Future-conditioned Unsupervised Pretraining for Decision Transformer

TL;DR

This paper introduces Pretrained Decision Transformer (PDT), a novel unsupervised pretraining method for offline reinforcement learning that uses future trajectory information to improve decision-making and generalization from reward-free data.

Contribution

The paper proposes PDT, a simple yet effective approach for unsupervised RL pretraining that leverages future trajectory information, enabling better generalization and sub-optimal data utilization.

Findings

PDT outperforms supervised pretraining on sub-optimal data.

PDT can extract diverse behaviors from offline data.

PDT allows controllable sampling of high-return behaviors during finetuning.

Abstract

Recent research in offline reinforcement learning (RL) has demonstrated that return-conditioned supervised learning is a powerful paradigm for decision-making problems. While promising, return conditioning is limited to training data labeled with rewards and therefore faces challenges in learning from unsupervised data. In this work, we aim to utilize generalized future conditioning to enable efficient unsupervised pretraining from reward-free and sub-optimal offline data. We propose Pretrained Decision Transformer (PDT), a conceptually simple approach for unsupervised RL pretraining. PDT leverages future trajectory information as a privileged context to predict actions during training. The ability to make decisions based on both present and future factors enhances PDT's capability for generalization. Besides, this feature can be easily incorporated into a return-conditioned framework for online finetuning, by assigning return values to possible futures and sampling future embeddings based on their respective values. Empirically, PDT outperforms or performs on par with its supervised pretraining counterpart, especially when dealing with sub-optimal data. Further analysis reveals that PDT can extract diverse behaviors from offline data and controllably sample high-return behaviors by online finetuning. Code is available at here.