Research map / Graph learning / Graph self-supervised learning
Low-Rank Few-Shot Node Classification by Node-Level Graph Diffusion
Yancheng Wang, Chengshuai Zhao, Dongfang Sun, huan liu, Yingzhen Yang
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 |
|---|---|---|---|
| add LR-FGDM to COLA | COLA · Coauthor-CS 5-way 5-shot | accuracy (%): 87.54 -> 89.83 | Improved |
| add LRA layer on top of LR-FGDM | COLA (LR-FGDM) · CoraFull 5-way 5-shot | accuracy (%): 89.66 -> 90.32 | Improved |
| add LR-FGDM to COSMIC | COSMIC · CoraFull 2-way 5-shot | accuracy (%): 94.51 -> 96.74 | Improved |
| add low-rank regularization (ablation) | COLA (LR-FGDM) · CoraFull 5-way 5-shot | accuracy (%): 88.74 -> 89.66 | Improved |
| add low-rank regularization (ablation) | COLA (LR-FGDM) · ogbn-arxiv 5-way 5-shot | accuracy (%): 68.60 -> 69.63 | Improved |
| add prototypical regularization (ablation) | COLA (LR-FGDM) · Coauthor-CS 5-way 5-shot | accuracy (%): 89.02 -> 89.83 | Improved |
| add prototypical regularization (ablation) | COLA (LR-FGDM) · DBLP 5-way 5-shot | accuracy (%): 88.64 -> 89.51 | Improved |
| add both regularizers (augmentation alone is already slightly helpful) (ablation) | COLA (LR-FGDM) · ogbn-arxiv 5-way 5-shot | accuracy (%): 67.91 -> 69.63 | Improved |
| use hierarchical edge reconstruction instead of BLND (ablation) | LR-FGDM · ogbn-arxiv 5-way 5-shot | accuracy (%): 68.47 -> 69.63 | Improved |
| use hierarchical edge reconstruction instead of regular GAE decoder (ablation) | HGAE · ogbn-arxiv training time | seconds: 1690 -> 301 | Improved |
| use hierarchical edge reconstruction for generation (ablation) | FGDM · ogbn-arxiv generation time | s/sample: 0.426 -> 0.130 | Improved |
| add LR-FGDM synthetic support nodes instead of SNS (GAN) | COLA · CoraFull 5-way 5-shot | accuracy (%): 88.49 -> 89.66 | Improved |
Closest papers by meaning
- Fast Graph Sharpness-Aware Minimization for Enhancing and Accelerating Few-Shot Node Classification (NeurIPS 2024 · similarity 0.88)
- Graph Few-Shot Learning via Adaptive Spectrum Experts and Cross-Set Distribution Calibration (NeurIPS 2025 · similarity 0.87)
- Normalize Then Propagate: Efficient Homophilous Regularization for Few-Shot Semi-Supervised Node Classification (AAAI 2025 · similarity 0.86)
- Few-Shot Diffusion Models Escape the Curse of Dimensionality (NeurIPS 2024 · similarity 0.86)
- Outlier-Aware Post-Training Quantization for Discrete Graph Diffusion Models (ICML 2025 · similarity 0.86)
- Unifying Generation and Prediction on Graphs with Latent Graph Diffusion (NeurIPS 2024 · similarity 0.85)
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.
