Learning in Games: Robustness of Fast Convergence

TL;DR

This paper demonstrates that learning algorithms with low approximate regret lead to rapid convergence to near-optimal outcomes in repeated games, broadening the scope of applicable algorithms and feedback models.

Contribution

It introduces a low approximate regret property that ensures fast convergence in various game settings, including bandit feedback and dynamic populations, with improved speed and broader applicability.

Findings

Fast convergence with high probability in repeated games

Applicability to bandit feedback and dynamic populations

Improved convergence speed proportional to the number of players

Abstract

We show that learning algorithms satisfying a $\textit{low approximate regret}$ property experience fast convergence to approximate optimality in a large class of repeated games. Our property, which simply requires that each learner has small regret compared to a $(1+\epsilon)$-multiplicative approximation to the best action in hindsight, is ubiquitous among learning algorithms; it is satisfied even by the vanilla Hedge forecaster. Our results improve upon recent work of Syrgkanis et al. [SALS15] in a number of ways. We require only that players observe payoffs under other players' realized actions, as opposed to expected payoffs. We further show that convergence occurs with high probability, and show convergence under bandit feedback. Finally, we improve upon the speed of convergence by a factor of $n$, the number of players. Both the scope of settings and the class of algorithms for which our analysis provides fast convergence are considerably broader than in previous work. Our framework applies to dynamic population games via a low approximate regret property for shifting experts. Here we strengthen the results of Lykouris et al. [LST16] in two ways: We allow players to select learning algorithms from a larger class, which includes a minor variant of the basic Hedge algorithm, and we increase the maximum churn in players for which approximate optimality is achieved. In the bandit setting we present a new algorithm which provides a "small loss"-type bound with improved dependence on the number of actions in utility settings, and is both simple and efficient. This result may be of independent interest.