Recommendations as Treatments: Debiasing Learning and Evaluation
This work addresses biases in recommender systems for users and platforms, offering a practical and scalable solution.
The paper tackles the problem of selection biases in recommender system data by adapting causal inference models, resulting in unbiased performance estimators and a matrix factorization method that significantly improves prediction performance on real-world data.
Most data for evaluating and training recommender systems is subject to selection biases, either through self-selection by the users or through the actions of the recommendation system itself. In this paper, we provide a principled approach to handling selection biases, adapting models and estimation techniques from causal inference. The approach leads to unbiased performance estimators despite biased data, and to a matrix factorization method that provides substantially improved prediction performance on real-world data. We theoretically and empirically characterize the robustness of the approach, finding that it is highly practical and scalable.