Unveiling Hidden Threats: Using Fractal Triggers to Boost Stealthiness of Distributed Backdoor Attacks in Federated Learning

2025-11-12

AI Summary

Traditional distributed backdoor attacks (DBAs) in federated learning rely heavily on large volumes of poisoned data, rendering them susceptible to detection. To address this, this paper proposes a novel fractal-structured DBA framework. First, it introduces fractal self-similarity into trigger design, generating sub-trigger patterns with enhanced feature representation and significantly improving per-sample poisoning efficacy. Second, it incorporates a dynamic angular perturbation mechanism that jointly suppresses detectability in both the frequency and gradient domains. Experimental results demonstrate that the proposed method achieves a 92.3% attack success rate using only 62.4% of the poisoning samples required by conventional approaches, while reducing detection probability by 22.8% and KL divergence by 41.2%. This framework thus unifies low poisoning overhead, high attack effectiveness, and strong stealthiness.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Distributed Search

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Traditional distributed backdoor attacks (DBA) in federated learning improve stealthiness by decomposing global triggers into sub-triggers, which however requires more poisoned data to maintian the attck strength and hence increases the exposure risk. To overcome this defect, This paper proposes a novel method, namely Fractal-Triggerred Distributed Backdoor Attack (FTDBA), which leverages the self-similarity of fractals to enhance the feature strength of sub-triggers and hence significantly reduce the required poisoning volume for the same attack strength. To address the detectability of fractal structures in the frequency and gradient domains, we introduce a dynamic angular perturbation mechanism that adaptively adjusts perturbation intensity across the training phases to balance efficiency and stealthiness. Experiments show that FTDBA achieves a 92.3% attack success rate with only 62.4% of the poisoning volume required by traditional DBA methods, while reducing the detection rate by 22.8% and KL divergence by 41.2%. This study presents a low-exposure, high-efficiency paradigm for federated backdoor attacks and expands the application of fractal features in adversarial sample generation.

Problem

• Enhancing stealthiness of distributed backdoor attacks in federated learning
• Reducing poisoning data volume while maintaining attack strength
• Mitigating detectability of fractal triggers through adaptive perturbation

Innovation

• Using fractal triggers to enhance sub-trigger feature strength
• Applying dynamic angular perturbation for stealthiness
• Reducing poisoning volume while maintaining attack success