Consortium; Space Systems Laboratory; Systems Analysis Tools (Professor David Miller); Generalized Information Network Analysis; Dynamics, Optics, Controls, and Structures; Uncertainty Propagation in System Modeling; Spaceflight Dynamics and Control Technologies;

Report to the President for year ended June 30, 2004, Department of Urban Studies and Planning

2025-11-22T03:17:32Z

Report to the President for year ended June 30, 2004, Department of Urban Studies and Planning Vale, Lawrence J. This report contains the following sections: Progress on Departmental Priorities; Faculty Achievements; DUSP’s Contribution to MIT–Wide Efforts; Research and Teaching on Urban Planning; City Design and Development; Environmental Policy Group; Housing Community and Economic Development; International Development and Regional Planning; Graduate Degree Program Enrollment and Activities; Undergraduate Program Activities; Awards; Outreach to Alumni; International Connection; Community Partnerships;

Report to the President for year ended June 30, 2004, Program in Media Arts and Sciences

2025-11-22T03:18:15Z

Report to the President for year ended June 30, 2004, Program in Media Arts and Sciences Mitchell, William J. This report contains the following sections: Education; Faculty and Staff; Students

Report to the President for year ended June 30, 2004, Department of Architecture

2025-11-22T03:17:42Z

Report to the President for year ended June 30, 2004, Department of Architecture Anderson, Stanford This report contains the following sections: Architectural Design; Building Technology; History, Theory, and Criticism; Computation; Undergraduate Program; Aga Khan Program for Islamic Architecture; Department of Architecture Enrollments;

Report to the President for year ended June 30, 2004, Museum Loan Network

2025-11-22T03:17:09Z

Report to the President for year ended June 30, 2004, Museum Loan Network Gross, Lori This report contains the following sections: Program Development; Website; Press and Promotion; Grants; Future Plans; Personnel Changes

Report to the President for year ended June 30, 2004, Francis Bitter Magnet Laboratory

2025-11-22T03:18:17Z

Report to the President for year ended June 30, 2004, Francis Bitter Magnet Laboratory Griffin, Robert G. This report contains the following sections: Research Activities; Facilities; Education and Personnel; Future Plans;

Report to the President for year ended June 30, 2004, Broad Institute

2025-11-22T03:17:34Z

Report to the President for year ended June 30, 2004, Broad Institute Lander, Eric S. This report contains the following sections: Mission; Research; Scientific Programs; Faculty; Core Members; Associate Members; Facility

Report to the President for year ended June 30, 2004, Center for Biomedical Engineering

2025-11-22T03:16:59Z

Report to the President for year ended June 30, 2004, Center for Biomedical Engineering Grodzinsky, Alan J. This report contains the following sections: Major Research Thrust Areas; Core Facilities in 500 Technology Square; Major New Initiatives;

Report to the President for year ended June 30, 2004, MIT OpenCourseWare

2025-11-22T03:17:38Z

Report to the President for year ended June 30, 2004, MIT OpenCourseWare Margulies, Anne This report contains the following sections: Achievements; Publication Process; Organization; Technology; Communications; Evaluation; Awards; Finances and Funding; Personnel Information

Report to the President for year ended June 30, 2004, Ombuds Office

2025-11-22T03:18:00Z

Report to the President for year ended June 30, 2004, Ombuds Office Robinson, Toni P.; Rowe, Mary P.

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Women Students by Course and Year

2025-11-22T03:18:22Z

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Women Students by Course and Year MIT Registrar's Office

Report to the President for year ended June 30, 2004, Treasurer of the Corporation

2025-11-22T03:18:01Z

Report to the President for year ended June 30, 2004, Treasurer of the Corporation Bufferd, Allan S.; Stone, Theresa M. This report contains the following sections: Investment Committee

Report to the President for year ended June 30, 2004, Finance: Office of the Controller

2025-11-22T03:17:28Z

Report to the President for year ended June 30, 2004, Finance: Office of the Controller Morgan, James L. This report contains the following sections: Highlights; Accounts Payable; Institute and Enterprise Reporting; Insurance Office; Property Office; Lincoln Fiscal Office; Business Process Changes; Buy/Pay; Payroll; Travel; Institute and Enterprise Reporting; Cashier; Representative Metrics;

Report to the President for year ended June 30, 2004, Finance: Office of Budget and Financial Planning

2025-11-22T03:17:45Z

Report to the President for year ended June 30, 2004, Finance: Office of Budget and Financial Planning Ruiz, Israel; Warner, Margaret This report contains the following sections: Current Goals and Objectives; SAPBUD: Budgeting in SAP; Business and Organizational Modeling; Stochastic Financial Planning; Capital Planning and Budgeting Model Development; Accomplishment; Administration Initiatives; Future Plans; Personnel Information

Report to the President for year ended June 30, 2004, Finance: Audit Division

2025-11-22T03:17:37Z

Report to the President for year ended June 30, 2004, Finance: Audit Division Fisher, Deborah L.

Report to the President for year ended June 30, 2004, Executive Vice President

2025-11-22T03:18:24Z

Report to the President for year ended June 30, 2004, Executive Vice President Curry, John R.

