Robust Reputation-Driven Crowdsourced Federated Learning

2026-08-09

AI Summary

This work addresses the vulnerability of existing reputation mechanisms in crowdsourced federated learning to adaptive stealthy attackers, who can evade detection and gradually accumulate trust to compromise the global model. To counter this threat, we propose R2CFL, a novel framework that dynamically couples robust reputation modeling with a nearest-neighbor mixture (R2-NNM) defense mechanism during aggregation, simultaneously filtering malicious updates and evolving participant reputations. This approach effectively disrupts attackers’ trust accumulation and ensures that reputation scores accurately reflect the statistical robustness of the underlying defense. Experimental results demonstrate that R2-NNM matches or exceeds state-of-the-art Byzantine-robust and backdoor defense methods under adaptive attacks, and when integrated with existing detectors, precisely characterizes their true and false positive behaviors.

Abstract

Crowdsourced Federated Learning (CrowdFL) extends traditional federated learning by enabling open and heterogeneous participation through a crowdsourcing paradigm. In this setting, reputation-driven incentive mechanisms are commonly employed to guide worker selection and enhance trustworthiness. While such approaches improve participant reliability, existing frameworks largely overlook the quantification of their robustness against stealthy adversaries, particularly those capable of evading standard detection mechanisms. To fill this gap, this paper proposes R2CFL, a robust reputation-driven CrowdFL framework. R2CFL introduces a robust reputation model coupled with a nearest neighbor mixing (R2-NNM) defense mechanism that links reputation evolution with the filtering of updates during aggregation. The proposed mechanism prevents stealthy attackers from gradually accumulating trust and influencing future tasks. Experimental results demonstrate that R2-NNM matches or surpasses state-of-the-art Byzantine-robust and backdoor defense mechanisms against adaptive attackers. Furthermore, when integrated with existing detect-and-filter defenses, the proposed reputation model faithfully captures the statistical robustness of the underlying defense by producing reputation scores that closely reflect its true positive and false positive characteristics.

Problem

• Crowdsourced Federated Learning
• reputation mechanism
• stealthy adversaries
• robustness
• Byzantine attacks

Innovation

• Robust Reputation
• Crowdsourced Federated Learning
• Nearest Neighbor Mixing
• Stealthy Adversary Defense
• Byzantine Robustness