Research map / Graph learning / GNN architectures
Demystifying Structural Disparity in Graph Neural Networks: Can One Size Fit All?
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Test on minority-pattern nodes instead of an i.i.d. random split | GCN · PubMed | accuracy: 89.18±0.15 -> 51.04±0.16 | Hurt |
| Test on minority-pattern nodes instead of an i.i.d. random split | GCN · Ogbn-Arxiv | accuracy: 72.99±0.14 -> 34.94±0.07 | Hurt |
| Test on minority-pattern nodes instead of an i.i.d. random split | GCN · Squirrel | accuracy: 58.09±0.71 -> 32.13±4.93 | Hurt |
| Test on minority-pattern nodes instead of an i.i.d. random split | GCN · Chameleon | accuracy: 75.09±0.79 -> 43.35±3.47 | Hurt |
| Use GCNII rather than GCN | GCN node classification · Ogbn-Arxiv (proposed OOD split) | accuracy: 34.94±0.07 -> 36.81±0.14 | Helped |
| Use GCNII rather than GCN | GCN node classification · Squirrel (proposed OOD split) | accuracy: 32.13±4.93 -> 37.15±1.39 | Helped |
| Use GCNII rather than GCN | GCN node classification · PubMed (proposed OOD split) | accuracy: 51.04±0.16 -> 67.76±0.36 | Helped |
| Use GCNII rather than GCN | GCN node classification · Chameleon (proposed OOD split) | accuracy: 43.35±3.47 -> 41.25±2.03 | Hurt |
| Replace EERM's GCN encoder with GCNII | EERM · PubMed (proposed OOD split) | accuracy: 65.37±1.35 -> 67.59±0.91 | Helped |
| Replace EERM's GCN encoder with GCNII | EERM · Ogbn-Arxiv (proposed OOD split) | accuracy: 34.23±0.46 -> 40.28±0.84 | Helped |
| Replace EERM's GCN encoder with GCNII | EERM · Squirrel (proposed OOD split) | accuracy: 40.93±0.57 -> 44.31±0.40 | Helped |
| Replace EERM's GCN encoder with GCNII | EERM · Chameleon (proposed OOD split) | accuracy: 45.84±1.05 -> 48.59±0.78 | Helped |
Closest papers by meaning
- When Do Graph Neural Networks Help with Node Classification? Investigating the Homophily Principle on Node Distinguishability (NeurIPS 2023 · similarity 0.93)
- What Is Missing For Graph Homophily? Disentangling Graph Homophily For Graph Neural Networks (NeurIPS 2024 · similarity 0.93)
- On the Interplay between Graph Structure and Learning Algorithms in Graph Neural Networks (ICML 2025 · similarity 0.92)
- Making Classic GNNs Strong Baselines Across Varying Homophily: A Smoothness–Generalization Perspective (NeurIPS 2025 · similarity 0.91)
- On the Impact of Feature Heterophily on Link Prediction with Graph Neural Networks (NeurIPS 2024 · similarity 0.91)
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.91)
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.