EBS-CFL: Efficient and Byzantine-robust Secure Clustered Federated Learning

2025-04-11

AAAI Conference on Artificial Intelligence

AI Summary

This paper addresses the challenge of coexisting cluster-identity concealment and Byzantine attacks in clustering federated learning (CFL) under data heterogeneity. Methodologically, it proposes the first privacy-preserving robust clustered collaborative training framework: (1) secure aggregation via secure multi-party computation to prevent cluster-identity leakage; (2) a gradient-correlation-based detection and weighted aggregation mechanism to ensure both Byzantine resilience and convergence; and (3) a lightweight client-side encoding verification scheme to jointly guarantee authentication and confidentiality. Theoretical analysis formally proves the framework’s security. Experiments demonstrate that, when the number of clusters (m = 1), the proposed method achieves at least an (O(log n)) speedup in computational efficiency over baselines; its communication and computational complexities are (O(ml + m^2)) and (O(m^2 l)), respectively.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Natural Language Processing: Safety and Robustness
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Security and Privacy: Privacy-enhancing technologies
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Despite federated learning (FL)'s potential in collaborative learning, its performance has deteriorated due to the data heterogeneity of distributed users. Recently, clustered federated learning (CFL) has emerged to address this challenge by partitioning users into clusters according to their similarity. However, CFL faces difficulties in training when users are unwilling to share their cluster identities due to privacy concerns. To address these issues, we present an innovative Efficient and Robust Secure Aggregation scheme for CFL, dubbed EBS-CFL. The proposed EBS-CFL supports effectively training CFL while maintaining users' cluster identity confidentially. Moreover, it detects potential poisonous attacks without compromising individual client gradients by discarding negatively correlated gradients and aggregating positively correlated ones using a weighted approach. The server also authenticates correct gradient encoding by clients. EBS-CFL has high efficiency with client-side overhead O(ml + m^2) for communication and O(m^2l) for computation, where m is the number of cluster identities, and l is the gradient size. When m = 1, EBS-CFL's computational efficiency of client is at least O(log(n)) times better than comparison schemes, where n is the number of clients. In addition, we validate the scheme through extensive experiments. Finally, we theoretically prove the scheme's security.

Problem

• Addresses data heterogeneity in federated learning via clustering
• Ensures privacy by securing user cluster identities
• Detects and mitigates poisonous attacks in gradient aggregation

Innovation

• Secure clustered federated learning with privacy
• Detects attacks by filtering malicious gradients
• Efficient client-side computation and communication