Research map / Graph learning / Graph self-supervised learning
Graph Contrastive Learning with Stable and Scalable Spectral Encoding
Deyu Bo, Yuan Fang, Yang Liu, Chuan Shi
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 |
|---|---|---|---|
| Replace EigenMLP with MLP in Sp2GCL (ablation) | Sp2GCL · Facebook | accuracy (%): 90.22 -> 90.43 | Improved |
| Replace EigenMLP with MLP (ablation) | Sp2GCL · PubMed | accuracy (%): 82.3 -> 82.3 | No clear effect |
| Replace EigenMLP with MLP (ablation) | Sp2GCL · molbace | ROC-AUC (%): 77.49 -> 78.76 | Improved |
| Replace EigenMLP with SAN (ablation) | Sp2GCL · Facebook | accuracy (%): 81.52 -> 90.43 | Improved |
| Replace EigenMLP with BasisNet (ablation) | Sp2GCL · Facebook | accuracy (%): 83.87 -> 90.43 | Improved |
| Add eigenvectors U to GRACE inputs (ablation) | GRACE · Facebook | accuracy (%): 89.32 -> 89.85 | Improved |
| Add eigenvectors and eigenvalues (+UL) to GRACE (ablation) | GRACE · Pubmed | accuracy (%): 80.6 -> 81.5 | Improved |
| Use spatial-spectral contrast Sp2GCL instead of feeding spectral features to GRACE (ablation) | Sp2GCL · molbace | ROC-AUC (%): 77.36 -> 78.76 | Improved |
| Add spectral view (Sp2GCL) to spatial GCL on Wiki-CS | Sp2GCL · Wiki-CS | accuracy (%): 80.14 -> 79.42 | Hurt |
| Use Sp2GCL instead of SpCo/SPAN on large graphs | Sp2GCL · arXiv, Flickr, PPI | accuracy (%): 71.83, 52.05, 74.28 | Improved |
| Apply random reflection and rotation to eigenvectors, MLP instead of EigenMLP (ablation) | EigenMLP · PubMed | accuracy change (%): -9.3 | Improved |
| Use EigenMLP instead of MLP under EVD tolerance perturbations (ablation) | EigenMLP · PubMed | accuracy change (%): -7.8 | Improved |
Closest papers by meaning
- Revisiting Graph Contrastive Learning from the Perspective of Graph Spectrum (NeurIPS 2022 · similarity 0.93)
- S3GCL: Spectral, Swift, Spatial Graph Contrastive Learning (ICML 2024 · similarity 0.93)
- PolyGCL: GRAPH CONTRASTIVE LEARNING via Learnable Spectral Polynomial Filters (ICLR 2024 · similarity 0.91)
- ReGCL: Rethinking Message Passing in Graph Contrastive Learning (AAAI 2024 · similarity 0.89)
- Spectral Graph Neural Networks are Incomplete on Graphs with a Simple Spectrum (NeurIPS 2025 · similarity 0.89)
- Graph Contrastive Invariant Learning from the Causal Perspective (AAAI 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.
