Local Learning Rules for Out-of-Equilibrium Physical Generative Models
Machine Learning
2025-08-28 v3 Mesoscale and Nanoscale Physics
Emerging Technologies
Neural and Evolutionary Computing
Abstract
We show that the out-of-equilibrium driving protocol of score-based generative models (SGMs) can be learned via local learning rules. The gradient with respect to the parameters of the driving protocol is computed directly from force measurements or from observed system dynamics. As a demonstration, we implement an SGM in a network of driven, nonlinear, overdamped oscillators coupled to a thermal bath. We first apply it to the problem of sampling from a mixture of two Gaussians in 2D. Finally, we train a 12x12 oscillator network on the MNIST dataset to generate images of handwritten digits 0 and 1.
Cite
@article{arxiv.2506.19136,
title = {Local Learning Rules for Out-of-Equilibrium Physical Generative Models},
author = {Cyrill Bösch and Geoffrey Roeder and Marc Serra-Garcia and Ryan P. Adams},
journal= {arXiv preprint arXiv:2506.19136},
year = {2025}
}
Comments
6 pages, 2 figures