Stochastic Gradient Methods with Compressed Communication for Decentralized Saddle Point Problems

TL;DR

This paper introduces two novel decentralized stochastic gradient algorithms with compression techniques to efficiently solve non-smooth saddle-point problems, providing theoretical guarantees and demonstrating competitive empirical performance.

Contribution

The paper develops two new compression-based stochastic gradient algorithms for decentralized saddle-point problems, with rigorous complexity analysis and empirical validation.

Findings

Algorithms achieve near-optimal complexity bounds.

Numerical experiments confirm theoretical guarantees.

Methods outperform existing approaches in decentralized settings.

Abstract

We develop two compression based stochastic gradient algorithms to solve a class of non-smooth strongly convex-strongly concave saddle-point problems in a decentralized setting (without a central server). Our first algorithm is a Restart-based Decentralized Proximal Stochastic Gradient method with Compression (C-RDPSG) for general stochastic settings. We provide rigorous theoretical guarantees of C-RDPSG with gradient computation complexity and communication complexity of order $\mathcal{O}( (1+\delta)^4 \frac{1}{L^2}{\kappa_f^2}\kappa_g^2 \frac{1}{\epsilon} )$, to achieve an $\epsilon$-accurate saddle-point solution, where $\delta$ denotes the compression factor, $\kappa_f$ and $\kappa_g$ denote respectively the condition numbers of objective function and communication graph, and $L$ denotes the smoothness parameter of the smooth part of the objective function. Next, we present a Decentralized Proximal Stochastic Variance Reduced Gradient algorithm with Compression (C-DPSVRG) for finite sum setting which exhibits gradient computation complexity and communication complexity of order $\mathcal{O} \left((1+\delta) \max \{\kappa_f^2, \sqrt{\delta}\kappa^2_f\kappa_g,\kappa_g \} \log\left(\frac{1}{\epsilon}\right) \right)$. Extensive numerical experiments show competitive performance of the proposed algorithms and provide support to the theoretical results obtained.