Research map / Graph learning / GNN architectures
Transfer Learning of Graph Neural Networks with Ego-graph Information Maximization
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 |
|---|---|---|---|
| train on Barabasi source graph instead of Forest-fire source (larger structural difference) (ablation) | EGI (GIN) · synthetic role identification, transferable features | accuracy: 0.710 -> 0.616 | Hurt |
| use non-transferable uniform node features instead of degree one-hot (ablation) | EGI (GIN) · synthetic role identification F-F | accuracy: 0.710 -> 0.376 | Hurt |
| pre-train EGI on Europe instead of using random-init GIN | GIN · Airport USA | accuracy: 0.616 -> 0.646 | Helped |
| pre-train DGI instead of random-init GIN | GIN · Airport USA | accuracy: 0.616 -> 0.549 | Hurt |
| EGI instead of GMI | GIN encoder · Airport Brazil | accuracy: 0.731 -> 0.732 | No clear effect |
| EGI instead of GMI in joint fine-tuning | GCN · YAGO 20-shot relation prediction | AUROC: 0.721 -> 0.787 | Helped |
| EGI instead of GMI in post-fine-tuning | GCN · YAGO 20-shot relation prediction | AUROC: 0.728 -> 0.739 | Helped |
| structural difference between source and target graphs above 1 | EGI transfer · Gene brain cancer networks | micro-F1 gain over untrained GIN | Hurt |
Closest papers by meaning
- Learning to Pre-train Graph Neural Networks (AAAI 2021 · similarity 0.90)
- Contextual Structure Knowledge Transfer for Graph Neural Networks (AAAI 2025 · similarity 0.89)
- How Much Can Transfer? BRIDGE: Bounded Multi-Domain Graph Foundation Model with Generalization Guarantees (ICML 2025 · similarity 0.89)
- Towards Pre-trained Graph Condensation via Optimal Transport (NeurIPS 2025 · similarity 0.88)
- Improving Graph Neural Networks by Learning Continuous Edge Directions (ICLR 2025 · similarity 0.88)
- Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective (ICLR 2025 · similarity 0.88)
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.