Fair Node Representation Learning via Adaptive Data Augmentation
This addresses fairness issues in graph-based machine learning, particularly for underrepresented groups, by providing a novel augmentation approach that is incremental but applicable to various GNN mechanisms.
The paper tackled bias in node representation learning from Graph Neural Networks by theoretically analyzing sources of bias in nodal features and graph structure, and developed fairness-aware data augmentation frameworks to reduce it, with experiments showing improved fairness in statistical parity and equal opportunity while maintaining comparable utility to state-of-the-art methods.
Node representation learning has demonstrated its efficacy for various applications on graphs, which leads to increasing attention towards the area. However, fairness is a largely under-explored territory within the field, which may lead to biased results towards underrepresented groups in ensuing tasks. To this end, this work theoretically explains the sources of bias in node representations obtained via Graph Neural Networks (GNNs). Our analysis reveals that both nodal features and graph structure lead to bias in the obtained representations. Building upon the analysis, fairness-aware data augmentation frameworks on nodal features and graph structure are developed to reduce the intrinsic bias. Our analysis and proposed schemes can be readily employed to enhance the fairness of various GNN-based learning mechanisms. Extensive experiments on node classification and link prediction are carried out over real networks in the context of graph contrastive learning. Comparison with multiple benchmarks demonstrates that the proposed augmentation strategies can improve fairness in terms of statistical parity and equal opportunity, while providing comparable utility to state-of-the-art contrastive methods.