Research map / Graph learning / GNN architectures
Understanding Attention and Generalization in 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 supervised attention to GIN with threshold pooling (ablation) | GIN · Colors Test-LargeC | accuracy (%): 26 -> 89 | Helped |
| add supervised attention to ChebyGIN (ablation) | ChebyGIN · Triangles Test-Large | accuracy (%): 30 -> 48 | Helped |
| add supervised attention to ChebyGIN (ablation) | ChebyGIN · MNIST-75sp Test-Noisy | accuracy (%): 80 -> 92.3 | Helped |
| add unsupervised attention pooling to ChebyGIN (ablation) | ChebyGIN · Colors Test-LargeC | accuracy (%): 15 -> 11 | Hurt |
| add unsupervised attention pooling to ChebyGIN (ablation) | ChebyGIN · Triangles Test-Large | accuracy (%): 30 -> 26 | Hurt |
| add weakly-supervised attention to ChebyGIN (ablation) | ChebyGIN · MNIST-75sp Test-Noisy | accuracy (%): 80 -> 88.8 | Helped |
| add weakly-supervised attention to ChebyGIN (ablation) | ChebyGIN · Colors Test-LargeC | accuracy (%): 11 -> 73 | Helped |
| replace top-k pooling with threshold pooling (supervised) | GIN · Colors Test-LargeC | accuracy (%): 28 -> 89 | Helped |
| replace top-k pooling with threshold pooling (supervised) | ChebyGIN · Triangles Test-Large | accuracy (%): 39 -> 48 | Helped |
| replace top-k pooling with threshold pooling (supervised) | ChebyGIN · MNIST-75sp Test-Noisy | accuracy (%): 90.6 -> 92.3 | Helped |
| weakly-supervised vs unsupervised attention (ablation) | GCN/ChebyNet · D&D-200 | accuracy (%): 51.9 -> 54.3 | Helped |
| weakly-supervised vs unsupervised attention (ablation) | GCN/ChebyNet · Proteins-25 | accuracy (%): 75.6 -> 76.2 | Helped |
Closest papers by meaning
- Towards Deep Attention in Graph Neural Networks: Problems and Remedies (ICML 2023 · similarity 0.91)
- Graph Attention Networks (ICLR 2018 · similarity 0.90)
- How to Find Your Friendly Neighborhood: Graph Attention Design with Self-Supervision (ICLR 2021 · similarity 0.90)
- Graph Attention is Not Always Beneficial: A Theoretical Analysis of Graph Attention Mechanisms via Contextual Stochastic Block Models (ICML 2025 · similarity 0.88)
- What Improves the Generalization of Graph Transformers? A Theoretical Dive into the Self-attention and Positional Encoding (ICML 2024 · similarity 0.88)
- Attention Mechanisms Perspective: Exploring LLM Processing of Graph-Structured Data (ICML 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.