Report to the President for year ended June 30, 2004, Enterprise Services

2025-11-22T03:17:26Z

Report to the President for year ended June 30, 2004, Enterprise Services Graham, Louis W., Jr.; Walsh, Phillip J.; Berlin, Richard D.; Dimond, Steven M.; Fitzgerald, Michael; Michaud, Daniel; Brutti, Larry; Dimond, Steven M. This report contains the following sections: Audio Visual Services; Campus Activities Complex; Campus Dining; Copy Technology Centers; MIT Endicott House; MIT Card Office; Parking and Transportation; TechCASH

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of International Students by Course and Year

2025-11-22T03:18:16Z

Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of International Students by Course and Year MIT Registrar's Office

Report to the President for year ended June 30, Appendix B: Enrollment Statistics, Fall 2003, Number of Students by Course and Year

2025-11-22T03:17:56Z

Report to the President for year ended June 30, Appendix B: Enrollment Statistics, Fall 2003, Number of Students by Course and Year; Report to the President for year ended June 30, 2004, Appendix B: Enrollment Statistics, Fall 2003, Number of Students by Course and Year MIT Registrar's Office

Report to the President for year ended June 30, 2004, Appendix A: Degrees Awarded 2003-2004

2025-11-22T03:17:06Z

Report to the President for year ended June 30, 2004, Appendix A: Degrees Awarded 2003-2004 MIT Registrar's Office

Report to the President for year ended June 30, 2004, Operations: MIT Police

2025-11-22T03:18:25Z

Report to the President for year ended June 30, 2004, Operations: MIT Police DiFava, John This report contains the following sections: Patrol Division; Professional Standards; Community Policing; Criminal Activity; Devpartment Highlights

Report to the President for year ended June 30, 2004, Office of Student Discipline

2025-11-22T03:17:00Z

Report to the President for year ended June 30, 2004, Office of Student Discipline Tyrell, Steven This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Summary of Staffing Changes

Report to the President for year ended June 30, 2004, Office of Academic Services

2025-11-22T03:18:12Z

Report to the President for year ended June 30, 2004, Office of Academic Services Vandiver, J. Kim This report contains the following sections: Academic Information and Communication; Academic Resource Center; Faculty and Alumni Support; Staffing Changes

Report to the President for year ended June 30, 2004, Cambridge–MIT Institute

2025-11-22T03:18:04Z

Report to the President for year ended June 30, 2004, Cambridge–MIT Institute Crawley, Ed This report contains the following sections: Noteworthy Events; Educational Programs; Research Programs and Industry; Special Interest Groups; Knowledge Exchange Activities; Future Plans;

Report to the President for year ended June 30, 2004, Operations: Facilities

2025-11-22T03:17:40Z

Report to the President for year ended June 30, 2004, Operations: Facilities Sirianni, Victoria This report contains the following sections: Infrastructure Renewal; Shared Services Center; Client Orientation; Collaboration; Sustainability; Accountability; Professionalism; Capital Projects; Design and Construction Services; Continuation of Project Controls and Information Tracking Efforts; Finance and Accounting; Operations; Utilities; Administration; Information Technology; Personnel Changes;

Report to the President for year ended June 30, 2004, Community Development and Substance Abuse Programs

2025-11-22T03:17:39Z

Report to the President for year ended June 30, 2004, Community Development and Substance Abuse Programs Trujillo, Daniel A. This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Summary of Staffing Changes

Report to the President for year ended June 30, 2004, Experimental Study Group

2025-11-22T03:17:04Z

Report to the President for year ended June 30, 2004, Experimental Study Group Slocum, Alexander; Dourmashkin, Peter; Sweet, Holly This report contains the following sections: Student Statistics; Staff and Faculty; Academic Initiatives; Awards; Alumni Involvement; Future Developments;

Report to the President for year ended June 30, 2004, Department of Physics

2025-11-22T03:17:12Z

Report to the President for year ended June 30, 2004, Department of Physics Kastner, Marc This report contains the following sections: Honors and Awards; Education; Diversity; Pappalardo Fellowships in Physics; Research Highlights;

Report to the President for year ended June 30, 2004, Center for International Studies

2025-11-22T03:18:21Z

Report to the President for year ended June 30, 2004, Center for International Studies Samuels, Richard J.; Van Evera, Stephen; Makinson, Carolyn This report contains the following sections: MIT Security Studies Program; MIT International Science and Technology Initiative; OpenCourseWare; MIT Mexico Program; HASS Minor in Applied International Studies; Program on Human Rights and Justice; The Inter-University Committee on International Migration; The Inter-University Initiative on Humanitarian Studies and Field Practice; Political Economy and Technology Policy Program; Crosscutting Working Groups; Seminar XXI—Outreach to the Washington Policy Community; Public Programs; Seminars, Colloquia, Workshops, and Conferences; Grant Programs; Publications; Personnel; CIS Affirmative Action Goals and Successes;

Report to the President for year ended June 30, 2004, Singapore–MIT Alliance

2025-11-22T03:17:33Z

Report to the President for year ended June 30, 2004, Singapore–MIT Alliance Patera, Anthony T. This report contains the following sections: Partnership; Management Structure; Summer Conference; Distance Learning; Entering Class; Noteworthy Events in 2003; Innovation in Manufacturing Systems and Technology; Molecular Engineering of Biological and Chemical Systems; Computer Science; Benefits and Goals

Report to the President for year ended June 30, 2004, Materials Processing Center/Microphotonics Center

2025-11-22T03:17:08Z

Report to the President for year ended June 30, 2004, Materials Processing Center/Microphotonics Center Lippegrenfell, Tamarleigh This report contains the following sections: Relationships; Activities; Equipment and Facilities; Outlook;

Report to the President for year ended June 30, 2004, Laboratory for Information and Decision Systems

2025-11-22T03:18:01Z

Report to the President for year ended June 30, 2004, Laboratory for Information and Decision Systems Chan, Vincent W. S. This report contains the following sections: Highlights; Faculty; Students; Research Overview; Research Areas;

Report to the President for year ended June 30, 2004, Concourse Program

2025-11-22T03:18:16Z

Report to the President for year ended June 30, 2004, Concourse Program Rose, Robert M. This report contains the following sections: Personnel Information; Enrollment; Teaching and Curriculum

Report to the President for year ended June 30, 2004, Biotechnology Process Engineering Center

