Research map / Adversarial robustness and security / Black-box and transfer attacks
A Restricted Black-Box Adversarial Framework Towards Attacking Graph Embedding Models
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 |
|---|---|---|---|
| Use GF-Attack rather than the strongest listed baseline | GCN · Cora | change in classification accuracy (percent): -5.20 -> -7.60 | Helped |
| Use GF-Attack rather than the strongest listed baseline | SGC · Cora | change in classification accuracy (percent): -5.62 -> -9.73 | Helped |
| Use GF-Attack rather than the strongest listed baseline | DeepWalk · Cora | change in classification accuracy (percent): -6.29 vs -5.31 | Hurt |
| Use GF-Attack rather than the strongest listed baseline | LINE · Cora | change in classification accuracy (percent): -12.40 -> -13.27 | Helped |
| Use GF-Attack rather than the strongest listed baseline | GCN · Citeseer | change in classification accuracy (percent): -6.50 -> -7.78 | Helped |
| Use GF-Attack rather than the strongest listed baseline | SGC · Citeseer | change in classification accuracy (percent): -4.08 -> -6.19 | Helped |
| Use GF-Attack rather than the strongest listed baseline | DeepWalk · Citeseer | change in classification accuracy (percent): -12.56 vs -12.50 | No clear effect |
| Use GF-Attack rather than the strongest listed baseline | LINE · Citeseer | change in classification accuracy (percent): -20.10 -> -22.11 | Helped |
| Use GF-Attack rather than the strongest listed baseline | GCN · Pubmed | change in classification accuracy (percent): -6.40 -> -7.96 | Helped |
| Use GF-Attack rather than the strongest listed baseline | SGC · Pubmed | change in classification accuracy (percent): -6.11 -> -7.20 | Helped |
| Use GF-Attack rather than the strongest listed baseline | DeepWalk · Pubmed | change in classification accuracy (percent): -3.05 -> -7.43 | Helped |
| Use GF-Attack rather than the strongest listed baseline | LINE · Pubmed | change in classification accuracy (percent): -13.21 -> -14.16 | Helped |
Closest papers by meaning
- Blindfolded Attackers Still Threatening: Strict Black-Box Adversarial Attacks on Graphs (AAAI 2022 · similarity 0.92)
- Adversarial Attacks on Node Embeddings via Graph Poisoning (ICML 2019 · similarity 0.91)
- Towards More Practical Adversarial Attacks on Graph Neural Networks (NeurIPS 2020 · similarity 0.90)
- Value at Adversarial Risk: A Graph Defense Strategy against Cost-Aware Attacks (AAAI 2024 · similarity 0.88)
- Highly Imperceptible Black-Box Graph Injection Attacks with Reinforcement Learning (AAAI 2025 · similarity 0.87)
- Adversarial Robust Generalization of Graph Neural Networks (ICML 2025 · similarity 0.87)
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.