Research map / Graph learning / GNN architectures
Orthogonal 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 |
|---|---|---|---|
| Apply Ortho-GConv to GCNII | GCNII · Full-supervised node classification, six datasets | accuracy: 2.2% | Helped |
| Apply Ortho-GConv to GCNII | GCNII · Texas | accuracy: 7.4% | Helped |
| Apply Ortho-GConv to Graph-U-Nets | Graph-U-Nets · D&D | accuracy: 83.00 -> 83.87 | Helped |
| Apply Ortho-GConv to Graph-U-Nets | Graph-U-Nets · PROTEINS | accuracy: 77.68 -> 78.78 | Helped |
| Apply Ortho-GConv to GCNII | GCNII · CiteSeer (semi-supervised) | accuracy: 1.1% | Helped |
| Remove hybrid weight initialization (ablation) | Ortho-GCN · Cora (semi-supervised) | accuracy | Hurt |
| Remove orthogonal transformation (ablation) | Ortho-GCN · Cora (semi-supervised) | accuracy | Hurt |
| Remove orthogonal regularization (ablation) | Ortho-GCN · Cora (semi-supervised) | accuracy | Hurt |
| Vary Newton iteration count away from T=4 (ablation) | Ortho-GCN · Cora, 2-layer and 8-layer | accuracy | Hurt |
| Increase Newton iteration count (ablation) | Ortho-GCN · Cora, 2-layer and 8-layer | training time | Hurt |
| Apply Ortho-GConv across 2-, 4-, and 8-layer backbones | GCN, JKNet, and GCNII · Full-supervised node classification | accuracy: 1.9%; 2.1%; 2.8%; 12.4% | Mixed |
Closest papers by meaning
- Simple Orthogonal Graph Representation Learning (Student Abstract) (AAAI 2024 · similarity 0.90)
- Generalizing CNNs to graphs with learnable neighborhood quantization (NeurIPS 2024 · similarity 0.87)
- Residual Connections and Normalization Can Provably Prevent Oversmoothing in GNNs (ICLR 2025 · similarity 0.87)
- Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence (ICML 2025 · similarity 0.87)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.87)
- Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024 · similarity 0.86)
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.