Decision Transformer: Reinforcement Learning via Sequence Modeling

TL;DR

This paper presents Decision Transformer, a novel approach that models reinforcement learning as a sequence prediction task using Transformers, enabling effective offline RL without traditional value functions or policy gradients.

Contribution

It introduces a Transformer-based architecture for RL that conditions on desired return, past states, and actions, simplifying the process and improving performance over existing methods.

Findings

Matches or exceeds state-of-the-art offline RL performance

Effective on Atari, OpenAI Gym, and Key-to-Door tasks

Simplifies RL by framing it as sequence modeling

Abstract

We introduce a framework that abstracts Reinforcement Learning (RL) as a sequence modeling problem. This allows us to draw upon the simplicity and scalability of the Transformer architecture, and associated advances in language modeling such as GPT-x and BERT. In particular, we present Decision Transformer, an architecture that casts the problem of RL as conditional sequence modeling. Unlike prior approaches to RL that fit value functions or compute policy gradients, Decision Transformer simply outputs the optimal actions by leveraging a causally masked Transformer. By conditioning an autoregressive model on the desired return (reward), past states, and actions, our Decision Transformer model can generate future actions that achieve the desired return. Despite its simplicity, Decision Transformer matches or exceeds the performance of state-of-the-art model-free offline RL baselines on Atari, OpenAI Gym, and Key-to-Door tasks.