You Can't Count on Luck: Why Decision Transformers and RvS Fail in Stochastic Environments
This addresses a critical limitation in offline reinforcement learning for stochastic domains, offering a practical improvement over existing methods.
The paper tackles the failure of Decision Transformers and RvS methods in stochastic environments by proposing ESPER, which clusters trajectories and conditions on average cluster returns to improve alignment between target and achieved returns, achieving significantly better performance and higher maximum scores than baselines in tasks like 2048 and Connect Four.
Recently, methods such as Decision Transformer that reduce reinforcement learning to a prediction task and solve it via supervised learning (RvS) have become popular due to their simplicity, robustness to hyperparameters, and strong overall performance on offline RL tasks. However, simply conditioning a probabilistic model on a desired return and taking the predicted action can fail dramatically in stochastic environments since trajectories that result in a return may have only achieved that return due to luck. In this work, we describe the limitations of RvS approaches in stochastic environments and propose a solution. Rather than simply conditioning on the return of a single trajectory as is standard practice, our proposed method, ESPER, learns to cluster trajectories and conditions on average cluster returns, which are independent from environment stochasticity. Doing so allows ESPER to achieve strong alignment between target return and expected performance in real environments. We demonstrate this in several challenging stochastic offline-RL tasks including the challenging puzzle game 2048, and Connect Four playing against a stochastic opponent. In all tested domains, ESPER achieves significantly better alignment between the target return and achieved return than simply conditioning on returns. ESPER also achieves higher maximum performance than even the value-based baselines.