Research map / Graph learning / GNN architectures
Demystifying Uneven Vulnerability of Link Stealing Attacks against Graph Neural Networks
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 group-based thresholds (separate inter-class threshold) instead of a single shared threshold for link stealing | Link stealing attack on GCN · COX2, DHFR, PROTEINS (ASR, multiple distances) | ASR of inter-class group: e.g. PROTEINS cosine 63.39 to 98.22 | Helped |
| Inter-class vs intra-class grouping of node pairs (uneven vulnerability) | GCN victim model · Ten datasets incl. Credit | AUC difference between groups: about 25% AUC difference on Credit | Helped |
| Use GraphSAGE instead of GCN as victim model | Victim architecture · GCN vs GAT vs GraphSAGE | inter/intra AUC difference | No clear effect |
Closest papers by meaning
- GraphIP–Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It? (NeurIPS 2026 · similarity 0.88)
- On provable privacy vulnerabilities of graph representations (NeurIPS 2024 · similarity 0.88)
- Can Graph Neural Networks Expose Training Data Properties? An Efficient Risk Assessment Approach (NeurIPS 2024 · similarity 0.87)
- Impact of Dataset Properties on Membership Inference Vulnerability of Deep Transfer Learning (NeurIPS 2025 · similarity 0.87)
- Practical Bayes-Optimal Membership Inference Attacks (NeurIPS 2025 · similarity 0.87)
- Networked Inequality: Preferential Attachment Bias in Graph Neural Network Link Prediction (ICML 2024 · 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.