Differentiable Simulations for Enhanced Sampling of Rare Events

Martin Sipka, Johannes C. B. Dietschreit, Lukáš Grajciar, Rafael Gomez-Bombarelli
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:31990-32007, 2023.

Abstract

Simulating rare events, such as the transformation of a reactant into a product in a chemical reaction typically requires enhanced sampling techniques that rely on heuristically chosen collective variables (CVs). We propose using differentiable simulations (DiffSim) for the discovery and enhanced sampling of chemical transformations without a need to resort to preselected CVs, using only a distance metric. Reaction path discovery and estimation of the biasing potential that enhances the sampling are merged into a single end-to-end problem that is solved by path-integral optimization. This is achieved by introducing multiple improvements over standard DiffSim such as partial backpropagation and graph mini-batching making DiffSim training stable and efficient. The potential of DiffSim is demonstrated in the successful discovery of transition paths for the Muller-Brown model potential as well as a benchmark chemical system - alanine dipeptide.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-sipka23a, title = {Differentiable Simulations for Enhanced Sampling of Rare Events}, author = {Sipka, Martin and Dietschreit, Johannes C. B. and Grajciar, Luk\'{a}\v{s} and Gomez-Bombarelli, Rafael}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {31990--32007}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/sipka23a/sipka23a.pdf}, url = {https://proceedings.mlr.press/v202/sipka23a.html}, abstract = {Simulating rare events, such as the transformation of a reactant into a product in a chemical reaction typically requires enhanced sampling techniques that rely on heuristically chosen collective variables (CVs). We propose using differentiable simulations (DiffSim) for the discovery and enhanced sampling of chemical transformations without a need to resort to preselected CVs, using only a distance metric. Reaction path discovery and estimation of the biasing potential that enhances the sampling are merged into a single end-to-end problem that is solved by path-integral optimization. This is achieved by introducing multiple improvements over standard DiffSim such as partial backpropagation and graph mini-batching making DiffSim training stable and efficient. The potential of DiffSim is demonstrated in the successful discovery of transition paths for the Muller-Brown model potential as well as a benchmark chemical system - alanine dipeptide.} }
Endnote
%0 Conference Paper %T Differentiable Simulations for Enhanced Sampling of Rare Events %A Martin Sipka %A Johannes C. B. Dietschreit %A Lukáš Grajciar %A Rafael Gomez-Bombarelli %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-sipka23a %I PMLR %P 31990--32007 %U https://proceedings.mlr.press/v202/sipka23a.html %V 202 %X Simulating rare events, such as the transformation of a reactant into a product in a chemical reaction typically requires enhanced sampling techniques that rely on heuristically chosen collective variables (CVs). We propose using differentiable simulations (DiffSim) for the discovery and enhanced sampling of chemical transformations without a need to resort to preselected CVs, using only a distance metric. Reaction path discovery and estimation of the biasing potential that enhances the sampling are merged into a single end-to-end problem that is solved by path-integral optimization. This is achieved by introducing multiple improvements over standard DiffSim such as partial backpropagation and graph mini-batching making DiffSim training stable and efficient. The potential of DiffSim is demonstrated in the successful discovery of transition paths for the Muller-Brown model potential as well as a benchmark chemical system - alanine dipeptide.
APA
Sipka, M., Dietschreit, J.C.B., Grajciar, L. & Gomez-Bombarelli, R.. (2023). Differentiable Simulations for Enhanced Sampling of Rare Events. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:31990-32007 Available from https://proceedings.mlr.press/v202/sipka23a.html.

Related Material