Research map / Graph learning / Graph self-supervised learning
Regularizing Graph Neural Networks via Consistency-Diversity Graph Augmentations
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 Neighbor Replace augmentation (ablation) | GCN-based NASA · Cora | accuracy: 84.5±0.1 -> 85.1±0.3 | Helped |
| add Neighbor Replace augmentation (ablation) | GCN-based NASA · Citeseer | accuracy: 75.0±0.1 -> 75.5±0.4 | Helped |
| replace Neighbor Replace with DropEdge (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 84.7±0.5; 75.5±0.4 -> 75.1±0.2 | Hurt |
| replace Neighbor Replace with DropNode (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 84.6±0.3; 75.5±0.4 -> 74.9±0.3 | Hurt |
| replace Neighbor Replace with dropout (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 84.5±0.2; 75.5±0.4 -> 74.7±0.2 | Hurt |
| use static rather than dynamic training (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 84.7±0.9; 75.5±0.4 -> 70.7±12.6 | Hurt |
| remove graph augmentation (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 84.5±0.1; 75.5±0.4 -> 75.0±0.1 | Hurt |
| remove neighbor-constrained regularization (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 83.4±0.4; 75.5±0.4 -> 73.1±0.7 | Hurt |
| remove sharpening (ablation) | GCN-based NASA · Cora and Citeseer | accuracy: 85.1±0.3 -> 83.7±0.5; 75.5±0.4 -> 72.7±0.5 | Hurt |
| apply NASA regularization and augmentation | GCN · Cora standard split | accuracy: 81.5±0.3 -> 85.1±0.3 | Helped |
| apply NASA regularization and augmentation | GCN · Citeseer standard split | accuracy: 70.3±0.9 -> 75.5±0.4 | Helped |
| apply NASA regularization and augmentation | GCN · Pubmed standard split | accuracy: 79.0±0.2 -> 80.2±0.3 | Helped |
Closest papers by meaning
- Graph Random Neural Networks for Semi-Supervised Learning on Graphs (NeurIPS 2020 · similarity 0.93)
- Normalize Then Propagate: Efficient Homophilous Regularization for Few-Shot Semi-Supervised Node Classification (AAAI 2025 · similarity 0.90)
- Graph Neural Network Generalization With Gaussian Mixture Model Based Augmentation (ICML 2025 · similarity 0.89)
- Towards Bridging Generalization and Expressivity of Graph Neural Networks (ICLR 2025 · similarity 0.89)
- Through the Dual-Prism: A Spectral Perspective on Graph Data Augmentation for Graph Classifications (AAAI 2025 · similarity 0.88)
- Exploring Consistency in Graph Representations: from Graph Kernels to Graph Neural Networks (NeurIPS 2024 · 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.