ASMR: Angular Support for Malfunctioning Client Resilience in Federated Learning

2025-08-04

AI Summary

In federated learning, faulty model updates from anomalous clients—caused by data heterogeneity, system failures, or adversarial attacks—severely degrade global model performance. Existing defense mechanisms often rely on strong prior assumptions (e.g., known upper bounds on the number of malicious clients), limiting their practical applicability. To address this, we propose a prior-free, dynamic robust aggregation framework. Our method introduces an angular distance–based client update evaluation paradigm, leveraging gradient direction analysis to adaptively construct decision boundaries for anomaly detection and filtering—without requiring hyperparameter tuning or pre-specified corruption rates. Evaluated on histopathological image classification tasks, our approach significantly improves both model accuracy and convergence stability under diverse attack and system failure scenarios, consistently outperforming state-of-the-art defense methods.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Federated Learning (FL) allows the training of deep neural networks in a distributed and privacy-preserving manner. However, this concept suffers from malfunctioning updates sent by the attending clients that cause global model performance degradation. Reasons for this malfunctioning might be technical issues, disadvantageous training data, or malicious attacks. Most of the current defense mechanisms are meant to require impractical prerequisites like knowledge about the number of malfunctioning updates, which makes them unsuitable for real-world applications. To counteract these problems, we introduce a novel method called Angular Support for Malfunctioning Client Resilience (ASMR), that dynamically excludes malfunctioning clients based on their angular distance. Our novel method does not require any hyperparameters or knowledge about the number of malfunctioning clients. Our experiments showcase the detection capabilities of ASMR in an image classification task on a histopathological dataset, while also presenting findings on the significance of dynamically adapting decision boundaries.

Problem

• Detects malfunctioning clients in Federated Learning using angular distance
• Eliminates need for hyperparameters or prior knowledge of malfunctioning clients
• Improves global model resilience against technical issues and malicious attacks

Innovation

• Dynamic exclusion of malfunctioning clients
• Angular distance-based resilience method
• No hyperparameters or prior knowledge needed