Improving Recommendation Fairness via Graph Structure and Representation Augmentation

TL;DR

This paper proposes a dual data augmentation framework for fair recommendation systems using graph structures and representation learning, effectively reducing bias while maintaining utility.

Contribution

It introduces a novel data augmentation approach based on prior hypotheses to improve fairness in GCN-based recommendation models.

Findings

Outperforms existing fairness methods on real-world datasets.

Effectively reduces bias without sacrificing recommendation utility.

Demonstrates the effectiveness of dual augmentation strategies.

Abstract

Graph Convolutional Networks (GCNs) have become increasingly popular in recommendation systems. However, recent studies have shown that GCN-based models will cause sensitive information to disseminate widely in the graph structure, amplifying data bias and raising fairness concerns. While various fairness methods have been proposed, most of them neglect the impact of biased data on representation learning, which results in limited fairness improvement. Moreover, some studies have focused on constructing fair and balanced data distributions through data augmentation, but these methods significantly reduce utility due to disruption of user preferences. In this paper, we aim to design a fair recommendation method from the perspective of data augmentation to improve fairness while preserving recommendation utility. To achieve fairness-aware data augmentation with minimal disruption to user preferences, we propose two prior hypotheses. The first hypothesis identifies sensitive interactions by comparing outcomes of performance-oriented and fairness-aware recommendations, while the second one focuses on detecting sensitive features by analyzing feature similarities between biased and debiased representations. Then, we propose a dual data augmentation framework for fair recommendation, which includes two data augmentation strategies to generate fair augmented graphs and feature representations. Furthermore, we introduce a debiasing learning method that minimizes the dependence between the learned representations and sensitive information to eliminate bias. Extensive experiments on two real-world datasets demonstrate the superiority of our proposed framework.