English

A Stochastic Variance Reduction Algorithm with Bregman Distances for Structured Composite Problems

Optimization and Control 2021-03-17 v1

Abstract

We develop a novel stochastic primal dual splitting method with Bregman distances for solving a structured composite problems involving infimal convolutions in non-Euclidean spaces. The sublinear convergence in expectation of the primal-dual gap is proved under mild conditions on stepsize for the general case. The linear convergence rate is obtained under additional condition like the strong convexity relative to Bregman functions.

Keywords

Cite

@article{arxiv.2103.08822,
  title  = {A Stochastic Variance Reduction Algorithm with Bregman Distances for Structured Composite Problems},
  author = {Nguyen Van Dung and Băng Công Vũ},
  journal= {arXiv preprint arXiv:2103.08822},
  year   = {2021}
}