English

Constrained Exploration via Reflected Replica Exchange Stochastic Gradient Langevin Dynamics

Machine Learning 2024-06-04 v2 Artificial Intelligence Machine Learning

Abstract

Replica exchange stochastic gradient Langevin dynamics (reSGLD) is an effective sampler for non-convex learning in large-scale datasets. However, the simulation may encounter stagnation issues when the high-temperature chain delves too deeply into the distribution tails. To tackle this issue, we propose reflected reSGLD (r2SGLD): an algorithm tailored for constrained non-convex exploration by utilizing reflection steps within a bounded domain. Theoretically, we observe that reducing the diameter of the domain enhances mixing rates, exhibiting a quadratic\textit{quadratic} behavior. Empirically, we test its performance through extensive experiments, including identifying dynamical systems with physical constraints, simulations of constrained multi-modal distributions, and image classification tasks. The theoretical and empirical findings highlight the crucial role of constrained exploration in improving the simulation efficiency.

Keywords

Cite

@article{arxiv.2405.07839,
  title  = {Constrained Exploration via Reflected Replica Exchange Stochastic Gradient Langevin Dynamics},
  author = {Haoyang Zheng and Hengrong Du and Qi Feng and Wei Deng and Guang Lin},
  journal= {arXiv preprint arXiv:2405.07839},
  year   = {2024}
}

Comments

28 pages, 13 figures

R2 v1 2026-06-28T16:25:32.162Z