Research map / Graph learning / Graph self-supervised learning
Generalizing Graph Neural Networks on Out-of-Distribution Graphs
Shaohua Fan, Xiao Wang, Chuan Shi, Peng Cui, Bai Wang
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 |
|---|---|---|---|
| add StableGNN framework (high-level variable learning + CVD) to GraphSAGE | GraphSAGE · Synthetic house-motif (mu=0.9) | accuracy: 65.2 -> 70.35 | Improved |
| add StableGNN framework to GCN | GCN · Synthetic house-motif (mu=0.9) | accuracy: 69.68 -> 76.35 | Improved |
| increase spurious correlation degree mu from 0.6 to 0.9 | GraphSAGE · Synthetic house-motif | F1: 67.83 -> 54.24 | Hurt |
| add StableGNN to GraphSAGE | GraphSAGE · Molhiv | ROC-AUC: 75.78 -> 77.63 | Improved |
| add StableGNN to GraphSAGE | GraphSAGE · Molesol | RMSE: 1.1852 -> 1.0092 | Improved |
| add StableGNN to GCN | GCN · Molesol | RMSE: 1.153 -> 0.9638 | Improved |
| add StableGNN to GraphSAGE | GraphSAGE · Moltox21 | ROC-AUC: 68.88 -> 69.14 | No clear effect |
| add CVD regularizer to StableGNN-NoCVD (ablation) | StableGNN-NoCVD · Molhiv | ROC-AUC: 76.47 -> 77.63 | Improved |
| add CVD regularizer to StableGNN-NoCVD (ablation) | StableGNN-NoCVD · Molesol | RMSE: 1.0819 -> 1.022 | Improved |
| use cluster-level CVD instead of per-dimension decorrelation (ablation) | StableGNN-NoCVD · Molhiv | ROC-AUC: 75.32 -> 77.63 | Improved |
| use sample reweighting CVD instead of HSIC disentanglement (ablation) | StableGNN-NoCVD · Molbbbp | ROC-AUC: 64.21 -> 68.47 | Improved |
| add per-dimension decorrelation to GraphSAGE (ablation) | GraphSAGE · Molbace | ROC-AUC: 78.51 -> 77.34 | Hurt |
Closest papers by meaning
- Out-of-Distribution Generalization on Graphs via Progressive Inference (AAAI 2025 · similarity 0.92)
- Rethinking Causal Relationships Learning in Graph Neural Networks (AAAI 2024 · similarity 0.92)
- Quantifying Distributional Invariance in Causal Subgraph for IRM-Free Graph Generalization (NeurIPS 2025 · similarity 0.90)
- A Recipe for Causal Graph Regression: Confounding Effects Revisited (ICML 2025 · similarity 0.90)
- Learning to Reweight for Generalizable Graph Neural Network (AAAI 2024 · similarity 0.90)
- Towards Bridging Generalization and Expressivity of Graph Neural Networks (ICLR 2025 · 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.
