Off-Policy Confidence Interval Estimation with Confounded Markov Decision Process
This addresses the challenge of reliable policy evaluation in real-world applications like healthcare and technology where confounding is common, offering a robust method for uncertainty quantification.
The paper tackles the problem of constructing confidence intervals for a target policy's value in offline settings with confounded Markov decision processes, where unmeasured variables can bias results, and shows that the value is identifiable using auxiliary variables, developing an efficient estimator validated on simulated and real ridesharing data.
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.