Research map / Adversarial robustness and security / Black-box and transfer attacks
Blindfolded Attackers Still Threatening: Strict Black-Box Adversarial Attacks on Graphs
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 STACK instead of GPGD | GCN node classification · Cora-ML | Macro-F1 decrease (%): 4.22$\pm$0.6 -> 5.27$\pm$0.3 | Helped |
| use STACK instead of GPGD | GCN node classification · Citeseer | Macro-F1 decrease (%): 2.14$\pm$0.9 -> 3.98$\pm$0.5 | Helped |
| use STACK instead of GPGD | GCN node classification · Polblogs | Macro-F1 decrease (%): 2.35$\pm$1.8 -> 5.32$\pm$1.1 | Helped |
| add edge-dependency updates and restart to STACK-r-d (ablation) | STACK · Cora-ML node classification (GCN) | Macro-F1 decrease (%): 4.03$\pm$0.6 -> 5.27$\pm$0.3 | Helped |
| add restart mechanism (ablation) | STACK · Cora-ML node classification (GCN) | Macro-F1 decrease (%): 5.02$\pm$0.4 -> 5.27$\pm$0.3 | Helped |
| add edge-dependency updates and restart to STACK-r-d (ablation) | STACK · Proteins graph classification (GIN) | Macro-F1 decrease (%): 11.48 -> 13.53 | Helped |
| add restart mechanism (ablation) | STACK · Proteins graph classification (GIN) | Macro-F1 decrease (%): 12.81 -> 13.53 | Helped |
| add edge-dependency updates and restart to STACK-r-d (ablation) | STACK · Enzymes graph classification (Diffpool) | Macro-F1 decrease (%): 38.18 -> 39.62 | Helped |
| add restart mechanism (ablation) | STACK · Enzymes graph classification (Diffpool) | Macro-F1 decrease (%): 40.18 -> 39.62 | Hurt |
| add restart mechanism (ablation) | STACK · Cora-ML node classification (GCN, 5% perturbation) | Macro-F1 decrease (%): 4.43 -> 4.30 | Hurt |
| add restart mechanism (ablation) | STACK · Cora-ML node classification (GCN, 15% perturbation) | Macro-F1 decrease (%): 5.77 -> 6.40 | Helped |
Closest papers by meaning
- Towards More Practical Adversarial Attacks on Graph Neural Networks (NeurIPS 2020 · similarity 0.92)
- A Restricted Black-Box Adversarial Framework Towards Attacking Graph Embedding Models (AAAI 2020 · similarity 0.92)
- Adversarial Attacks on Graph Classifiers via Bayesian Optimisation (NeurIPS 2021 · similarity 0.91)
- Value at Adversarial Risk: A Graph Defense Strategy against Cost-Aware Attacks (AAAI 2024 · similarity 0.90)
- Highly Imperceptible Black-Box Graph Injection Attacks with Reinforcement Learning (AAAI 2025 · similarity 0.89)
- Adversarial Contrastive Graph Masked AutoEncoder Against Graph Structure and Feature Dual Attacks (AAAI 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.