Off-policy evaluation for MDPs with unknown structure

TL;DR

This paper introduces G-SCOPE, an efficient off-policy evaluation algorithm for MDPs with unknown factored structure, enabling reliable policy comparison without direct testing.

Contribution

The paper proposes G-SCOPE, a novel algorithm that exploits factored environment dynamics for efficient off-policy evaluation in high-dimensional MDPs.

Findings

G-SCOPE is computationally and sample efficient.

The algorithm scales well on high-dimensional problems.

Finite sample analysis supports its effectiveness.

Abstract

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.