Structure-Aware Distributed Backdoor Attacks in Federated Learning

Researchers have discovered that model architecture significantly affects the success of backdoor attacks in federated learning systems. The study introduces new metrics to measure model vulnerability and develops a framework showing that certain network structures can amplify malicious perturbations even with minimal poisoning.

• →Model architecture plays a crucial role in determining the effectiveness of backdoor attacks in federated learning systems.
• →Networks with multi-path feature fusion can amplify and retain malicious perturbations even under low poisoning ratios.
• →Two new metrics (SRS and SCC) were introduced to measure model sensitivity and preference for fractal perturbations.
• →The Structural Compatibility Coefficient strongly correlates with attack success rates and can predict perturbation survivability.
• →These findings suggest that defensive strategies should consider model architecture alongside traditional security measures.