Probing reaction channels via reinforcement learning
Abstract
Abstract Chemical reaction are dynamical processes involving the correlated reorganization of atomic configurations, driving the conversion of an initial reactant into a result product. By virtue of the metastability of both the reactants and products, chemical reactions are rare events, proceeding fleetingly. Reaction pathways can be modelled probabilistically by using the notion of reactive density in the phase space of the molecular system. Such density is related to a function known as the committor function, which describes the likelihood of a configuration evolving to one of the nearby metastable regions. In theory, the committor function can be obtained by solving the backward Kolmogorov equation, which is a partial differential equation defined in the full dimensional phase space. However, using traditional methods to solve this problem is not practical for high dimensional systems. In this work, we propose a reinforcement learning based method to identify important configurations that connect reactant and product states along chemical reaction paths. By shooting multiple trajectories from these configurations, we can generate an ensemble of states that concentrate on the transition path ensemble. This configuration ensemble can be effectively employed in a neural network-based partial differential equation solver to obtain an approximation solution of a restrictedmore »
- Authors:
- Publication Date:
- Sponsoring Org.:
- USDOE Office of Science (SC), Basic Energy Sciences (BES); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- OSTI Identifier:
- 2008008
- Alternate Identifier(s):
- OSTI ID: 2001350
- Grant/Contract Number:
- AC02-05CH11231
- Resource Type:
- Published Article
- Journal Name:
- Machine Learning: Science and Technology
- Additional Journal Information:
- Journal Name: Machine Learning: Science and Technology Journal Volume: 4 Journal Issue: 4; Journal ID: ISSN 2632-2153
- Publisher:
- IOP Publishing
- Country of Publication:
- United Kingdom
- Language:
- English
Citation Formats
Liang, Senwei, Singh, Aditya N., Zhu, Yuanran, Limmer, David T., and Yang, Chao. Probing reaction channels via reinforcement learning. United Kingdom: N. p., 2023.
Web. doi:10.1088/2632-2153/acfc33.
Liang, Senwei, Singh, Aditya N., Zhu, Yuanran, Limmer, David T., & Yang, Chao. Probing reaction channels via reinforcement learning. United Kingdom. https://doi.org/10.1088/2632-2153/acfc33
Liang, Senwei, Singh, Aditya N., Zhu, Yuanran, Limmer, David T., and Yang, Chao. Fri .
"Probing reaction channels via reinforcement learning". United Kingdom. https://doi.org/10.1088/2632-2153/acfc33.
@article{osti_2008008,
title = {Probing reaction channels via reinforcement learning},
author = {Liang, Senwei and Singh, Aditya N. and Zhu, Yuanran and Limmer, David T. and Yang, Chao},
abstractNote = {Abstract Chemical reaction are dynamical processes involving the correlated reorganization of atomic configurations, driving the conversion of an initial reactant into a result product. By virtue of the metastability of both the reactants and products, chemical reactions are rare events, proceeding fleetingly. Reaction pathways can be modelled probabilistically by using the notion of reactive density in the phase space of the molecular system. Such density is related to a function known as the committor function, which describes the likelihood of a configuration evolving to one of the nearby metastable regions. In theory, the committor function can be obtained by solving the backward Kolmogorov equation, which is a partial differential equation defined in the full dimensional phase space. However, using traditional methods to solve this problem is not practical for high dimensional systems. In this work, we propose a reinforcement learning based method to identify important configurations that connect reactant and product states along chemical reaction paths. By shooting multiple trajectories from these configurations, we can generate an ensemble of states that concentrate on the transition path ensemble. This configuration ensemble can be effectively employed in a neural network-based partial differential equation solver to obtain an approximation solution of a restricted Backward Kolmogorov equation, even when the dimension of the problem is very high. The resulting solution provides an approximation for the committor function that encodes mechanistic information for the reaction, paving a new way for understanding of complex chemical reactions and evaluation of reaction rates.},
doi = {10.1088/2632-2153/acfc33},
journal = {Machine Learning: Science and Technology},
number = 4,
volume = 4,
place = {United Kingdom},
year = {Fri Oct 06 00:00:00 EDT 2023},
month = {Fri Oct 06 00:00:00 EDT 2023}
}
https://doi.org/10.1088/2632-2153/acfc33
Works referenced in this record:
Computing committor functions for the study of rare events using deep learning
journal, August 2019
- Li, Qianxiao; Lin, Bo; Ren, Weiqing
- The Journal of Chemical Physics, Vol. 151, Issue 5
Forward flux sampling for rare event simulations
journal, October 2009
- Allen, Rosalind J.; Valeriani, Chantal; Rein ten Wolde, Pieter
- Journal of Physics: Condensed Matter, Vol. 21, Issue 46
Reinforcement learning of rare diffusive dynamics
journal, October 2021
- Das, Avishek; Rose, Dominic C.; Garrahan, Juan P.
