Generative discovery of de novo chemical designs using diffusion modeling and transformer deep neural networks with application to deep eutectic solvents
Name
234103_1_5.0155890.pdf
Description
Published version
Size
2.98 MB
Format
Adobe PDF
Checksum (MD5)
5fdc3e07ce0fe1733c5ebeffd6b02613
Author(s)
Luu, Rachel K
Wysokowski, Marcin
Buehler, Markus J
Date Issued
June 5, 2023
Journal
Applied Physics Letters
Publisher
AIP Publishing
Citation
Rachel K. Luu, Marcin Wysokowski, Markus J. Buehler; Generative discovery of de novo chemical designs using diffusion modeling and transformer deep neural networks with application to deep eutectic solvents. Appl. Phys. Lett. 5 June 2023; 122 (23): 234103.
Version
Final published version
Abstract
We report a series of deep learning models to solve complex forward and inverse design problems in molecular modeling and design. Using both diffusion models inspired by nonequilibrium thermodynamics and attention-based transformer architectures, we demonstrate a flexible framework to capture complex chemical structures. First trained on the Quantum Machines 9 (QM9) dataset and a series of quantum mechanical properties (e.g., homo, lumo, free energy, and heat capacity), we then generalize the model to study and design key properties of deep eutectic solvents (DESs). In addition to separate forward and inverse models, we also report an integrated fully prompt-based multi-task generative pretrained transformer model that solves multiple forward, inverse design, and prediction tasks, flexibly and within one model. We show that the multi-task generative model has the overall best performance and allows for flexible integration of multiple objectives, within one model, and for distinct chemistries, suggesting that synergies emerge during training of this large language model. Trained jointly in tasks related to the QM9 dataset and DESs, the model can predict various quantum mechanical properties and critical properties to achieve deep eutectic solvent behavior. Several combinations of DESs are proposed based on this framework.
MIT Department
Massachusetts Institute of Technology. Laboratory for Atomistic and Molecular Mechanics
Massachusetts Institute of Technology. Center for Computational Science and Engineering
Terms of Use
Creative Commons Attribution
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1063/5.0155890