2025-11-22T03:17:46Z

Report to the President for year ended June 30, 2004, Biotechnology Process Engineering Center Griffith, Linda This report contains the following sections: Goals, Objectives, and Priorities; Accomplishments; Administrative Initiatives; Finances and Funding; Future Plans; Personnel;

Report to the President for year ended June 30, 2004, Department of Electrical Engineering and Computer Science

2025-11-22T03:18:19Z

Report to the President for year ended June 30, 2004, Department of Electrical Engineering and Computer Science Guttag, John V. This report contains the following sections: Graduate Program; Undergraduate Program; 6-A Internship Program; Faculty Notes; Faculty Awards and Honors

Report to the President for year ended June 30, 2004, Media Laboratory

2025-11-22T03:18:06Z

Report to the President for year ended June 30, 2004, Media Laboratory Bender, Walter This report contains the following sections: Research Achievements; Exhibitions and Performances; Collaboration within MIT; Media Lab Europe; Sponsors; Human Resources/Administration;

Report to the President for year ended June 30, 2004, Department of Athletics, Physical Education, and Recreation

2025-11-22T03:17:02Z

Report to the President for year ended June 30, 2004, Department of Athletics, Physical Education, and Recreation Royer, Candace L. This report contains the following sections: Summary Statement; Highlights of the Year; New Initiatives; Summary of Staffing Changes

Report to the President for year ended June 30, 2004, Student Services Information Technology

2025-11-22T03:17:51Z

Report to the President for year ended June 30, 2004, Student Services Information Technology Stevenson, JoAnne This report contains the following sections: Accomplishments; Staffing Changes

Report to the President for year ended June 30, 2004, MIT Careers Office

2025-11-22T03:17:35Z

Report to the President for year ended June 30, 2004, MIT Careers Office Reed, Elizabeth This report contains the following sections: Accomplishments during FY2004; Staffing Changes;

Report to the President for year ended June 30, 2004, Director, Libraries

2025-11-22T03:17:58Z

Report to the President for year ended June 30, 2004, Director, Libraries Wolpert, Ann J.; Gass, Steve; Fleishauer, Carol; Glavash, Keith; Smith, MacKenzie This report contains the following sections: Director, Libraries; Public Services; Collection Services; Administrative Services; Technology Planning and Administration

Report to the President for year ended June 30, 2004, Center for Space Research

2025-11-22T03:16:53Z

Report to the President for year ended June 30, 2004, Center for Space Research Hewitt, Jacqueline N. This report contains the following sections: Research Highlights; Education and Public Outreach

Report to the President for year ended June 30, 2004, Laboratory for Nuclear Science

2025-11-22T03:17:44Z

Report to the President for year ended June 30, 2004, Laboratory for Nuclear Science Matthews, June L. This report contains the following sections: Experimental High-Energy Physics; Experimental Nuclear Physics; Theoretical Nuclear and Particle Physics; Education

Report to the President for year ended June 30, 2004, George R. Harrison Spectroscopy Laboratory

2025-11-22T03:17:04Z

Report to the President for year ended June 30, 2004, George R. Harrison Spectroscopy Laboratory Feld, Michael S. This report contains the following sections: Research Highlights;

Report to the President for year ended June 30, 2004, Department of Mathematics

2025-11-22T03:16:58Z

Report to the President for year ended June 30, 2004, Department of Mathematics Vogan, David A., Jr. This report contains the following sections: Students; Faculty Changes; Administration; Research; Honors, Prizes, and Awards; Education;

Report to the President for year ended June 30, 2004, Department of Earth, Atmospheric, and Planetary Sciences

2025-11-22T03:17:54Z

Report to the President for year ended June 30, 2004, Department of Earth, Atmospheric, and Planetary Sciences Zuber, Maria T. This report contains the following sections: Educational Activities; Faculty; Current Research;

Report to the President for year ended June 30, 2004, Program in Science, Technology, and Society

2025-11-22T03:17:53Z

Report to the President for year ended June 30, 2004, Program in Science, Technology, and Society Williams, Rosalind H. This report contains the following sections: Doctoral Program; Projects, Grants, and Initiatives; Educational Activities; Ongoing Activities of the Program; Knight Science Journalism Fellowship Program; Faculty Activities;

Report to the President for year ended June 30, 2004, Special Projects Office

2025-11-22T03:17:25Z

Report to the President for year ended June 30, 2004, Special Projects Office Enders, Margaret S. This report contains the following sections: Administration of the MIT Communication Requirement; The Cambridge-MIT Undergraduate Student Exchange Program; Planning for an Office of Study Abroad and Foreign Scholarships; The MacVicar Faculty Fellows Program; The d’Arbeloff Grants Program; Support to the Task Force on the Undergraduate Educational Commons; Staffing Changes

Report to the President for year ended June 30, 2004, Office of Minority Education

2025-11-22T03:17:16Z

Report to the President for year ended June 30, 2004, Office of Minority Education Beamon, Kim This report contains the following sections: Project Interphase; Seminar XL; Second Summer Program; Tutorial Service Room; Industrial Advisory Council for Minority Education; Office of Minority Education Student Advisory Council; Minority Scholarships; Minority Awards Banquet;

Report to the President for year ended June 30, 2004, Admissions Office

2025-11-22T03:16:51Z

Report to the President for year ended June 30, 2004, Admissions Office Jones, Marilee This report contains the following sections: Accomplishments; Staffing Changes

Report to the President for year ended June 30, 2004, International Students Office

2025-11-22T03:17:51Z

Report to the President for year ended June 30, 2004, International Students Office Guichard‐Ashbrook, Danielle This report contains the following sections: International Admissions; International Student Advising; Orientation Programs for International Students; Host to International Students Program; Future Goals;