- The Journal of Chemical Physics, Vol. 155, Issue 13
Statistical mechanics of isomerization dynamics in liquids and the transition state approximation
journal, January 1978
- Chandler, David
- The Journal of Chemical Physics, Vol. 68, Issue 6
Predicting rare events using neural networks and short-trajectory data
journal, September 2023
- Strahan, John; Finkel, Justin; Dinner, Aaron R.
- Journal of Computational Physics, Vol. 488
Algorithms for solving high dimensional PDEs: from nonlinear Monte Carlo to machine learning
journal, December 2021
- E., Weinan; Han, Jiequn; Jentzen, Arnulf
- Nonlinearity, Vol. 35, Issue 1
A reinforcement learning approach to rare trajectory sampling
journal, January 2021
- Rose, Dominic C.; Mair, Jamie F.; Garrahan, Juan P.
- New Journal of Physics, Vol. 23, Issue 1
Inexact iterative numerical linear algebra for neural network-based spectral estimation and rare-event prediction
journal, July 2023
- Strahan, John; Guo, Spencer C.; Lorpaiboon, Chatipat
- The Journal of Chemical Physics, Vol. 159, Issue 1
Well-Tempered Metadynamics: A Smoothly Converging and Tunable Free-Energy Method
journal, January 2008
- Barducci, Alessandro; Bussi, Giovanni; Parrinello, Michele
- Physical Review Letters, Vol. 100, Issue 2
Kinetic Pathways of Ion Pair Dissociation in Water
journal, May 1999
- Geissler, Phillip L.; Dellago, Christoph; Chandler, David
- The Journal of Physical Chemistry B, Vol. 103, Issue 18
Identifying functional brain networks from spatial-temporal wide-field calcium imaging data via a recurrent autoencoder
conference, April 2022
- Zhang, Xiaohui; Landsness, Eric C.; Culver, Joseph P.
- Neural Imaging and Sensing 2022
Variational deep learning of equilibrium transition path ensembles
journal, July 2023
- Singh, Aditya N.; Limmer, David T.
- The Journal of Chemical Physics, Vol. 159, Issue 2
Local and global perspectives on diffusion maps in the analysis of molecular systems
journal, January 2020
- Trstanova, Z.; Leimkuhler, B.; Lelièvre, T.
- Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 476, Issue 2233
T RANSITION P ATH S AMPLING : Throwing Ropes Over Rough Mountain Passes, in the Dark
journal, October 2002
- Bolhuis, Peter G.; Chandler, David; Dellago, Christoph
- Annual Review of Physical Chemistry, Vol. 53, Issue 1
Machine-guided path sampling to discover mechanisms of molecular self-organization
journal, April 2023
- Jung, Hendrik; Covino, Roberto; Arjun, A.
- Nature Computational Science, Vol. 3, Issue 4
WESTPA: An Interoperable, Highly Scalable Software Package for Weighted Ensemble Simulation and Analysis
journal, January 2015
- Zwier, Matthew C.; Adelman, Joshua L.; Kaus, Joseph W.
- Journal of Chemical Theory and Computation, Vol. 11, Issue 2
Explaining reaction coordinates of alanine dipeptide isomerization obtained from deep neural networks using Explainable Artificial Intelligence (XAI)
journal, April 2022
- Kikutsuji, Takuma; Mori, Yusuke; Okazaki, Kei-ichi
- The Journal of Chemical Physics, Vol. 156, Issue 15
Weighted-ensemble Brownian dynamics simulations for protein association reactions
journal, January 1996
- Huber, G. A.; Kim, S.
- Biophysical Journal, Vol. 70, Issue 1
Illustration of transition path theory on a collection of simple examples
journal, August 2006
- Metzner, Philipp; Schütte, Christof; Vanden-Eijnden, Eric
- The Journal of Chemical Physics, Vol. 125, Issue 8
Frequency Principle: Fourier Analysis Sheds Light on Deep Neural Networks
journal, June 2020
- Xu, Zhi-Qin John
- Communications in Computational Physics, Vol. 28, Issue 5
Towards a Theory of Transition Paths
journal, May 2006
- E., Weinan; Vanden-Eijnden, Eric
- Journal of Statistical Physics, Vol. 123, Issue 3
Variational Approach to Enhanced Sampling and Free Energy Calculations
journal, August 2014
- Valsson, Omar; Parrinello, Michele
- Physical Review Letters, Vol. 113, Issue 9
Transition Path Sampling as Markov Chain Monte Carlo of Trajectories: Recent Algorithms, Software, Applications, and Future Outlook
journal, March 2021
- Bolhuis, Peter G.; Swenson, David W. H.
