FairSIN: Achieving Fairness in Graph Neural Networks through Sensitive Information Neutralization

2024-03-19

AAAI Conference on Artificial Intelligence

AI Summary

Graph Neural Networks (GNNs) often exhibit prediction bias stemming from sensitive attributes (e.g., gender, race). Existing fairness-aware methods—primarily based on filtering (e.g., edge pruning, feature masking)—frequently degrade non-sensitive information, resulting in suboptimal trade-offs between accuracy and fairness. To address this, we propose a novel “sensitive information neutralization” paradigm: before message passing, fairness-promoting features (F3) are injected to actively neutralize sensitive bias in node representations. We theoretically prove that F3 can be realized via aggregation over heterogeneous neighbors—those differing in sensitive attributes—and provide three plug-and-play implementations grounded in both data- and model-level perspectives. Our approach is architecture-agnostic, compatible with mainstream GNN backbones including GCN, GAT, and GIN. Extensive experiments on five benchmark datasets demonstrate that our method reduces the average demographic disparity (ADG) by up to 42% while incurring less than a 0.8% drop in accuracy—significantly outperforming filtering-based baselines.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Computer Vision: Bias, Fairness & Privacy

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Despite the remarkable success of graph neural networks (GNNs) in modeling graph-structured data, like other machine learning models, GNNs are also susceptible to making biased predictions based on sensitive attributes, such as race and gender. For fairness consideration, recent state-of-the-art (SOTA) methods propose to filter out sensitive information from inputs or representations, e.g., edge dropping or feature masking. However, we argue that such filtering-based strategies may also filter out some non-sensitive feature information, leading to a sub-optimal trade-off between predictive performance and fairness. To address this issue, we unveil an innovative neutralization-based paradigm, where additional Fairness-facilitating Features (F3) are incorporated into node features or representations before message passing. The F3 are expected to statistically neutralize the sensitive bias in node representations and provide additional nonsensitive information. We also provide theoretical explanations for our rationale, concluding that F3 can be realized by emphasizing the features of each node’s heterogeneous neighbors (neighbors with different sensitive attributes). We name our method as FairSIN, and present three implementation variants from both data-centric and model-centric perspectives. Experimental results on five benchmark datasets with three different GNN backbones show that FairSIN significantly improves fairness metrics while maintaining high prediction accuracies. Codes and appendix can be found at https://github.com/BUPT-GAMMA/FariSIN.

Problem

• Addressing biased predictions in GNNs from sensitive attributes
• Improving fairness without sacrificing predictive performance through neutralization
• Mitigating sensitive information bias while preserving non-sensitive features

Innovation

• Neutralization-based paradigm for bias reduction
• Incorporating Fairness-facilitating Features before message passing
• Emphasizing heterogeneous neighbors' features statistically