Particle-based Generalised Stochastic Optimisation
Abstract
We develop a class of diffusion-based stochastic particle optimisation methods for loss functions with intractable gradients. Specifically, we consider problems in which the loss gradient is an integral with respect to a parameter-dependent distribution, a structure that includes training generative models, fine-tuning, and learning latent-variable models. We introduce mean-field dynamics and its interacting-particle approximations, which contain several existing algorithms as special cases and provides a route to constructing new methods. Under well-posedness and joint contractivity assumptions, we prove exponential convergence and show that the continuous-time particle system admits a non-asymptotic error bound. We illustrate it by developing momentum and higher-order Langevin variants and evaluating them on maximum marginal-likelihood estimation and energy-based-model training.
Cite
@article{arxiv.2608.02844,
title = {Particle-based Generalised Stochastic Optimisation},
author = {Jiechen Jackie Zhang and O. Deniz Akyildiz},
journal= {arXiv preprint arXiv:2608.02844},
year = {2026}
}