Deep Generative Markov State Models

TL;DR

DeepGenMSM is a deep learning framework that models metastable dynamical systems, enabling accurate long-term predictions and generation of realistic molecular configurations, even in unobserved regions.

Contribution

It introduces a novel deep generative Markov State Model that combines probabilistic encoding, Markov chain dynamics, and generative sampling for metastable systems.

Findings

Accurately estimates long-time kinetics of molecular systems.

Generates physically realistic structures beyond training data.

Operates efficiently with long time-steps in configuration space.

Abstract

We propose a deep generative Markov State Model (DeepGenMSM) learning framework for inference of metastable dynamical systems and prediction of trajectories. After unsupervised training on time series data, the model contains (i) a probabilistic encoder that maps from high-dimensional configuration space to a small-sized vector indicating the membership to metastable (long-lived) states, (ii) a Markov chain that governs the transitions between metastable states and facilitates analysis of the long-time dynamics, and (iii) a generative part that samples the conditional distribution of configurations in the next time step. The model can be operated in a recursive fashion to generate trajectories to predict the system evolution from a defined starting state and propose new configurations. The DeepGenMSM is demonstrated to provide accurate estimates of the long-time kinetics and generate valid distributions for molecular dynamics (MD) benchmark systems. Remarkably, we show that DeepGenMSMs are able to make long time-steps in molecular configuration space and generate physically realistic structures in regions that were not seen in training data.