Decoupled SGDA for Games with Intermittent Strategy Communication

TL;DR

This paper introduces Decoupled SGDA, a communication-efficient algorithm for multiplayer games with intermittent strategy updates, achieving near-optimal convergence and significantly reducing communication costs in various game settings.

Contribution

The paper proposes Decoupled SGDA, a novel method that allows independent strategy updates with periodic synchronization, improving communication efficiency in multiplayer minimax games.

Findings

Achieves near-optimal communication complexity in SCSC games.

Reduces communication costs in weakly coupled games.

Outperforms federated minimax methods in imbalanced noise settings.

Abstract

We focus on reducing communication overhead in multiplayer games, where frequently exchanging strategies between players is not feasible and players have noisy or outdated strategies of the other players. We introduce Decoupled SGDA, a novel adaptation of Stochastic Gradient Descent Ascent (SGDA). In this approach, players independently update their strategies based on outdated opponent strategies, with periodic synchronization to align strategies. For Strongly-Convex-Strongly-Concave (SCSC) games, we demonstrate that Decoupled SGDA achieves near-optimal communication complexity comparable to the best-known GDA rates. For weakly coupled games where the interaction between players is lower relative to the non-interactive part of the game, Decoupled SGDA significantly reduces communication costs compared to standard SGDA. Our findings extend to multi-player games. To provide insights into the effect of communication frequency and convergence, we extensively study the convergence of Decoupled SGDA for quadratic minimax problems. Lastly, in settings where the noise over the players is imbalanced, Decoupled SGDA significantly outperforms federated minimax methods.