Multi-Game Decision Transformers

TL;DR

This paper demonstrates that a single transformer model trained offline can learn to play up to 46 Atari games at near-human performance, showing scalability and adaptability similar to language and vision models.

Contribution

It introduces a multi-game decision transformer that learns multiple Atari games simultaneously with a single set of weights, advancing generalist reinforcement learning.

Findings

Single transformer can play multiple Atari games at near-human levels

Performance scales with model size and improves with fine-tuning on new games

Multi-game models outperform other RL and behavioral cloning approaches

Abstract

A longstanding goal of the field of AI is a method for learning a highly capable, generalist agent from diverse experience. In the subfields of vision and language, this was largely achieved by scaling up transformer-based models and training them on large, diverse datasets. Motivated by this progress, we investigate whether the same strategy can be used to produce generalist reinforcement learning agents. Specifically, we show that a single transformer-based model - with a single set of weights - trained purely offline can play a suite of up to 46 Atari games simultaneously at close-to-human performance. When trained and evaluated appropriately, we find that the same trends observed in language and vision hold, including scaling of performance with model size and rapid adaptation to new games via fine-tuning. We compare several approaches in this multi-game setting, such as online and offline RL methods and behavioral cloning, and find that our Multi-Game Decision Transformer models offer the best scalability and performance. We release the pre-trained models and code to encourage further research in this direction.