Report to the President for year ended June 30, 2004, MIT Press

2025-11-22T03:18:13Z

Report to the President for year ended June 30, 2004, MIT Press Faran, Ellen W. This report contains the following sections: FY2004 Highlights; FY2004 Financial Results; MIT Press Management Board, 2003–2004; MIT Press Editorial Board, 2003–2004; MIT Press Acquisitions Editors; Books Division; Production Department; Journals Division; MIT Faculty Journal Editors; MIT Press Bookstore

Report to the President for year ended June 30, 2004, Picower Center for Learning and Memory

2025-11-22T03:17:50Z

Report to the President for year ended June 30, 2004, Picower Center for Learning and Memory Tonegawa, Susumu This report contains the following sections: Major Research Breakthroughs; New Building and Faculty Hiring; Public Relations; Promotions; Awards; Research Highlights;

Report to the President for year ended June 30, 2004, Center for Cancer Research

2025-11-22T03:17:34Z

Report to the President for year ended June 30, 2004, Center for Cancer Research Jacks, Tyler This report contains the following sections: Animal Models of Cancer; Stem Cells, Development, and Cancer; RNAi Technology; Integrative Analysis of Cancer Pathways; Faculty Awards

Report to the President for year ended June 30, 2004, Program in Writing and Humanistic Studies

2025-11-22T03:18:07Z

Report to the President for year ended June 30, 2004, Program in Writing and Humanistic Studies Paradis, James This report contains the following sections: Research and Publications; Academic Programs and Initiatives; Service, Grants, and Awards; Personnel;

Report to the President for year ended June 30, 2004, Music and Theater Arts

2025-11-22T03:17:55Z

Report to the President for year ended June 30, 2004, Music and Theater Arts Ziporyn, Evan This report contains the following sections: Highlights of the Year; Honors and Awards; Program Highlights; Achievements; Personnel;

Report to the President for year ended June 30, 2004, Office of Professional Education Programs

2025-11-22T03:17:44Z

Report to the President for year ended June 30, 2004, Office of Professional Education Programs Stine, Jennifer This report contains the following sections: Current Goals, Objectives, Priorities; Accomplishments and Program Developments; Funding; Future Plans; Personnel Information

Report to the President for year ended June 30, 2004, Anthropology

2025-11-22T03:18:29Z

Report to the President for year ended June 30, 2004, Anthropology Jackson, Jean This report contains the following sections: Personnel and Administrative Changes; Program Contributions to MIT and Outside Communities; Educational Activities; Presentations; Publications; Other Program Accomplishments; Grants, Honors, and Awards

Report to the President for year ended June 30, 2004, Laboratory for Manufacturing and Productivity

2025-11-22T03:18:27Z

Report to the President for year ended June 30, 2004, Laboratory for Manufacturing and Productivity Gutowski, Timothy G. This report contains the following sections: Research and Education Highlights, Awards;

Report to the President for year ended June 30, 2004, Industrial Performance Center

2025-11-22T03:17:52Z

Report to the President for year ended June 30, 2004, Industrial Performance Center Lester, Richard K. This report contains the following sections: Research Highlights; People;

Report to the President for year ended June 30, 2004, Department of Materials Science and Engineering

2025-11-22T03:17:31Z

Report to the President for year ended June 30, 2004, Department of Materials Science and Engineering Suresh, Subra This report contains the following sections: Research Initiatives; Undergraduate Education; Graduate Education; Master of Engineering in Materials; Other Educational Initiatives; Student Organizations; Personnel; Research Highlights; Awards and Honors; AY2004 Undergraduate Awards; AY2004 Graduate Awards; Faculty Notes; Future Plans;

Report to the President for year ended June 30, 2004, Engineering Systems Division

2025-11-22T03:17:11Z

Report to the President for year ended June 30, 2004, Engineering Systems Division Hastings, Daniel; Allen, Tom; Hanson, Bill; Simchi‐Levi, David; Newman, Dava; Nordal, Nils; Moavenzadeh, Fred; Clay, Phillip L.; Heywood, John B.; Cusumano, Michael A.; MacDuffie, John Paul; Cutcher‐Gershenfeld, Joel; Kochan, Thomas A.; Nightingale, Deborah; Carroll, John; Bryan, Frederick “Terry”; Harris, Wesley L.; Roth, Richard; Ashford, Nicholas; Sheffi, Yossi This report contains the following sections: Ongoing Initiatives; Faculty Notes; Student Honors; Program Honors; INCOSE; Conference on Systems Engineering Research; ESD Administrative Staff; Major Meetings; Leaders for Manufacturing; System Design and Management; Technology and Policy Program; Technology, Management, and Policy Program; Center for Innovation in Product Development; Center for Technology, Policy, and Industrial Development; Ford–MIT Alliance; International Motor Vehicle Program; Labor Aerospace Research Agenda; Lean Aerospace Initiative; Lean Sustainment Initiative; Materials Systems Laboratory; MIT Information Quality Program; Technology and Law Program; MIT Center for Transportation and Logistics;

Report to the President for year ended June 30, 2004, Department of Civil and Environmental Engineering

2025-11-22T03:17:45Z

Report to the President for year ended June 30, 2004, Department of Civil and Environmental Engineering Jaillet, Patrick This report contains the following sections: Initiatives; Educational Activities; Undergraduate Program; Graduate Programs; Faculty Notes; Student Notes; Departmental Awards;

Report to the President for year ended June 30, 2004, Dean, School of Engineering

2025-11-22T03:18:05Z

Report to the President for year ended June 30, 2004, Dean, School of Engineering Magnanti, Thomas L. This report contains the following sections: Continuing Initiatives; Emerging Technologies; Educational Innovation and Diversity; New Initiative: Electronic Outreach to Alumni; Notable Events; Organizational Reviews and Changes; Personnel; Awards; Statistics for AY2004;

