More Efficient Off-Policy Evaluation through Regularized Targeted Learning
This work addresses the problem of accurately estimating policy performance from historical data for researchers and practitioners in reinforcement learning, representing a strong specific gain rather than a foundational advancement.
The paper tackles off-policy evaluation in reinforcement learning by introducing a novel doubly-robust estimator based on Targeted Maximum Likelihood Estimation, which uniformly outperforms existing methods across multiple environments and model misspecification levels, achieving impressive performance gains.
We study the problem of off-policy evaluation (OPE) in Reinforcement Learning (RL), where the aim is to estimate the performance of a new policy given historical data that may have been generated by a different policy, or policies. In particular, we introduce a novel doubly-robust estimator for the OPE problem in RL, based on the Targeted Maximum Likelihood Estimation principle from the statistical causal inference literature. We also introduce several variance reduction techniques that lead to impressive performance gains in off-policy evaluation. We show empirically that our estimator uniformly wins over existing off-policy evaluation methods across multiple RL environments and various levels of model misspecification. Finally, we further the existing theoretical analysis of estimators for the RL off-policy estimation problem by showing their $O_P(1/\sqrt{n})$ rate of convergence and characterizing their asymptotic distribution.