- Advanced Theory and Simulations, Vol. 4, Issue 4
Variational control forces for enhanced sampling of nonequilibrium molecular dynamics simulations
journal, December 2019
- Das, Avishek; Limmer, David T.
- The Journal of Chemical Physics, Vol. 151, Issue 24
Transition-Path Theory and Path-Finding Algorithms for the Study of Rare Events
journal, March 2010
- E., Weinan; Vanden-Eijnden, Eric
- Annual Review of Physical Chemistry, Vol. 61, Issue 1
Galerkin approximation of dynamical quantities using trajectory data
journal, June 2019
- Thiede, Erik H.; Giannakis, Dimitrios; Dinner, Aaron R.
- The Journal of Chemical Physics, Vol. 150, Issue 24
Metadynamics: Metadynamics
journal, February 2011
- Barducci, Alessandro; Bonomi, Massimiliano; Parrinello, Michele
- Wiley Interdisciplinary Reviews: Computational Molecular Science, Vol. 1, Issue 5
Direct Evaluation of Rare Events in Active Matter from Variational Path Sampling
journal, January 2022
- Das, Avishek; Kuznets-Speck, Benjamin; Limmer, David T.
- Physical Review Letters, Vol. 128, Issue 2
Automatic Method for Identifying Reaction Coordinates in Complex Systems †
journal, April 2005
- Ma, Ao; Dinner, Aaron R.
- The Journal of Physical Chemistry B, Vol. 109, Issue 14
Deep reinforcement learning of transition states
journal, January 2021
- Zhang, Jun; Lei, Yao-Kun; Zhang, Zhen
- Physical Chemistry Chemical Physics, Vol. 23, Issue 11
Deep Network With Approximation Error Being Reciprocal of Width to Power of Square Root of Depth
journal, March 2021
- Shen, Zuowei; Yang, Haizhao; Zhang, Shijun
- Neural Computation, Vol. 33, Issue 4
Solving for high-dimensional committor functions using artificial neural networks
journal, October 2018
- Khoo, Yuehaw; Lu, Jianfeng; Ying, Lexing
- Research in the Mathematical Sciences, Vol. 6, Issue 1
Diffusion Maps, Reduction Coordinates, and Low Dimensional Representation of Stochastic Systems
journal, January 2008
- Coifman, R. R.; Kevrekidis, I. G.; Lafon, S.
- Multiscale Modeling & Simulation, Vol. 7, Issue 2
Computing committors via Mahalanobis diffusion maps with enhanced sampling data
journal, December 2022
- Evans, L.; Cameron, M. K.; Tiwary, P.
- The Journal of Chemical Physics, Vol. 157, Issue 21
Learning stochastic dynamics with statistics-informed neural network
journal, February 2023
- Zhu, Yuanran; Tang, Yu-Hang; Kim, Changho
- Journal of Computational Physics, Vol. 474
String method for the study of rare events
journal, August 2002
- E., Weinan; Ren, Weiqing; Vanden-Eijnden, Eric
- Physical Review B, Vol. 66, Issue 5
OpenMM 7: Rapid development of high performance algorithms for molecular dynamics
journal, July 2017
- Eastman, Peter; Swails, Jason; Chodera, John D.
- PLOS Computational Biology, Vol. 13, Issue 7
Using Diffusion Maps to Analyze Reaction Dynamics for a Hydrogen Combustion Benchmark Dataset
journal, August 2023
- Ko, Taehee; Heindel, Joseph P.; Guan, Xingyi
- Journal of Chemical Theory and Computation, Vol. 19, Issue 17
Ghost Point Diffusion Maps for Solving Elliptic PDEs on Manifolds with Classical Boundary Conditions
journal, December 2021
- Jiang, Shixiao Willing; Harlim, John
- Communications on Pure and Applied Mathematics, Vol. 76, Issue 2
Supervised learning and the finite-temperature string method for computing committor functions and reaction rates
journal, November 2022
- Hasyim, Muhammad R.; Batton, Clay H.; Mandadapu, Kranthi K.
- The Journal of Chemical Physics, Vol. 157, Issue 18