Accountable Off-Policy Evaluation With Kernel Bellman Statistics
arXiv:2008.06668
Abstract
We consider off-policy evaluation (OPE), which evaluates the performance of a new policy from observed data collected from previous experiments, without requiring the execution of the new policy. This finds important applications in areas with high execution cost or safety concerns, such as medical diagnosis, recommendation systems and robotics. In practice, due to the limited information from off-policy data, it is highly desirable to construct rigorous confidence intervals, not just point estimation, for the policy performance. In this work, we propose a new variational framework which reduces the problem of calculating tight confidence bounds in OPE into an optimization problem on a feasible set that catches the true state-action value function with high probability. The feasible set is constructed by leveraging statistical properties of a recently proposed kernel Bellman loss (Feng et al., 2019). We design an efficient computational approach for calculating our bounds, and extend it to perform post-hoc diagnosis and correction for existing estimators. Empirical results show that our method yields tight confidence intervals in different settings.
22 pages, 4 figures, ICML 2020
References in corpus (9)
- Unbiased Offline Evaluation of Contextual-bandit-based News Article Recommendation Algorithms
- Provably Efficient Reinforcement Learning with Linear Function Approximation
- Breaking the Curse of Horizon: Infinite-Horizon Off-Policy Estimation
- GenDICE: Generalized Offline Estimation of Stationary Values
- Modelling transition dynamics in MDPs with RKHS embeddings
- A Kernel Loss for Solving the Bellman Equation
- Minimax Value Interval for Off-Policy Evaluation and Policy Optimization
- Black-box Off-policy Estimation for Infinite-Horizon Reinforcement Learning
- Off-Policy Estimation of Long-Term Average Outcomes with Applications to Mobile Health