Probabilistic programming with programmable inference
Name
MansinghkaEtAl_pldi18.pdf
Description
Accepted version
Size
961.33 KB
Format
Adobe PDF
Checksum (MD5)
a25e9843a377ee4f307e487dc58b1825
Author(s)
Mansinghka, Vikash K.
Schaechtle, Ulrich
Handa, Shivam
Radul, Alexey
Chen, Yutian
Rinard, Martin
Date Issued
June 11, 2018
Publisher
ACM
Citation
Mansinghka, Vikash K., Schaechtle, Ulrich, Handa, Shivam, Radul, Alexey, Chen, Yutian et al. 2018. "Probabilistic programming with programmable inference."
Version
Author's final manuscript
Abstract
© 2018 Copyright held by the owner/author(s). We introduce inference metaprogramming for probabilistic programming languages, including new language constructs, a formalism, and the first demonstration of effectiveness in practice. Instead of relying on rigid black-box inference algorithms hard-coded into the language implementation as in previous probabilistic programming languages, inference metaprogramming enables developers to 1) dynamically decompose inference problems into subproblems, 2) apply inference tactics to subproblems, 3) alternate between incorporating new data and performing inference over existing data, and 4) explore multiple execution traces of the probabilistic program at once. Implemented tactics include gradient-based optimization, Markov chain Monte Carlo, variational inference, and sequental Monte Carlo techniques. Inference metaprogramming enables the concise expression of probabilistic models and inference algorithms across diverse fields, such as computer vision, data science, and robotics, within a single probabilistic programming language.
MIT Department
MIT-IBM Watson AI Lab
Massachusetts Institute of Technology. Department of Brain and Cognitive Sciences
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1145/3192366.3192409
