Research map / Graph learning / Graph self-supervised learning
Learning Invariant Graph Representations for Out-of-Distribution Generalization
Haoyang Li, Ziwei Zhang, Xin Wang, Wenwu Zhu
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 |
|---|---|---|---|
| use GIL (latent-environment invariant learning) instead of ERM | GIN backbone · SP-Motif r_test=1/3, r_train=0.8 | accuracy (%): 37.84 -> 51.24 | Improved |
| use GIL instead of DIR | DIR · SP-Motif r_test=0.2, r_train=0.8 | accuracy (%): 39.92 -> 46.36 | Improved |
| use GIL instead of GSAT under strong spurious correlation | GSAT · SP-Motif r_test=0.2, r_train=0.9 | accuracy (%): 33.87 -> 35.80 | Improved |
| use GIL instead of ERM on real graphs | ERM · MOLHIV | ROC-AUC (%): 76.20 -> 79.08 | Improved |
| use GIL instead of ERM | ERM · MNIST-75sp | accuracy (%): 14.94 -> 21.94 | Improved |
| use GIL instead of DIR | DIR · Graph-SST2 | accuracy (%): 83.29 -> 83.44 | No clear effect |
| wrap GIL around stronger backbone CIN | CIN · MOLHIV | ROC-AUC (%): 80.94 -> 81.15 | Improved |
| wrap GIL around stronger backbone HIG | HIG · MOLHIV | ROC-AUC (%): 84.03 -> 84.23 | Improved |
| set number of inferred environments to ground truth (3) (ablation) | GIL · SP-Motif | accuracy (%) | Improved |
| use more than 2 inferred environments (ablation) | GIL · MNIST-75sp | accuracy (%) | Hurt |
Closest papers by meaning
- Disentangling Invariant Subgraph via Variance Contrastive Estimation under Distribution Shifts (ICML 2025 · similarity 0.91)
- Subgraph Invariant Learning Towards Large-Scale Graph Node Classification (AAAI 2025 · similarity 0.91)
- Dissecting the Failure of Invariant Learning on Graphs (NeurIPS 2024 · similarity 0.91)
- Quantifying Distributional Invariance in Causal Subgraph for IRM-Free Graph Generalization (NeurIPS 2025 · similarity 0.91)
- Graph Invariant Learning with Subgraph Co-mixup for Out-of-Distribution Generalization (AAAI 2024 · similarity 0.91)
- Graph Contrastive Invariant Learning from the Causal Perspective (AAAI 2024 · similarity 0.90)
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.