Report to the President for year ended June 30, 2004, Associate Provost for the Arts

2025-11-22T03:16:58Z

Report to the President for year ended June 30, 2004, Associate Provost for the Arts Brody, Alan This report contains the following sections: Resources and Programs; Laboratory for the Performing Arts; Budget Crunch; Student Art Association and Wiesner Gallery; Council for the Arts and the McDermott Award; Office of Arts Communication; Museum Loan Network; MIT Museum; List Visual Arts Center;

Report to the President for year ended June 30, 2004, Council on Primary and Secondary Education

2025-11-22T03:17:30Z

Report to the President for year ended June 30, 2004, Council on Primary and Secondary Education Latanision, R. M. This report contains the following sections: MIT/Wellesley Teacher Education Program; Teacher Sabbaticals; Educational Program Outreach Directory; Programs by the CPSE Chairman; Association of American Universities Task Force on K–16 Education; Science and Engineering Program for Teachers;

Report to the President for year ended June 30, 2004, MIT Washington Office

2025-11-22T03:17:17Z

Report to the President for year ended June 30, 2004, MIT Washington Office Crowley, John C. This report contains the following sections: Mission; Advocacy Coalitions and Working Groups; Legislative Initiatives; MIT Lincoln Laboratory, Center for Fusion and Plasma Science, Bates Laboratory; MIT Congressional Staff Seminar on Science and Technology; Executive Branch

Report to the President for year ended June 30, 2004, Dean, School of Architecture and Planning

2025-11-22T03:16:55Z

Report to the President for year ended June 30, 2004, Dean, School of Architecture and Planning Santos, Adèle Naudé; Knight, Terry This report contains the following sections: Faculty; Space; Educational Initiatives; Events and Awards; Goals for 2005;

Report to the President for year ended June 30, 2004, Office of the Arts

2025-11-22T03:18:12Z

Report to the President for year ended June 30, 2004, Office of the Arts Cohen, Susan R.; Haller, Mary L.; Billingsley, Glenn; Oshima, Michèle This report contains the following sections: Council for the Arts; Arts Communication; Office of the Arts Development; Student & Artist-in-Residence Programs;

Report to the President for year ended June 30, 2004, MIT Museum

2025-11-22T03:18:26Z

Report to the President for year ended June 30, 2004, MIT Museum Curtis, Jack; Douglas, Debbie; Hunt, Stephanie; Hasselbalch, Kurt; Leen, Mary; OʹNeill, Jenny; Rosenthal, Beryl; Van Zante, Gary; Whitlow, Joan This report contains the following sections: Collections; Education and Outreach; Exhibitions; Administration;

Report to the President for year ended June 30, 2004, Lincoln Laboratory

2025-11-22T03:17:23Z

Report to the President for year ended June 30, 2004, Lincoln Laboratory Briggs, David L. This report contains the following sections: Laboratory Operations; Technical Program Highlights;

Report to the President for year ended June 30, 2004, Associate Provost

2025-11-22T03:16:54Z

Report to the President for year ended June 30, 2004, Associate Provost Canizares, Claude R.

Report to the President for year ended June 30, 2004, Technology and Development Program

2025-11-22T03:18:24Z

Report to the President for year ended June 30, 2004, Technology and Development Program Moavenzadeh, Fred This report contains the following sections: Current Research Programs; Future Research Initiatives; Current Educational Initiatives; Organization

Report to the President for year ended June 30, 2004, Haystack Observatory

2025-11-22T03:18:27Z

Report to the President for year ended June 30, 2004, Haystack Observatory Salah, Joseph E. This report contains the following sections: Instrumentation; Radio Astronomy; Instrumentation Development; Atmospheric Science; Educational Programs;

Report to the President for year ended June 30, 2004, Computational and Systems Biology Initiative

2025-11-22T03:17:07Z

Report to the President for year ended June 30, 2004, Computational and Systems Biology Initiative Sorger, Peter; Tadmor, Brigitta This report contains the following sections: Goals and Priorities; Education and Training; Research; Technology Development; Junior Faculty Startup; Outreach; Leadership; Finances and Funding

Report to the President for year ended June 30, 2004, Office of Educational Opportunity Programs

2025-11-22T03:18:02Z

Report to the President for year ended June 30, 2004, Office of Educational Opportunity Programs Crichlow, Ronald S.; Layne, Evette M. This report contains the following sections: MIT/Wellesley Upward Bound

Report to the President for year ended June 30, 2004, Center for Archaeological Materials / Center for Materials Research in Archaeology and Ethnology

2025-11-22T03:16:57Z

Report to the President for year ended June 30, 2004, Center for Archaeological Materials / Center for Materials Research in Archaeology and Ethnology Lechtman, Heather

Report to the President for year ended June 30, 2004, Finance: Office of Sponsored Programs

2025-11-22T03:17:14Z

Report to the President for year ended June 30, 2004, Finance: Office of Sponsored Programs Norris, Julie This report contains the following sections: Research Volume; Compliance Issues; Costing Issues; Negotiation of Rates; Other Costing Activities; Export Control Laws and Related Issues; Administrative Theme Initiatives;

Report to the President for year ended June 30, 2004, Teaching and Learning Laboratory

2025-11-22T03:17:57Z

Report to the President for year ended June 30, 2004, Teaching and Learning Laboratory Breslow, Lori This report contains the following sections: Instructional Support; Assessment and Evaluation; Research and Scholarship; Staff Changes;

Report to the President for year ended June 30, 2004, Student Financial Services

2025-11-22T03:16:52Z

Report to the President for year ended June 30, 2004, Student Financial Services Hicks, Betsy This report contains the following sections: Operating Activities; Student Receivables; Undergraduate Financial Aid; Undergraduate Parent Loans; Graduate Need-based Financial Aid; Accomplishments; Staffing;

