Unsupervised Domain Adaptation Meets Offline Recommender Learning

To construct a well-performing recommender offline, eliminating selection biases of the rating feedback is critical. A current promising solution to the challenge is the causality approach using the propensity scoring method. However, the performance of existing propensity-based algorithms can be significantly affected by the propensity estimation bias. To alleviate the problem, we formulate the missing-not-at-random recommendation as the unsupervised domain adaptation problem and drive the propensity-agnostic generalization error bound. We further propose a corresponding algorithm minimizing the bound via adversarial learning. Empirical evaluation using the Yahoo! R3 dataset demonstrates the effectiveness and the real-world applicability of the proposed approach.