Robust Batch Policy Learning in Markov Decision Processes
This work addresses the problem of robust policy learning from pre-collected data for researchers and practitioners in reinforcement learning, representing an incremental improvement with a novel method for a known bottleneck.
The paper tackles the offline data-driven sequential decision-making problem in Markov decision processes by proposing a method to learn a robust policy that maximizes the smallest average reward across distributions centered at the policy-induced stationary distribution, achieving a rate-optimal regret bound up to a logarithmic factor in terms of total decision points.
We study the offline data-driven sequential decision making problem in the framework of Markov decision process (MDP). In order to enhance the generalizability and adaptivity of the learned policy, we propose to evaluate each policy by a set of the average rewards with respect to distributions centered at the policy induced stationary distribution. Given a pre-collected dataset of multiple trajectories generated by some behavior policy, our goal is to learn a robust policy in a pre-specified policy class that can maximize the smallest value of this set. Leveraging the theory of semi-parametric statistics, we develop a statistically efficient policy learning method for estimating the de ned robust optimal policy. A rate-optimal regret bound up to a logarithmic factor is established in terms of total decision points in the dataset.