Report to the President for year ended June 30, 2004, Office of the Registrar

2025-11-22T03:17:00Z

Report to the President for year ended June 30, 2004, Office of the Registrar Callahan, Mary This report contains the following sections: Accomplishments; Operational Highlights; Classroom Management Highlights; Registration; Degrees Awarded; Personnel Changes;

Report to the President for year ended June 30, 2004, Dean for Graduate Students

2025-11-22T03:18:03Z

Report to the President for year ended June 30, 2004, Dean for Graduate Students Colbert, Isaac M.; Staton, Blanche; Wurie, Brima; Charles, Roy A. This report contains the following sections: Graduate Students Office; Indicators of Change; Steps Forward; Strategic Collaborations; Renewed Commitment to Graduate Student Diversity; Graduate Fellowships; Programs and Services;

Report to the President for year ended June 30, 2004, Earth System Initiative

2025-11-22T03:17:05Z

Report to the President for year ended June 30, 2004, Earth System Initiative Chisholm, Penny; Hodges, Kip This report contains the following sections: Research; Education and Outreach; Personnel;

Report to the President for year ended June 30, 2004, Department of Brain and Cognitive Sciences

2025-11-22T03:18:05Z

Report to the President for year ended June 30, 2004, Department of Brain and Cognitive Sciences Sur, Mriganka This report contains the following sections: Education; Faculty Highlights; Research Advances; Learning and Memory; Brain Development and Plasticity; Language and Cognition; New Undertakings and Ongoing Events

Report to the President for year ended June 30, 2004, Department of Political Science

2025-11-22T03:18:25Z

Report to the President for year ended June 30, 2004, Department of Political Science Cohen, Joshua This report contains the following sections: Significant Events 2003–2004; Student Recruitment, Placement, and Enrollment; Faculty/Personnel; Promotions/Personnel Activity in AY2004 and Upcoming Faculty Searches; Faculty Leaves, Departures, Upcoming Searches, and Visitors; Faculty Research and Publications;

Report to the President for year ended June 30, 2004, Department of Linguistics and Philosophy

2025-11-22T03:17:03Z

Report to the President for year ended June 30, 2004, Department of Linguistics and Philosophy Marantz, Alec This report contains the following sections: Research: Linguistics; Research: Philosophy; Publications; Honors and Awards; Leaves of Absence; Personnel

Report to the President for year ended June 30, 2004, Department of Economics

2025-11-22T03:17:29Z

Report to the President for year ended June 30, 2004, Department of Economics Holmström, Bengt This report contains the following sections: Highlights of the Year; Future Plans; Personnel; Honors and Awards; Research Achievements;

Report to the President for year ended June 30, 2004, Microsystems Technology Laboratories

2025-11-22T03:16:57Z

Report to the President for year ended June 30, 2004, Microsystems Technology Laboratories Schmidt, Martin A. This report contains the following sections: Highlights; Future Plans;

Report to the President for year ended June 30, 2004, Biological Engineering Division

2025-11-22T03:17:05Z

Report to the President for year ended June 30, 2004, Biological Engineering Division Lauffenburger, Douglas A. This report contains the following sections: Undergraduate Education; Graduate Education;

Report to the President for year ended June 30, 2004, Center for Advanced Visual Studies

2025-11-22T03:17:28Z

Report to the President for year ended June 30, 2004, Center for Advanced Visual Studies Wodiczko, Krzysztof This report contains the following sections: Activities

Report to the President for year ended June 30, 2004, Technology Licensing Office

2025-11-22T03:16:54Z

Report to the President for year ended June 30, 2004, Technology Licensing Office Nelsen, Lita

Report to the President for year ended June 30, 2004, Sea Grant College Program

2025-11-22T03:17:29Z

Report to the President for year ended June 30, 2004, Sea Grant College Program Chryssostomidis, Chryssostomos This report contains the following sections: Education; Graduate Student Research Assistants; Undergraduate Research Opportunities Program; K-12 Education; New Core Research Projects; Ongoing Core Research Projects; Advisory Services; Outreach and Communications; Program Management

Report to the President for year ended June 30, 2004, Research Laboratory of Electronics

2025-11-22T03:18:28Z

Report to the President for year ended June 30, 2004, Research Laboratory of Electronics Shapiro, Jeffrey H. This report contains the following sections: Circuits, Systems, Signals and Communications; Physical Sciences; Quantum Computation and Communication; Nanostructures; Photonic Materials, Devices, and Systems; Communication Biophysics; RLE Conference Facility; Appointments, Awards, and Events; Affirmative Action

Report to the President for year ended June 30, 2004, Nuclear Reactor Laboratory

2025-11-22T03:18:08Z

Report to the President for year ended June 30, 2004, Nuclear Reactor Laboratory Moncton, David E. This report contains the following sections: MIT Research Reactor; Reactor Administration and Organization; Organizational Diversity; Safety and Security; Relicensing and Redesign; Major Reactor Services; Research Activities;

Report to the President for year ended June 30, 2004, MIT/WHOI Joint Program in Oceanography and Applied Ocean Science and Engineering

2025-11-22T03:17:41Z

Report to the President for year ended June 30, 2004, MIT/WHOI Joint Program in Oceanography and Applied Ocean Science and Engineering Rizzoli, Paola; Schwartz, Ronni This report contains the following sections: New Program in Marine Meteorology; Student Apartment in University Housing; Continuation of Presidential Fellowships; External Review

Report to the President for year ended June 30, 2004, Academic Media Production Services

