A General Offline Reinforcement Learning Framework for Interactive Recommendation
This addresses the challenge of building effective recommender systems without costly online exploration, which is incremental as it builds on existing offline RL techniques.
The paper tackles the problem of learning interactive recommender systems from logged feedback without online exploration by proposing a general offline reinforcement learning framework, achieving superior performance over existing methods in experiments on two real-world datasets.
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.