Identifying nonequilibrium degrees of freedom in high-dimensional stochastic systems

2025/08/11 by Catherine Ji, Ji, Catherine, Raj, Ravin +4
Computer Science · Physics and Astronomy · #Advanced Thermodynamics and Statistical Mechanics #Biological Physics (physics.bio-ph) #FOS: Physical sciences #Neural Networks and Reservoir Computing #Quantum many-body systems #Statistical Mechanics (cond-mat.stat-mech)

paper · pdf · doi:10.48550/arxiv.2508.08247

openalex publication_date 2025/08/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Any coarse-grained description of a nonequilibrium system should faithfully represent its latent irreversible degrees of freedom. However, standard dimensionality reduction methods typically prioritize accurate reconstruction over physical relevance. Here, we introduce a model-free approach to identify irreversible degrees of freedom in stochastic systems that are in a nonequilibrium steady state. Our method leverages the insight that a black-box classifier, trained to differentiate between forward and time-reversed trajectories, implicitly estimates the local entropy production rate. By parameterizing this classifier as a quadratic form of learned state representations, we obtain nonlinear embeddings of high-dimensional state-space dynamics, which we term Latent Embeddings of Nonequilibrium Systems (LENS). LENS effectively identifies low-dimensional irreversible flows and provides a scalable, learning-based strategy for estimating entropy production rates directly from high-dimensional time series data.

Citations

• Learning Stochastic Thermodynamics Directly from Correlation and Trajectory-Fluctuation Currents
• Learning rate matrix and information-thermodynamic trade-off relation
• Model-free learning of probability flows: Elucidating the nonequilibrium dynamics of flocking
• Accelerating Goal-Conditioned RL Algorithms and Research
• Deep learning probability flows and entropy production rates in active matter
• Measuring irreversibility from learned representations of biological patterns
• Scale-dependent irreversibility in living matter
• Implicit Neural Representations with Periodic Activation Functions
• Noise Contrastive Estimation and Negative Sampling for Conditional Models: Consistency and Statistical Efficiency
• An Intriguing Failing of Convolutional Neural Networks and the CoordConv Solution
• Layer Normalization
• Gaussian Error Linear Units (GELUs)
• f-GAN: Training Generative Neural Samplers using Variational Divergence Minimization
• Deep Residual Learning for Image Recognition
• Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images
• Stochastic thermodynamics, fluctuation theorems and molecular machines
• Fourier's Law for a Harmonic Crystal with Self-Consistent Stochastic Reservoirs
• Exponential Time Differencing for Stiff Systems

Related

• Model-free learning of probability flows: Elucidating the nonequilibrium dynamics of flocking
• Identifiable learning of dissipative dynamics
• Inferring entropy production in many-body systems using nonequilibrium maximum entropy
• Emergence of Nonequilibrium Latent Cycles in Unsupervised Generative Modeling
• Universal bounds on entropy production from fluctuating coarse-grained trajectories