Confounding-Robust Policy Evaluation in Infinite-Horizon Reinforcement Learning

Abstract

Off-policy evaluation of sequential decision policies from observational data is necessary in applications of batch reinforcement learning such as education and healthcare. In such settings, however, unobserved variables confound observed actions, rendering exact evaluation of new policies impossible, i.e., unidentifiable. We develop a robust approach that estimates sharp bounds on the (unidentifiable) value of a given policy in an infinite-horizon problem given data from another policy with unobserved confounding, subject to a sensitivity model. We consider stationary or baseline unobserved confounding and compute bounds by optimizing over the set of all stationary state-occupancy ratios that agree with a new partially identified estimating equation and the sensitivity model. We prove convergence to the sharp bounds as we collect more confounded data. Although checking set membership is a linear program, the support function is given by a difficult nonconvex optimization problem. We develop approximations based on nonconvex projected gradient descent and demonstrate the resulting bounds empirically.

References in corpus (6)

• Causal Effect Inference with Deep Latent-Variable Models
• Safe Policy Improvement by Minimizing Robust Baseline Regret
• Deep Reinforcement Learning for Sepsis Treatment
• Consistent On-Line Off-Policy Evaluation
• Deconfounding Reinforcement Learning in Observational Settings
• Policy Evaluation with Latent Confounders via Optimal Balance

Cited by in corpus (6)

• Sharp Sensitivity Analysis for Inverse Propensity Weighting via Quantile Balancing
• Off-policy Evaluation in Infinite-Horizon Reinforcement Learning with Latent Confounders
• Instrumental Variable Value Iteration for Causal Offline Reinforcement Learning
• On Covariate Shift of Latent Confounders in Imitation and Reinforcement Learning
• Sensitivity Analysis for Unmeasured Confounding in Medical Product Development and Evaluation Using Real World Evidence
• Universal Off-Policy Evaluation