Learning Strategy Representation for Imitation Learning in Multi-Agent Games

TL;DR

This paper introduces STRIL, a novel framework for learning strategy representations in multi-agent imitation learning, which improves data filtering and enhances performance without requiring player identification.

Contribution

STRIL is a new plug-in framework that effectively learns strategy representations and indicators in multi-agent games, improving imitation learning by filtering sub-optimal data.

Findings

Successfully learns strategy representations in multi-agent scenarios

Effectively filters sub-optimal trajectories using indicators

Significantly improves imitation learning performance in various games

Abstract

The offline datasets for imitation learning (IL) in multi-agent games typically contain player trajectories exhibiting diverse strategies, which necessitate measures to prevent learning algorithms from acquiring undesirable behaviors. Learning representations for these trajectories is an effective approach to depicting the strategies employed by each demonstrator. However, existing learning strategies often require player identification or rely on strong assumptions, which are not appropriate for multi-agent games. Therefore, in this paper, we introduce the Strategy Representation for Imitation Learning (STRIL) framework, which (1) effectively learns strategy representations in multi-agent games, (2) estimates proposed indicators based on these representations, and (3) filters out sub-optimal data using the indicators. STRIL is a plug-in method that can be integrated into existing IL algorithms. We demonstrate the effectiveness of STRIL across competitive multi-agent scenarios, including Two-player Pong, Limit Texas Hold'em, and Connect Four. Our approach successfully acquires strategy representations and indicators, thereby identifying dominant trajectories and significantly enhancing existing IL performance across these environments.