A General Offline Reinforcement Learning Framework for Interactive Recommendation

No. 5: AAAI-21 Technical Tracks 5

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 35

AAAI Technical Track on Data Mining and Knowledge Management

This paper studies the problem of learning interactive recommender systems from logged feedbacks without any exploration in online environments. We address the problem by proposing a general offline reinforcement learning framework for recommendation, which enables maximizing cumulative user rewards without online exploration. Specifically, we first introduce a probabilistic generative model for interactive recommendation, and then propose an effective inference algorithm for discrete and stochastic policy learning based on logged feedbacks. In order to perform offline learning more effectively, we propose five approaches to minimize the distribution mismatch between the logging policy and recommendation policy: support constraints, supervised regularization, policy constraints, dual constraints and reward extrapolation. We conduct extensive experiments on two public real-world datasets, demonstrating that the proposed methods can achieve superior performance over existing supervised learning and reinforcement learning methods for recommendation.

10.1609/aaai.v35i5.16579