Research map / Graph learning / GNN architectures
Rethinking Graph Regularization for 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 P-reg to GCN | GCN · CORA random split | accuracy: 79.47 -> 82.83 | Helped |
| Add P-reg to GCN | GCN · CiteSeer random split | accuracy: 69.45 -> 71.63 | Helped |
| Add P-reg to GCN | GCN · PubMed random split | accuracy: 76.26 -> 77.37 | Helped |
| Add P-reg to GCN | GCN · CS random split | accuracy: 91.48 -> 92.58 | Helped |
| Add P-reg to GCN | GCN · Physics random split | accuracy: 93.66 -> 94.41 | Helped |
| Add P-reg to GCN | GCN · Computers random split | accuracy: 80.73 -> 81.66 | Helped |
| Add P-reg to GCN | GCN · Photo random split | accuracy: 89.24 -> 91.24 | Helped |
| Add P-reg to GAT | GAT · PubMed random split | accuracy: 75.61 -> 76.39 | Mixed |
| Add Laplacian regularization | GCN · node classification random splits | accuracy | No clear effect |
| Replace cross-entropy P-reg with squared-error P-reg (ablation) | GCN · CORA random split | accuracy: 82.83 -> 82.85 | No clear effect |
| Replace cross-entropy P-reg with KL-divergence P-reg (ablation) | GCN · CiteSeer random split | accuracy: 71.63 -> 69.76 | Hurt |
| Replace cross-entropy P-reg with squared-error P-reg (ablation) | GAT · Computers random split | accuracy: 83.68 -> 76.89 | Hurt |
Closest papers by meaning
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.89)
- Generalizing CNNs to graphs with learnable neighborhood quantization (NeurIPS 2024 · similarity 0.88)
- Normalize Then Propagate: Efficient Homophilous Regularization for Few-Shot Semi-Supervised Node Classification (AAAI 2025 · similarity 0.88)
- Two Heads Are Better Than One: Boosting Graph Sparse Training via Semantic and Topological Awareness (ICML 2024 · similarity 0.88)
- Residual Connections and Normalization Can Provably Prevent Oversmoothing in GNNs (ICLR 2025 · similarity 0.88)
- Locality-Aware Graph Rewiring in GNNs (ICLR 2024 · 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.