Multi-Instance Adversarial Attack on GNN-Based Malicious Domain Detection

Explore a 15-minute IEEE conference talk examining multi-instance adversarial attacks on Graph Neural Network (GNN) based malicious domain detection systems. Delve into the vulnerabilities of GNN models in cybersecurity applications and learn about innovative attack strategies that exploit these weaknesses. Gain insights into the challenges faced by current malicious domain detection methods and understand the potential implications for network security. Discover cutting-edge research findings and potential countermeasures to enhance the robustness of GNN-based detection systems against sophisticated adversarial threats.