SelfTICA: contrastive learning of dynamical representations for rare-event sampling and characterization
Abstract
Rare events govern many important molecular processes but remain difficult to characterize within accessible simulation timescales. This challenge has motivated machine-learning methods to construct low-dimensional collective variables for specific objectives, including state discrimination, slow-mode identification, and committor-function approximation. Here we introduce SelfTICA, a self-supervised framework that uses contrastive learning on time-lagged configurations to learn a latent representation of the slow modes governing relevant transitions. Once learned, this representation is frozen and reused across downstream tasks, including collective variables construction, enhanced sampling, free-energy estimation, and committor learning. Compared with direct slow-mode optimization, SelfTICA improves training stability and provides collective variables even from limited and exploratory trajectories, accelerating rare-event sampling and enabling accurate free-energy estimation. The dynamical information encoded during pretraining also accelerates committor learning and enables characterization of transition-state regions. These results show that contrastive learning of slow dynamical representations provides a common foundation for rare-event sampling and mechanistic analysis.
Related
- Machine-Learned Dynamical Representations for Accelerated RiteWeight Convergence
- End-to-end multimodal structure elucidation from raw spectra combining contrastive learning and evolutionary algorithms
- Contrastive Reward Based Reinforcement Learning for Goal Directed 3D Molecular Design
- Contrastive Regularization of Machine Learning Potentials
- AET5: A transcriptome-guided molecular generation framework with contrastive self-supervised learning
BibTeX
@misc{zhu2026selftica,
title = {{SelfTICA: contrastive learning of dynamical representations for rare-event sampling and characterization}},
author = {Kai Zhu and Jintu Zhang and Pietro Novelli and Tingjun Hou and Luigi Bonati},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2606.15495},
eprint = {2606.15495},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.15495},
}
RIS
TY - UNPB
TI - SelfTICA: contrastive learning of dynamical representations for rare-event sampling and characterization
AU - Kai Zhu
AU - Jintu Zhang
AU - Pietro Novelli
AU - Tingjun Hou
AU - Luigi Bonati
PY - 2026
DA - 2026/09/09
PB - arXiv
DO - 10.48550/arxiv.2606.15495
UR - https://arxiv.org/abs/2606.15495
AB - Rare events govern many important molecular processes but remain difficult to characterize within accessible simulation timescales. This challenge has motivated machine-learning methods to construct low-dimensional collective variables for specific objectives, including state discrimination, slow-mode identification, and committor-function approximation. Here we introduce SelfTICA, a self-supervised framework that uses contrastive learning on time-lagged configurations to learn a latent representation of the slow modes governing relevant transitions. Once learned, this representation is frozen and reused across downstream tasks, including collective variables construction, enhanced sampling, free-energy estimation, and committor learning. Compared with direct slow-mode optimization, SelfTICA improves training stability and provides collective variables even from limited and exploratory trajectories, accelerating rare-event sampling and enabling accurate free-energy estimation. The dynamical information encoded during pretraining also accelerates committor learning and enables characterization of transition-state regions. These results show that contrastive learning of slow dynamical representations provides a common foundation for rare-event sampling and mechanistic analysis.
ER -