Balanced Neighborhoods for Fairness-Aware Collaborative Recommendation

Como, Italy

Recent work on fairness in machine learning has begun to be extended to recommender systems. While there is a tension between the goals of fairness and of personalization, there are contexts in which a global evaluations of outcomes is possible and where equity across such outcomes is a desirable goal. In this paper, we introduce the concept of a balanced neighborhood as a mechanism to preserve personalization in recommendation while enhancing the fairness of recommendation outcomes.We show that a modified version of the SLIM algorithm can be used to improve the balance of user neighborhoods, with the result of achieving greater outcome fairness in a real-world dataset with minimal loss in ranking performance.