Research map / Adversarial robustness and security / Attacks on text and graphs
Certifiably Robust Graph Contrastive Learning
Minhua Lin, Teng Xiao, Enyan Dai, Xiang Zhang, Suhang Wang
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 |
|---|---|---|---|
| apply RES to GRACE | GRACE · Cora, random-attack graph | robust accuracy: 74.5 -> 79.1 | Improved |
| apply RES to GRACE | GRACE · Cora, CLGA-poisoned graph | robust accuracy: 74.9 -> 78.2 | Improved |
| apply RES to GRACE | GRACE · Pubmed, PRBCD graph | robust accuracy: 73.2 -> 78.8 | Improved |
| apply RES to GRACE | GRACE · Physics, PRBCD graph | robust accuracy: 89.2 -> 94.1 | Improved |
| apply RES to GRACE | GRACE · OGB-arxiv, PRBCD graph | robust accuracy: 55.7 -> 58.3 | Improved |
| apply RES to GRACE on clean graph | GRACE · Cora, raw graph | accuracy: 77.1 -> 79.7 | Improved |
| apply RES to GRACE on clean graph | GRACE · Pubmed, raw graph | accuracy: 79.5 -> 79.5 | No clear effect |
| apply RES to GRACE on clean graph | GRACE · Physics, raw graph | accuracy: 94.0 -> 94.7 | Improved |
| apply RES to GRACE on clean graph | GRACE · OGB-arxiv, raw graph | accuracy: 65.1 -> 65.2 | Improved |
| RES-GRACE vs robust GCL baseline Ariel on clean Cora | RES-GRACE · Cora, raw graph | accuracy: 79.8 -> 79.7 | Hurt |
| replace randomized edgedrop noise with binary flip noise (FLIP) (ablation) | RES · Cora, Nettack budget 3 | robust accuracy: 79.4 -> 27 | Hurt |
| replace randomized edgedrop noise with binary flip noise (FLIP) (ablation) | RES · Pubmed, Nettack budget 3 | robust accuracy: 82.5 -> 42 | Hurt |
Closest papers by meaning
- Does GCL Need a Large Number of Negative Samples? Enhancing Graph Contrastive Learning with Effective and Efficient Negative Sampling (AAAI 2025 · similarity 0.90)
- GNN-Transformer Cooperative Architecture for Trustworthy Graph Contrastive Learning (AAAI 2025 · similarity 0.89)
- ReGCL: Rethinking Message Passing in Graph Contrastive Learning (AAAI 2024 · similarity 0.89)
- Bounding the Expected Robustness of Graph Neural Networks Subject to Node Feature Attacks (ICLR 2024 · similarity 0.89)
- Refine then Classify: Robust Graph Neural Networks with Reliable Neighborhood Contrastive Refinement (AAAI 2025 · similarity 0.89)
- Hierarchical Topology Isomorphism Expertise Embedded Graph Contrastive Learning (AAAI 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.