2025-11-22T03:18:09Z

Report to the President for year ended June 30, 2004, Academic Media Production Services Mitra, Amitava “Babi”; Kumar, M. S. Vijay This report contains the following sections: Vision; Strategic Support; AMPS Advisory Board; Stellar Faculty Advisory Board; Services Offered; Organizing for Service; MIT Video Productions and Digital Technologies; Initiatives to Advance Operational Efficiency; People; Educational Design and Development Group; Changes in Personnel and Funding; Web Tools and Operations; Stellar Course Management System; Financial Operations and Administrative Liaison Unit; Projects; Facilities; Awards, Conferences, and Programs; Future Plans

Report to the President for year ended June 30, 2004, Provost

2025-11-22T03:17:31Z

Report to the President for year ended June 30, 2004, Provost Brown, Robert A. This report contains the following sections: People; Academic Programs; MIT OpenCourseWare; Broad Institute; McGovern Institute for Brain Research; Facilities; Faculty; Graduate Student Fellowships; Finances; Research

Report to the President for year ended June 30, 2004, Vice President and Secretary of the Corporation

2025-11-22T03:17:01Z

Report to the President for year ended June 30, 2004, Vice President and Secretary of the Corporation Willmore, Kathryn A.; Gallagher, Gayle M.; Lisanti, Suzana; Jones, Arthur L.; Lee, Monica; Kiang, Stuart; Lester, Susan A. This report contains the following sections: Public Relations Services; Conference Services, Events, and Information Center; MIT Home Page Team; News Office; Publishing Services Bureau; Reference Publications Office| Office of the Secretary of the Corporation

Report to the President for year ended June 30, 2004, Vice President for Resource Development

2025-11-22T03:18:30Z

Report to the President for year ended June 30, 2004, Vice President for Resource Development Stowe, Barbara G.; Dare, Stephen A.; Eastment, Katherine E.; Serfes, Pamela Dumas; Koster, Karl F.; Scott, Robert D.; Rinaldi, Christine M.; Oldham, John E.; Sager, Judith V.; Miller, Lucy V. This report contains the following sections: Highlights; Corporation and Foundation Giving; Principal Gifts; Stewardship; Volunteer Partnerships; Unrestricted and Core Support; Human Resources; Summary of Private Support; Campaign for MIT; Campaign Giving; Communications and Donor Relations; Corporate Relations; Development Research and Systems; Donor Partnerships and Special Projects; Foundation Relations and Academic Development Support; Gift Planning; Principal Gifts

Report to the President for year ended June 30, 2004, Vice President for Information Services and Technology

2025-11-22T03:18:08Z

Report to the President for year ended June 30, 2004, Vice President for Information Services and Technology Grochow, Jerrold M. This report contains the following sections: Client Orientation; Collaboration; Sustainability; Accountability; Professionalism;

Report to the President for year ended June 30, 2004, Chair of the Faculty

2025-11-22T03:18:18Z

Report to the President for year ended June 30, 2004, Chair of the Faculty Bras, Rafael L.; Burns, Lily U. This report contains the following sections: Faculty Policy Committee; Committee on the Undergraduate Program; Subcommittee on the Communication Requirement; Committee on Academic Performance; Committee on Curricula; Committee on Discipline; Harold E. Edgerton Award Committee; Committee on Faculty–Administration; Killian Award Committee; Committee on the Library System; Committee on Nominations; Committee on Student Life; Student–Faculty Interaction; Student Family Health Care; Committee on Outside Professional Activities; Committee on Undergraduate Admissions and Financial Aid;

Report to the President for year ended June 30, 2004, ROTC Programs

2025-11-22T03:17:36Z

Report to the President for year ended June 30, 2004, ROTC Programs Rojko, Paul; Baker, Brian L.; Holland, Robert D. This report contains the following sections: Air Force Reserve Officers’ Training Corps; Army Reserve Officers’ Training Corps; Naval Reserve Officers Training Corps

Report to the President for year ended June 30, 2004, Dean for Undergraduate Education

2025-11-22T03:17:27Z

Report to the President for year ended June 30, 2004, Dean for Undergraduate Education Redwine, Robert P.

Report to the President for year ended June 30, 2004, Dean, MIT Sloan School of Management

2025-11-22T03:17:36Z

Report to the President for year ended June 30, 2004, Dean, MIT Sloan School of Management Schmalensee, Richard This report contains the following sections: Academic Program Updates; Other Initiatives

Report to the President for year ended June 30, 2004, Comparative Media Studies

2025-11-22T03:17:49Z

Report to the President for year ended June 30, 2004, Comparative Media Studies Jenkins, Henry This report contains the following sections: Research; Fundraising; Governance; Graduate Admissions; Undergraduate Education; Events and Programs; Honors and Awards; Visiting Scholars; Publications

Report to the President for year ended June 30, 2004, Laboratory for Electromagnetic and Electronic Systems

2025-11-22T03:17:15Z

Report to the President for year ended June 30, 2004, Laboratory for Electromagnetic and Electronic Systems Kassakian, John G. This report contains the following sections: Automotive Electrical and Electronic Systems; Modeling, Monitoring, and Control of Power Systems; Power Electronics and Electromechanics; Sensors, Nanotechnology, and Microelectromechanical Systems; Enhanced Ultracapacitor Analysis and Development; From Bioelectromechanics to Biomedicine; Honors and Awards

Dopamine modulation of aggression

2026-03-08T03:26:40Z

Dopamine modulation of aggression Dai, Bing; Lin, Dayu Rationale Aggression is an innate social behavior prevalent across animal species. However, in modern human society, inter-personal aggression is considered disruptive and detrimental to both families and communities. Clinically, antipsychotics, which primarily target dopamine (DA) receptors, have been widely used to suppress hyper-aggression. However, the mechanisms underlying the effect of the antipsychotics remain incompletely understood. Objectives We reviewed key steps in brain DA synthesis and summarized genetic and pharmacological evidence supporting the role of the mesolimbic DA system in aggression. Next, we discussed recent circuit studies that elucidate the DA action in modulating aggression-related brain regions. These lines of evidence collectively suggest that DA acts on different brain regions to facilitate aggression and self-learning, and signals the valence of the fighting experience.

