Beyond Passive Aggregation: Active Auditing and Topology-Aware Defense in Decentralized Federated Learning

TL;DR

This paper introduces an active, topology-aware defense framework for decentralized federated learning that uses proactive auditing metrics and a dynamical model to detect and mitigate adaptive backdoor attacks more effectively than passive methods.

Contribution

The paper proposes a novel active auditing framework with new metrics and a topology-aware defense placement strategy, advancing beyond traditional passive detection in federated learning.

Findings

Effective detection of latent backdoors using proactive metrics

Improved resilience against adaptive backdoor attacks

Maintains primary task utility while defending against stealthy threats

Abstract

Decentralized Federated Learning (DFL) remains highly vulnerable to adaptive backdoor attacks designed to bypass traditional passive defense metrics. To address this limitation, we shift the defensive paradigm toward a novel active, interventional auditing framework. First, we establish a dynamical model to characterize the spatiotemporal diffusion of adversarial updates across complex graph topologies. Second, we introduce a suite of proactive auditing metrics, stochastic entropy anomaly, randomized smoothing Kullback-Leibler divergence, and activation kurtosis. These metrics utilize private probes to stress-test local models, effectively exposing latent backdoors that remain invisible to conventional static detection. Furthermore, we implement a topology-aware defense placement strategy to maximize global aggregation resilience. We provide theoretical property for the system's convergence under co-evolving attack and defense dynamics. Numeric empirical evaluations across diverse architectures demonstrate that our active framework is highly competitive with state-of-the-art defenses in mitigating stealthy, adaptive backdoors while preserving primary task utility.