Skip to content

Machine learning kinetics from molecular dynamics data

Sep 2026 · arXiv.org · Vol abs/2609.17736 · 0 citations · 283 references
Physics Computer Science

TL;DR

A common operator viewpoint is developed connecting generator-based partial differential equations, variational principles, Markov state models, dynamical Galerkin approximation, and neural networks with an emphasis on self-supervised methods that learn solutions of their defining dynamical equations rather than relying on labeled shooting data.

Abstract

Most molecular transitions occur on time scales far beyond direct molecular dynamics simulations. The committor, the probability that a configuration reaches a product state before a reactant state, is a central kinetic statistic, providing a mechanism-independent reaction coordinate and a foundation for transition path theory and the calculation of rates. This review surveys modern approaches for estimating the committor and related kinetic statistics from molecular simulations, with an emphasis on self-supervised methods that learn solutions of their defining dynamical equations rather than relying on labeled shooting data. We develop a common operator viewpoint connecting generator-based partial differential equations, variational principles, Markov state models, dynamical Galerkin approximation, and neural networks. Empirical and theoretical evidence points to the efficiency of these methods. We provide theoretical and practical guidance for realizing their full potential in applications, including strategies for treating non-Markovian effects and for sampling. We conclude by identifying opportunities for further research, including connections to reinforcement learning and generative modeling.

View source

Similar papers

Preprint Sep 2026

Learning multistate kinetics with a variational multistate committor network

The long-time dynamics of complex molecular systems often involves rare transitions across networks of metastable states. Building on transition-path theory, which provides a rigorous framework for describing rare transitions between two metastable states, we introduce the variational multistate committor network (VMCN...

Chen-Yu Tang, Cheng Giuseppe Chen, Benoit Roux et al. · 0 citations
Sep 2026

DIME: Dynamics Inferred from Monte Carlo Ensembles via Continuous-Time Markov Chains.

A framework for constructing a continuous-time Markov chain (CTMC) from a static, Boltzmann-weighted ensemble is presented, enabling the generation of physically plausible kinetic trajectories without requiring extensive molecular dynamics simulation.

Krishna Praneet Mulukutla, Marimuthu Krishnan · 0 citations
#machine learning Preprint Aug 2026

Propensity Straight-Through Gradients for Discrete Stochastic Systems

This work exploits the affine state update to obtain the exact one-step conditional-mean sensitivity by differentiating normalized reaction propensities, and defines the propensity straight-through (PST) estimator, a temperature- and Gumbel-free path to scalable gradient-based learning through exact stochastic trajecto...

Jose M. G. Vilar, Leonor Saiz · 0 citations
Preprint Sep 2026

Deep Generative Markov State Models with Experimental Restraints

BICePs-reweighted Reversible DeepMSMs are introduced, combining Bayesian Inference of Conformational Populations (BICePs), variational learning of Markov processes, and maximum entropy (MaxEnt)/maximum caliber (MaxCal) principles to infer consistent thermodynamics and minimally perturbed kinetics.

Robert M. Raddi, Vincent A. Voelz · 0 citations
#protein folding Open access Sep 2026

Breaking timescales with generative sampling of conformational transitions

Gen-COMPAS is introduced, a generative committor-guided path-sampling framework that reconstructs rare biomolecular transition pathways and reveals the underlying thermodynamics and kinetics without using predefined collective variables or brute-force sampling, at an acceptable computational cost.

Chen-Yu Tang, M. P. Pandey, Cheng Giuseppe Chen et al. · 7 citations · ⚡1
Sep 2026

ML-RKIM: Physics-Informed Machine Learning for Reaction Kinetics Identification and Modeling

The method identified sparse, closed-form kinetic representations while reducing reliance on a single predefined kinetic model, and demonstrates the potential of this approach to accelerate kinetic characterization across polymer, food, and biomolecular processing applications, where accurate kinetic models are essenti...

Paulina Portales Picazo, Navid Zobeiry · 0 citations

Related blog posts

MIT News · Artificial Intelligence Nov 17, 2026

Learning the language of machine learning

By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.