![Theory of Reinforcement Learning_hi-res logo](/sites/default/files/styles/workshop_banner_sm_1x/public/2023-05/Theory%20of%20Reinforcement%20Learning_hi-res.png.jpg?itok=LJoCeC_d)
Abstract
In this talk, we consider high-confidence behavior-agnostic off-policy evaluation in reinforcement learning, where the goal is to estimate a confidence interval on a target policy’s value, given only access to a static experience dataset collected by unknown behavior policies. Starting from a function space embedding of the linear program formulation of the Q-function, we obtain an optimization problem with generalized estimating equation constraints. By applying the generalized empirical likelihood method to the resulting Lagrangian, we propose CoinDICE, a novel and efficient algorithm for computing confidence intervals. Theoretically, we prove the obtained confidence intervals are valid, in both asymptotic and finite-sample regimes. Empirically, we show in a variety of benchmarks that the confidence interval estimates are tighter and more accurate than existing methods.
This is a joint work with Ofir Nachum, Yinlam Chow, Lihong Li, Csaba Szepesvari, and Dale Schuurmans.