Learning multistate kinetics with a variational multistate committor network
Abstract
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), a neural framework that learns the probabilities of reaching each metastable state directly from molecular simulation data. From this representation, VMCN identifies state-specific commitment, candidate transition regions and committor-consistent pathways between state pairs, and an effective kinetic network characterized by transition rates. The model is trained using finite time-lag trajectory data together with boundary conditions defined on conservative state cores. Applications to a triple-well potential, trialanine isomerization, and the c--ring rotation in the Vo domain of a vacuolar ATPase show that VMCN recovers metastable organization, provides committor-consistent descriptions of transition mechanisms, and estimates state-to-state kinetics. VMCN further provides diagnostics for incomplete state decompositions and enables adaptive exploration of candidate metastable states and their connecting regions. By integrating VMCN with generative committor-guided path sampling (Gen-COMPAS) for chignolin, we start from two end point structures, identify a misfolded state and a candidate folding intermediate, and we direct subsequent sampling toward the resulting multistate transition network.
Related
- Extended Onsager's variational principle for nonlinear chemical reaction kinetics in active matter
- All Paths All at Once: A Quantum Monte Carlo and Neural Committor Approach to Reaction Pathway Search
- Gradient-estimator design overcomes trainability barriers in neural-network-based variational optimization
- Reaction-Network Kinetics of Lignin Hydrothermal Liquefaction: Effect of Cu Incorporation Strategy
- Tensor-based Approximation of Molecular Kinetics: Generator Learning, Reaction Coordinates and Incremental Updating
BibTeX
@misc{tang2026learning,
title = {{Learning multistate kinetics with a variational multistate committor network}},
author = {Chenyu Tang and Cheng Giuseppe Chen and Benoît Roux and Christophe Chipot},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.04449},
eprint = {2609.04449},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.04449},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Learning multistate kinetics with a variational multistate committor network
AU - Chenyu Tang
AU - Cheng Giuseppe Chen
AU - Benoît Roux
AU - Christophe Chipot
PY - 2026
DA - 2026/09/03
PB - arXiv
DO - 10.48550/arxiv.2609.04449
UR - https://arxiv.org/abs/2609.04449
AB - 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), a neural framework that learns the probabilities of reaching each metastable state directly from molecular simulation data. From this representation, VMCN identifies state-specific commitment, candidate transition regions and committor-consistent pathways between state pairs, and an effective kinetic network characterized by transition rates. The model is trained using finite time-lag trajectory data together with boundary conditions defined on conservative state cores. Applications to a triple-well potential, trialanine isomerization, and the c--ring rotation in the V_o domain of a vacuolar ATPase show that VMCN recovers metastable organization, provides committor-consistent descriptions of transition mechanisms, and estimates state-to-state kinetics. VMCN further provides diagnostics for incomplete state decompositions and enables adaptive exploration of candidate metastable states and their connecting regions. By integrating VMCN with generative committor-guided path sampling (Gen-COMPAS) for chignolin, we start from two end point structures, identify a misfolded state and a candidate folding intermediate, and we direct subsequent sampling toward the resulting multistate transition network.
ER -