Research map / Graph learning / GNN architectures
On the Topology Awareness and Generalization Performance of Graph Neural Networks
Junwei Su, Chuan Wu
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 |
|---|---|---|---|
| choose training sets with smaller structural distance to the remaining nodes (ablation) | GCN · Cora | accuracy (%): 65.42 -> 71.40 | Improved |
| choose training sets with smaller structural distance to the remaining nodes (ablation) | GCNII · Cora | accuracy (%): 61.42 -> 72.18 | Improved |
| choose training sets with smaller structural distance to the remaining nodes (ablation) | GAT · Citeseer | accuracy (%): 56.78 -> 64.63 | Improved |
| select initial labeled set with coverage-based sampling (farthest-first k-center) instead of random | GCN · Cora (0.5% labels) | test accuracy (%): 37.08 -> 49.03 | Improved |
| select initial labeled set with coverage-based sampling instead of random | GCN · Cora (0.5% labels) | maximum discrepancy (lower is better): 13.23 -> 8.23 | Improved |
| select initial labeled set with coverage-based sampling instead of PageRank | GCN · CiteSeer (0.5% labels) | test accuracy (%): 42.27 -> 50.98 | Improved |
| select initial labeled set with coverage-based sampling instead of PageRank | GCNII · Cora (0.5% labels) | test accuracy (%): 34.04 -> 35.61 | Improved |
| select initial labeled set with coverage-based sampling instead of random | GCNII · CiteSeer (0.5% labels) | maximum discrepancy (lower is better): 17.32 -> 13.04 | Mixed |
Closest papers by meaning
- A Topological Perspective on Demystifying GNN-Based Link Prediction Performance (ICLR 2024 · similarity 0.90)
- The Underappreciated Power of Vision Models for Graph Structural Understanding (NeurIPS 2025 · similarity 0.90)
- Weisfeiler and Lehman Go Paths: Learning Topological Features via Path Complexes (AAAI 2024 · similarity 0.90)
- Generalization of Graph Neural Networks Is Robust to Model Mismatch (AAAI 2025 · similarity 0.90)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.89)
- Two Heads Are Better Than One: Boosting Graph Sparse Training via Semantic and Topological Awareness (ICML 2024 · similarity 0.89)
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.
