analiza aiqbench
FEAT introduces a unified framework that merges equilibrium and non-equilibrium methods to solve free energy estimation, a foundational task in molecular simulations and quantum field theory. By utilizing learned stochastic interpolants, the approach achieves consistent, minimum-variance estimators that outperform current learning-based techniques across diverse scientific domains. This work advances the theoretical underpinnings of neural free energy calculations with practical implementations available for immediate use.
FEAT: Free energy Estimators with Adaptive Transport
arXiv:2504.11516v4 Announce Type: replace-cross Abstract: We present Free energy Estimators with Adaptive Transport (FEAT), a novel framework for free energy estimation -- a critical challenge across scientific domains. FEAT leverages learned transports implemented via stochastic interpolants and provides consistent, minimum-variance estimators based on escorted Jarzynski equality and controlled Crooks theorem, alongside variational upper and lower bounds on free energy differences. Unifying equilibrium and non-equilibrium methods under a single theoretical framework, FEAT establishes a principled foundation for neural free energy calculations. Experimental validation on toy examples, molecular simulations, and quantum field theory demonstrates improvements over existing learning-based methods. Our PyTorch implementation is available at https://github.com/jiajunhe98/FEAT.
Abstract:We present Free energy Estimators with Adaptive Transport (FEAT), a novel framework for free energy estimation -- a critical challenge across scientific domains. FEAT leverages learned transports implemented via stochastic interpolants and provides consistent, minimum-variance estimators based on escorted Jarzynski equality and controlled Crooks theorem, alongside variational upper and lower bounds on free energy differences. Unifying equilibrium and non-equilibrium methods under a single theoretical framework, FEAT establishes a principled foundation for neural free energy calculations. Experimental validation on toy examples, molecular simulations, and quantum field theory demonstrates improvements over existing learning-based methods. Our PyTorch implementation is available at this https URL.
Accepted to NeurIPS 2025; the first two authors contribute equally to this work; add an erratum to Appendix F.4 clarifying the applicability of the TSM loss
Machine Learning (stat.ML); Machine Learning (cs.LG); Chemical Physics (physics.chem-ph); Computational Physics (physics.comp-ph)
arXiv:2504.11516 [stat.ML]
(or arXiv:2504.11516v4 [stat.ML] for this version)
Text integral sub licența deschisă a publicației · 147 cuvinte
