Prompting Decision Transformer for Few-Shot Policy Generalization

Prompt-DT is a strong few-shot learner w/o any extra finetuning on unseen target tasks.

Proposes a Prompt-based Decision Transformer (Prompt-DT), which leverages the sequential modeling ability of the Transformer architecture and the prompt framework to achieve few-shot adaptation in offline RL. It outperforms its variants and strong meta offline RL baselines by a large margin with a trajectory prompt containing only a few timesteps. It is robust to prompt length changes and can generalize to out-of-distribution (OOD) environments.