Research map / Graph learning / Graph self-supervised learning
Spectral Feature Augmentation for Graph Contrastive Learning and Beyond
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 spectral feature augmentation to InfoNCE | Graph contrastive learning · Cora node classification | accuracy: 80.04 -> 85.89 | Helped |
| Add spectral feature augmentation to InfoNCE | Graph contrastive learning · CiteSeer node classification | accuracy: 71.35 -> 75.31 | Helped |
| Add spectral feature augmentation to SimCLR | Image contrastive learning · CIFAR10 | top-1 accuracy: 90.5 -> 91.6 | Helped |
| Add spectral feature augmentation to SimCLR | Image contrastive learning · CIFAR100 | top-1 accuracy: 65.5 -> 66.7 | Helped |
| Add spectral feature augmentation to SimCLR | Image contrastive learning · ImageNet-100 | top-1 accuracy: 76.8 -> 77.7 | Helped |
| Add graph augmentation and SFA together (ablation) | Graph contrastive learning · Cora | accuracy: 80.04 -> 85.90 | Helped |
| Add graph augmentation and SFA together (ablation) | Graph contrastive learning · Am-Comput. | accuracy: 85.01 -> 88.74 | Helped |
| Replace random channel feature augmentation with SFA alongside graph augmentation (ablation) | Graph contrastive learning · Cora | accuracy: 83.35 -> 85.90 | Helped |
| Use k=0 instead of the incomplete power iteration (ablation) | SFA · Am-Computer | accuracy: 88.83 -> 87.25 | Hurt |
| Use k=0 instead of the incomplete power iteration (ablation) | SFA · Cora | accuracy: 85.90 -> 83.04 | Hurt |
| Use k=0 instead of the incomplete power iteration (ablation) | SFA · CiteSeer | accuracy: 73.56 -> 71.35 | Hurt |
| Add SFA under noisy input features (ablation) | Graph contrastive learning · Cora, ε=10^-1 | accuracy: 59.23 -> 62.55 | Helped |
Closest papers by meaning
- Revisiting Graph Contrastive Learning from the Perspective of Graph Spectrum (NeurIPS 2022 · similarity 0.91)
- Community-Invariant Graph Contrastive Learning (ICML 2024 · similarity 0.90)
- Perfect Alignment May be Poisonous to Graph Contrastive Learning (ICML 2024 · similarity 0.89)
- Through the Dual-Prism: A Spectral Perspective on Graph Data Augmentation for Graph Classifications (AAAI 2025 · similarity 0.89)
- S3GCL: Spectral, Swift, Spatial Graph Contrastive Learning (ICML 2024 · similarity 0.87)
- Graph Contrastive Learning with Joint Spectral Augmentation of Attribute and Topology (AAAI 2025 · 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.