Black-box Off-policy Estimation for Infinite-Horizon Reinforcement Learning
Abstract
Off-policy estimation for long-horizon problems is important in many real-life applications such as healthcare and robotics, where high-fidelity simulators may not be available and on-policy evaluation is expensive or impossible. Recently, \citet{liu18breaking} proposed an approach that avoids the \emph{curse of horizon} suffered by typical importance-sampling-based methods, but are limited in practice as it requires that data be collected by a \emph{single} and \emph{known} behavior policy. In this work, we propose a novel approach that eliminates such limitations. In particular, we formulate the problem as one of solving for the fixed point of a ``backward flow'' operator, the solution of which gives the desired importance ratios of stationary distributions between the target and behavior policies. Experiments on benchmarks verify the effectiveness of the approach.