Research map / Graph learning / GNN architectures
Sheaf Graph Neural Networks via PAC-Bayes Spectral Optimization
Yoonhyuk Choi, Jiho Choi, Taewook Ko, JongWook Kim, Chong-Kwon Kim
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use SGPC over best baseline | SGPC · Actor | accuracy: 25.9 -> 38.1 | Improved |
| use SGPC over best baseline | SGPC · Cora | accuracy: 83.3 -> 83.0 | Hurt |
| use SGPC over best baseline | SGPC · Pubmed | accuracy: 80.2 -> 79.9 | Hurt |
| combine SVR diffusion with AFM branch (ablation) | SGPC · Cora, Citeseer, Chameleon, Texas | accuracy: about 2% over stronger single branch | Improved |
| add Wasserstein-Entropic lift with sheaf Laplacian in deep stacks (ablation) | SGPC · Cora (8 layers) | accuracy: 81.8% (drop of only 1.2%) vs about 16.6% and 22.1% drops | Improved |
| enlarge sheaf spectral gap lambda2 (ablation) | SGPC · Chameleon | test error | Improved |
| set lambda_KL very small (<=0.01) (ablation) | SGPC · node classification | accuracy | Hurt |
| vary spectral regularizer weight lambda_spec (ablation) | SGPC · node classification | accuracy | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SimCalib — SimCalib: Graph Neural Network Calibration Based on Similarity between Nodes (AAAI 2024) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Enhancing Spectral GNNs: From Topology and Perturbation Perspectives (ICML 2025 · similarity 0.89)
- Mitigating Over-squashing without Rewiring: A Sheaf Effective Resistance Perspective (NeurIPS 2026 · similarity 0.88)
- A Manifold Perspective on the Statistical Generalization of Graph Neural Networks (ICML 2025 · similarity 0.87)
- Compositional PAC-Bayes: Generalization of GNNs with persistence and beyond (NeurIPS 2024 · similarity 0.87)
- Bonsai: Gradient-free Graph Condensation for Node Classification (ICLR 2025 · similarity 0.87)
- How Graph Neural Networks Learn: Lessons from Training Dynamics (ICML 2024 · similarity 0.87)
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.
