Research map / Adversarial robustness and security / Attacks on text and graphs
Towards Reasonable Budget Allocation in Untargeted Graph Structure Attacks via Gradient Debias
Zihan Liu, Yun Luo, Lirong Wu, Zicheng Liu, Stan Z. Li
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 |
|---|---|---|---|
| Replace negative cross-entropy with gradient-debias objective | GraD versus Meta-Self · Weak transfer, Cora, Cora-ML, Citeseer, and Polblogs | classification accuracy: +2.7, +4.4, +1.6, +1.2, +3.8, +4.6, +5.0, +6.3 | Improved |
| Replace negative cross-entropy with gradient-debias objective | GraD versus strongest baseline · Strong transfer, GAT victims on Cora, Cora-ML, Citeseer, and Polblogs | classification accuracy: +3.4, +1.3, +0.2, +11.0 | Improved |
| Replace negative cross-entropy with gradient-debias objective | GraD versus strongest baseline · Strong transfer, GraphSage victims on Cora, Cora-ML, Citeseer, and Polblogs | classification accuracy: +4.3, +1.7, +2.0, -1.9 | Mixed |
| Weight the negative cross-entropy loss by pseudo-label confidence to cancel confidence-dependent gradient magnitude | Untargeted graph structure attack objective · Citeseer and Cora | partial-gradient L2 norm versus node confidence | Improved |
Closest papers by meaning
- Rethinking Independent Cross-Entropy Loss For Graph-Structured Data (ICML 2024 · similarity 0.86)
- Deterministic Certification of Graph Neural Networks against Graph Poisoning Attacks with Arbitrary Perturbations (CVPR 2025 · similarity 0.86)
- Clean-Label Graph Backdoor Attack in the Node Classification Task (AAAI 2025 · similarity 0.86)
- Robustness Inspired Graph Backdoor Defense (ICLR 2025 · similarity 0.86)
- Stealthy Yet Effective: Distribution-Preserving Backdoor Attacks on Graph Classification (NeurIPS 2025 · similarity 0.85)
- Attack by Yourself: Effective and Unnoticeable Multi-Category Graph Backdoor Attacks with Subgraph Triggers Pool (NeurIPS 2025 · 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.
