Transformers as Game Players: Provable In-context Game-playing Capabilities of Pre-trained Models
This work addresses a gap in theoretical understanding for multi-agent game-playing, which is important for researchers in reinforcement learning and game theory, though it appears incremental as it builds on prior single-agent results.
The paper tackles the problem of extending in-context learning capabilities of pre-trained transformer models to competitive multi-agent games, specifically two-player zero-sum games, and provides theoretical guarantees that these models can provably approximate Nash equilibrium in both decentralized and centralized settings.
The in-context learning (ICL) capability of pre-trained models based on the transformer architecture has received growing interest in recent years. While theoretical understanding has been obtained for ICL in reinforcement learning (RL), the previous results are largely confined to the single-agent setting. This work proposes to further explore the in-context learning capabilities of pre-trained transformer models in competitive multi-agent games, i.e., in-context game-playing (ICGP). Focusing on the classical two-player zero-sum games, theoretical guarantees are provided to demonstrate that pre-trained transformers can provably learn to approximate Nash equilibrium in an in-context manner for both decentralized and centralized learning settings. As a key part of the proof, constructional results are established to demonstrate that the transformer architecture is sufficiently rich to realize celebrated multi-agent game-playing algorithms, in particular, decentralized V-learning and centralized VI-ULCB.