Off-policy evaluation for MDPs with unknown structure
This addresses the challenge of proving policy superiority in dynamic decision problems for reinforcement learning practitioners, but appears incremental as it builds on existing off-policy evaluation methods.
The paper tackles the problem of evaluating a new policy in Markov Decision Processes without testing it, by introducing the G-SCOPE algorithm that uses data from an existing policy, and shows it is computationally and sample efficient with good scaling on high-dimensional problems.
Off-policy learning in dynamic decision problems is essential for providing strong evidence that a new policy is better than the one in use. But how can we prove superiority without testing the new policy? To answer this question, we introduce the G-SCOPE algorithm that evaluates a new policy based on data generated by the existing policy. Our algorithm is both computationally and sample efficient because it greedily learns to exploit factored structure in the dynamics of the environment. We present a finite sample analysis of our approach and show through experiments that the algorithm scales well on high-dimensional problems with few samples.