Research map / Graph learning / GNN architectures
Subgroup Generalization and Fairness of Graph Neural Networks
Jiaqi Ma, Junwei Deng, Qiaozhu Mei
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Evaluate test nodes in subgroups with increasing aggregated-feature distance from the training set | GCN, GAT, SGC, and APPNP · Cora, Citeseer, and Pubmed | test accuracy | Mixed |
| Evaluate test nodes in subgroups with increasing shortest-path distance from the training set | GNN models · Cora, Citeseer, and Pubmed | test accuracy | Improved |
| Group test nodes by centrality rather than distance to the training set | GNN models · Cora, Citeseer, and Pubmed | test accuracy | No clear effect |
| Evaluate MLP test nodes in subgroups with increasing aggregated-feature or geodesic distance from training nodes | MLP · Cora, Citeseer, and Pubmed | test accuracy | Mixed |
| Add independent noise to node features to reduce feature homophily (ablation) | MLP · Cora, Citeseer, and Pubmed | test accuracy across aggregated-feature-distance subgroups | Mixed |
| Add independent noise to node features to reduce feature homophily (ablation) | GCN, GAT, SGC, and APPNP · Cora, Citeseer, and Pubmed | test accuracy across aggregated-feature-distance subgroups | No clear effect |
| Select training nodes with high centrality for one dominant class and low centrality for other classes (ablation) | GNN models · Cora, Citeseer, and Pubmed | false positive rate for the dominant class | Hurt |
Closest papers by meaning
- Chasing Fairness in Graphs: A GNN Architecture Perspective (AAAI 2024 · similarity 0.90)
- Fair Graph U-Net: A Fair Graph Learning Framework Integrating Group and Individual Awareness (AAAI 2025 · similarity 0.89)
- FairSIN: Achieving Fairness in Graph Neural Networks through Sensitive Information Neutralization (AAAI 2024 · similarity 0.88)
- Generalization of Graph Neural Networks Is Robust to Model Mismatch (AAAI 2025 · similarity 0.88)
- Making Classic GNNs Strong Baselines Across Varying Homophily: A Smoothness–Generalization Perspective (NeurIPS 2025 · similarity 0.88)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
