Non-Stationary Off-Policy Optimization
arXiv:2006.08236
Abstract
Off-policy learning is a framework for evaluating and optimizing policies without deploying them, from data collected by another policy. Real-world environments are typically non-stationary and the offline learned policies should adapt to these changes. To address this challenge, we study the novel problem of off-policy optimization in piecewise-stationary contextual bandits. Our proposed solution has two phases. In the offline learning phase, we partition logged data into categorical latent states and learn a near-optimal sub-policy for each state. In the online deployment phase, we adaptively switch between the learned sub-policies based on their performance. This approach is practical and analyzable, and we provide guarantees on both the quality of off-policy optimization and the regret during online deployment. To show the effectiveness of our approach, we compare it to state-of-the-art baselines on both synthetic and real-world datasets. Our approach outperforms methods that act only on observed context.
AISTATS 2021; 16 pages, 2 figures
References in corpus (7)
- Doubly Robust Policy Evaluation and Learning
- On Upper-Confidence Bound Policies for Non-Stationary Bandit Problems
- Efficient Contextual Bandits in Non-stationary Worlds
- Learning Contextual Bandits in a Non-stationary Environment
- Nearly Optimal Adaptive Procedure with Change Detection for Piecewise-Stationary Bandit
- Off-Policy Evaluation and Learning from Logged Bandit Feedback: Error Reduction via Surrogate Policy
- Offline Evaluation of Ranking Policies with Click Models