Defending Against Backdoor Attacks via Alignment Checking in Model-Contrastive Federated Learning

2026-07-29

AI Summary

This work addresses the vulnerability of federated learning in edge environments to stealthy backdoor attacks, a challenge exacerbated by the limitations of existing defenses that fail to account for benign update deviations caused by statistical heterogeneity. To overcome this, the authors propose FedDAB, a two-stage defense mechanism that first enhances the consistency of benign local updates in both direction and magnitude through model contrastive regularization, and then identifies and filters out malicious updates via a combined global-direction and parameter-level alignment detection strategy. FedDAB is the first approach to integrate model contrastive learning with a historical alignment mechanism, effectively detecting stealthy backdoors while accommodating statistical heterogeneity. The method provides theoretical convergence guarantees at an 𝒪(1/T) rate. Experimental results demonstrate that FedDAB significantly outperforms current state-of-the-art defenses.

Abstract

Federated Learning (FL) is vulnerable to backdoor attacks because of its distributed nature in edge computing scenarios. Existing defense methods show limited efficacy as they overlook the deviations among benign local updates caused by statistical heterogeneity and the stealthiness of backdoor attacks. To tackle these issues, we propose FedDAB, a two-phase method that combines local contrastive regularization with alignment checking, to defend against backdoor attacks. In the first phase, FedDAB introduces a novel model-contrastive term into the local objective to enhance direction and magnitude consistency among benign updates. In the second phase, FedDAB employs an alignment checking strategy to evaluate each local update in terms of overall-direction alignment and parameter-level alignment with historical information, excluding updates that exhibit abnormal alignment patterns from global aggregation. We theoretically prove FedDAB's robustness with a convergence rate of $\mathcal{O}(1/T)$. Extensive experiments show that FedDAB outperforms existing defense methods against backdoor attacks.

Problem

• Backdoor Attacks
• Federated Learning
• Statistical Heterogeneity
• Edge Computing
• Model Robustness

Innovation

• Federated Learning
• Backdoor Attack
• Model-Contrastive Learning
• Alignment Checking
• Robust Aggregation