Research map / Graph learning / Graph self-supervised learning
Knowledge Distillation Improves Graph Structure Augmentation 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 KDGA distillation to GraphAug | GCN + GraphAug · Squirrel | accuracy (%): 37.2 -> 46.9 | Helped |
| add KDGA distillation to GraphAug | GCN + GraphAug · Texas | accuracy (%): 59.4 -> 68.0 | Helped |
| add KDGA distillation to GraphAug | GCN + GraphAug · Cora | accuracy (%): 83.2 -> 85.2 | Helped |
| add KDGA distillation to GraphAug | GCN + GraphAug · Citeseer | accuracy (%): 73.2 -> 73.9 | Helped |
| add KDGA distillation to GAUG | GCN + GAUG · Wisconsin | accuracy (%): 57.5 -> 65.0 | Helped |
| add KDGA distillation to MH-Aug | GAT + MH-Aug · Squirrel | accuracy (%): 29.5 -> 37.3 | Helped |
| apply GAUG graph augmentation alone on heterophily graphs | GCN · Texas | accuracy (%): 59.5 -> 58.0 | Hurt |
| apply GAUG graph augmentation alone on heterophily graphs | GCN · Cornell | accuracy (%): 57.0 -> 55.8 | Hurt |
| apply AdaEdge heuristic augmentation | GCN · Citeseer | accuracy (%): 71.6 -> 69.7 | Hurt |
| apply DropEdge heuristic augmentation | GCN · Cora | accuracy (%): 81.5 -> 82.2 | Helped |
| apply GraphAug alone without KDGA | GCN · Actor | accuracy (%): 30.3 -> 28.8 | Hurt |
| use parameter-independent student instead of parameter-shared (ablation) | KDGA GCN GraphAug · Squirrel | accuracy (%): 46.9 -> 40.5 | Hurt |
Closest papers by meaning
- Regularizing Graph Neural Networks via Consistency-Diversity Graph Augmentations (AAAI 2022 · similarity 0.91)
- Through the Dual-Prism: A Spectral Perspective on Graph Data Augmentation for Graph Classifications (AAAI 2025 · similarity 0.89)
- Community-Invariant Graph Contrastive Learning (ICML 2024 · similarity 0.89)
- Contextual Structure Knowledge Transfer for Graph Neural Networks (AAAI 2025 · similarity 0.89)
- VQGraph: Rethinking Graph Representation Space for Bridging GNNs and MLPs (ICLR 2024 · similarity 0.89)
- Large Language Model Meets Graph Neural Network in Knowledge Distillation (AAAI 2025 · 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.