Stochastic Gradient Langevin Dynamics with Variance Reduction
Machine Learning
2024-07-08 v1 Optimization and Control
Abstract
Stochastic gradient Langevin dynamics (SGLD) has gained the attention of optimization researchers due to its global optimization properties. This paper proves an improved convergence property to local minimizers of nonconvex objective functions using SGLD accelerated by variance reductions. Moreover, we prove an ergodicity property of the SGLD scheme, which gives insights on its potential to find global minimizers of nonconvex objectives.
Cite
@article{arxiv.2102.06759,
title = {Stochastic Gradient Langevin Dynamics with Variance Reduction},
author = {Zhishen Huang and Stephen Becker},
journal= {arXiv preprint arXiv:2102.06759},
year = {2024}
}