Off-Policy Confidence Interval Estimation with Confounded Markov Decision Process

TL;DR

This paper develops a method to accurately estimate confidence intervals for policy values in offline reinforcement learning settings where unmeasured confounders exist, using auxiliary variables to ensure identifiability.

Contribution

It introduces a novel approach for off-policy value estimation in confounded Markov decision processes, addressing a key gap in existing methods.

Findings

Method is robust to model misspecification

Provides rigorous uncertainty quantification

Validated on simulated and real ridesharing data

Abstract

This paper is concerned with constructing a confidence interval for a target policy's value offline based on a pre-collected observational data in infinite horizon settings. Most of the existing works assume no unmeasured variables exist that confound the observed actions. This assumption, however, is likely to be violated in real applications such as healthcare and technological industries. In this paper, we show that with some auxiliary variables that mediate the effect of actions on the system dynamics, the target policy's value is identifiable in a confounded Markov decision process. Based on this result, we develop an efficient off-policy value estimator that is robust to potential model misspecification and provide rigorous uncertainty quantification. Our method is justified by theoretical results, simulated and real datasets obtained from ridesharing companies. A Python implementation of the proposed procedure is available at https://github.com/Mamba413/cope.