Report to the President for year ended June 30, 2004, Computer Science and Artificial Intelligence Laboratory

2025-11-22T03:17:33Z

Report to the President for year ended June 30, 2004, Computer Science and Artificial Intelligence Laboratory Brooks, Rodney This report contains the following sections: Highlights; World Wide Web Consortium; Distinguished Lecture Series; Awards/Honors; Affirmative Action

Report to the President for year ended June 30, 2004, Department of Ocean Engineering

2025-11-22T03:18:15Z

Report to the President for year ended June 30, 2004, Department of Ocean Engineering Schmidt, Henrik This report contains the following sections: Current Goals and Objectives, Focus, and Priorities; Accomplishments; Administrative Initiatives; Strategic Planning; Future Plans; Personnel Information; Student Awards; Teaching and Curriculum; Current Research Projects;

Report to the President for year ended June 30, 2004, Department of Mechanical Engineering

2025-11-22T03:17:21Z

Report to the President for year ended June 30, 2004, Department of Mechanical Engineering Abeyaratne, Rohan This report contains the following sections: Undergraduate Program; Graduate Program; Faculty Notes;

Report to the President for year ended June 30, 2004, Institute for Soldier Nanotechnologies

2025-11-22T03:17:38Z

Report to the President for year ended June 30, 2004, Institute for Soldier Nanotechnologies Thomas, Edwin L. This report contains the following sections: Research; Soldier Design Competition; Industrial Collaboration; Facilities; Outreach; Appointments, Visitors, and Awards; Future Plans

Report to the President for year ended June 30, 2004, Operations Research Center

2025-11-22T03:18:29Z

Report to the President for year ended June 30, 2004, Operations Research Center Orlin, James B.; Tsitsiklis, John N. This report contains the following sections: Faculty, Students, Staff; Outreach and Professional Service; Operational Issues; Future Plans; Diversity; Professional Activities;

Report to the President for year ended June 30, 2004, Whitaker College of Health Sciences and Technology

2025-11-22T03:17:22Z

Report to the President for year ended June 30, 2004, Whitaker College of Health Sciences and Technology Samson, Leona D.; Dedon, Peter C.; Wurtman, Richard J.; Fox, James G.; Gray, Martha L. This report contains the following sections: Center for Environmental Health Sciences; Research Cores; Core Facilities; Pilot Project Program; Plans for 2005; Clinical Research Center; Administration; Education; Affirmative Action; Research Activities; Center for Experimental Pharmacology and Therapeutics; Computer Facility; Core Laboratory/Mass Spectrometry Facility; Research Highlights; CRC Investigator-Initiated Programs; Division of Comparative Medicine; Facility Management and Animal Care; Research Activities; Academic Activities; Committee on Animal Care Activities; Harvard-MIT Division of Health Sciences and Technology

Report to the President for year ended June 30, 2004, Vice President for Research and Associate Provost

2025-11-22T03:17:19Z

Report to the President for year ended June 30, 2004, Vice President for Research and Associate Provost Gast, Alice P.

Report to the President for year ended June 30, 2004, Office of Government and Community Relations

2025-11-22T03:17:50Z

Report to the President for year ended June 30, 2004, Office of Government and Community Relations Gallop, Sarah E.; Parravano, Paul This report contains the following sections: Local Government Relations; Federal Government Relations; Community Relations; Schools; Business; Community Service Fund

Report to the President for year ended June 30, 2004, Office of the President: In Special Recognition

2025-11-22T03:17:56Z

Report to the President for year ended June 30, 2004, Office of the President: In Special Recognition This report contains the following sections: Honors and Awards; In Memoriam;

Report to the President for year ended June 30, 2004, Humanities, Arts, and Social Sciences Education Office

2025-11-22T03:17:48Z

Report to the President for year ended June 30, 2004, Humanities, Arts, and Social Sciences Education Office Davis, Bette This report contains the following sections: HASS Enrollment Statistics by Field and Subject—Recent Trends; HASS Concentrations: Patterns of Popularity; HASS Minor Programs; Harvard Cross-Registration; Undergraduate Degrees Granted in SHASS; Undergraduate Majors in SHASS; Honors and Awards Granted to Undergraduate Majors in SHASS;

Report to the President for year ended June 30, 2004, School of Humanities, Arts, and Social Sciences

2025-11-22T03:17:13Z

Report to the President for year ended June 30, 2004, School of Humanities, Arts, and Social Sciences Khoury, Philip S. This report contains the following sections: Undergraduate Education; Affirmative Action; Honors and Awards; Fundraising; Faculty Promotions, Administrative Changes, and Retirements;

Report to the President for year ended June 30, 2004, iCampus

2025-11-22T03:17:43Z

Report to the President for year ended June 30, 2004, iCampus Bisbee, Rebecca This report contains the following sections: iCampus Research; Learning Web Services; iLab: Sharing Laboratory Equipment via Web Services; iMoat: Shared Services for Writing Instruction; Reinventing the Classroom with Educational Technology; Computer Science; Physics; Overview of Other Faculty Research; Student Projects;

Report to the President for year ended June 30, 2004, List Visual Arts Center

2025-11-22T03:17:18Z

Report to the President for year ended June 30, 2004, List Visual Arts Center Farver, Jane This report contains the following sections: Current Goals; Accomplishments; Exhibitions; Dean’s Gallery at the Sloan School; Interpretive Program Highlights; Collections; Permanent Collection; Percent for Art; Student Loan Art Collection; Administrative Changes; Finances/Funding; Future Goals; Personnel Information; Advisory Committee