EqGNN: Equalized Node Opportunity in Graphs

TL;DR

This paper introduces EqGNN, a novel graph neural network framework designed to optimize for equalized odds fairness, effectively reducing bias related to sensitive attributes while maintaining high predictive utility.

Contribution

It is the first to optimize GNNs specifically for the equalized odds fairness criterion, combining a GNN classifier, a sampler, and a discriminator with a novel permutation loss.

Findings

Achieves state-of-the-art fairness results on multiple datasets.

Effectively reduces sensitive attribute information in node representations.

Maintains high prediction accuracy while enforcing fairness.

Abstract

Graph neural networks (GNNs), has been widely used for supervised learning tasks in graphs reaching state-of-the-art results. However, little work was dedicated to creating unbiased GNNs, i.e., where the classification is uncorrelated with sensitive attributes, such as race or gender. Some ignore the sensitive attributes or optimize for the criteria of statistical parity for fairness. However, it has been shown that neither approaches ensure fairness, but rather cripple the utility of the prediction task. In this work, we present a GNN framework that allows optimizing representations for the notion of Equalized Odds fairness criteria. The architecture is composed of three components: (1) a GNN classifier predicting the utility class, (2) a sampler learning the distribution of the sensitive attributes of the nodes given their labels. It generates samples fed into a (3) discriminator that discriminates between true and sampled sensitive attributes using a novel "permutation loss" function. Using these components, we train a model to neglect information regarding the sensitive attribute only with respect to its label. To the best of our knowledge, we are the first to optimize GNNs for the equalized odds criteria. We evaluate our classifier over several graph datasets and sensitive attributes and show our algorithm reaches state-of-the-art results.