Research map / Graph learning / GNN architectures
Substructure Aware Graph Neural Networks
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 |
|---|---|---|---|
| add Ego subgraph information (ablation) | SAGIN · MUTAG | accuracy: 94.68±4.38 -> 95.23±2.99 | Helped |
| add Cut subgraph information (ablation) | SAGIN · MUTAG | accuracy: 94.12±3.09 -> 95.23±2.99 | Helped |
| add global graph information (ablation) | SAGIN · MUTAG | accuracy: 93.10±5.62 -> 95.23±2.99 | Helped |
| add Ego subgraph information (ablation) | SAGIN · ZINC | MAE: 0.094±0.002 -> 0.072±0.002 | Helped |
| add Cut subgraph information (ablation) | SAGIN · ZINC | MAE: 0.081±0.002 -> 0.072±0.002 | Helped |
| add global graph information (ablation) | SAGIN · ZINC | MAE: 0.079±0.001 -> 0.072±0.002 | Helped |
| add Ego subgraph information (ablation) | SAGIN · ZINC-FULL | MAE: 0.028±0.001 -> 0.016±0.002 | Helped |
| add Cut subgraph information (ablation) | SAGIN · ZINC-FULL | MAE: 0.031±0.001 -> 0.016±0.002 | Helped |
| add global graph information (ablation) | SAGIN · ZINC-FULL | MAE: 0.019±0.001 -> 0.016±0.002 | Helped |
| add Ego subgraph information (ablation) | SAGIN · MolHIV | ROC: 76.97±2.26 -> 80.64±1.28 | Helped |
| add Cut subgraph information (ablation) | SAGIN · MolHIV | ROC: 77.78±1.92 -> 80.64±1.28 | Helped |
| add global graph information (ablation) | SAGIN · MolHIV | ROC: 80.13±0.68 -> 80.64±1.28 | Helped |
Closest papers by meaning
- Substructure Assembling Network for Graph Classification (AAAI 2018 · similarity 0.92)
- Structure-Aware Transformer for Graph Representation Learning (ICML 2022 · similarity 0.91)
- Union Subgraph Neural Networks (AAAI 2024 · similarity 0.90)
- A Flexible, Equivariant Framework for Subgraph GNNs via Graph Products and Graph Coarsening (NeurIPS 2024 · similarity 0.90)
- Translating Subgraphs to Nodes Makes Simple GNNs Strong and Efficient for Subgraph Representation Learning (ICML 2024 · similarity 0.89)
- Subgraphormer: Unifying Subgraph GNNs and Graph Transformers via Graph Products (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.