A new research paper details the first comprehensive evaluation of gradient leakage attacks (GLAs) on graph neural networks (GNNs) used in circuit design and hardware security. The study reveals that GLAs can expose sensitive information like gate types and hardware Trojan properties, potentially aiding adversaries. While some GNN architectures like GIN offer more resilience, others like GAT can exacerbate leakage. Existing defense techniques show limited effectiveness and can degrade model performance, indicating a need for more robust privacy-preserving solutions. AI
IMPACT Highlights potential security vulnerabilities in AI models used for critical infrastructure, necessitating development of more robust privacy-preserving techniques.
RANK_REASON The cluster contains a research paper detailing a new security vulnerability and evaluation of existing defenses for graph neural networks.
- adversarial training
- differential privacy
- gradient clipping
- Gradient Leakage Attacks
- GAT
- Graph Neural Networks
- GraphSAGE
- ISCAS'85
- secure aggregation
- EPFL
- TrustHub
- model compression
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
