Fairness-aware Cross-Domain Recommendation

TL;DR

This paper introduces FairCDR, a fairness-aware model for cross-domain recommendation that improves user fairness by leveraging non-overlapping data and influence-based reweighting, enhancing both fairness and accuracy.

Contribution

The paper proposes a novel fairness-aware mapping function for cross-domain recommendation, utilizing influence functions to balance fairness and accuracy with limited overlapping data.

Findings

FairCDR improves user fairness in cross-domain recommendation.

The influence-based reweighing reduces unfairness effectively.

Extensive experiments validate the model's effectiveness.

Abstract

Cross-Domain Recommendation (CDR) is an effective way to alleviate the cold-start problem. However, previous work severely ignores fairness and bias when learning the mapping function, which is used to obtain the representations for fresh users in the target domain. To study this problem, in this paper, we propose a Fairness-aware Cross-Domain Recommendation model, called FairCDR. Our method achieves user-oriented group fairness by learning the fairness-aware mapping function. Since the overlapping data are quite limited and distributionally biased, FairCDR leverages abundant non-overlapping users and interactions to help alleviate these problems. Considering that each individual has different influence on model fairness, we propose a new reweighing method based on Influence Function (IF) to reduce unfairness while maintaining recommendation accuracy. Extensive experiments are conducted to demonstrate the effectiveness of our model.