Decision Transformer: Reinforcement Learning via Sequence Modeling

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.

PDF Abstract NeurIPS 2021 PDF NeurIPS 2021 Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Atari Games Atari 2600 Breakout DT Score 267.5 # 44
Atari Games Atari 2600 Pong DT Score 17.1 # 43
Atari Games Atari 2600 Q*Bert DT Score 25.1 # 56
Atari Games Atari 2600 Seaquest DT Score 2.4 # 56
Offline RL D4RL Decision Transformer (DT) Average Reward 73.5 # 3
D4RL D4RL Decision Transformer (DT) Average Reward 72.2 # 5