VeriFL: Communication-Efficient and Fast Verifiable Aggregation for Federated Learning

• Published in: IEEE transactions on information forensics and security Vol. 16; pp. 1736 - 1751
• Main Authors: Guo, Xiaojie, Liu, Zheli, Li, Jin, Gao, Jiqiang, Hou, Boyu, Dong, Changyu, Baker, Thar
• Format: Journal Article
• Language: English
• Published: IEEE 2021
• Subjects: Bandwidth; Collaborative work; commitment; Federated learning; linearly homomorphic hash; machine learning; Privacy; Protocols; Servers; Synchronization; Training; verifiable aggregation
• ISSN: 1556-6013; 1556-6021
• DOI: 10.1109/TIFS.2020.3043139

Federated learning (FL) enables a large number of clients to collaboratively train a global model through sharing their gradients in each synchronized epoch of local training. However, a centralized server used to aggregate these gradients can be compromised and forge the result in order to violate privacy or launch other attacks, which incurs the need to verify the integrity of aggregation. In this work, we explore how to design communication-efficient and fast verifiable aggregation in FL. We propose V eri FL, a verifiable aggregation protocol, with <inline-formula> <tex-math notation="LaTeX">O(N) </tex-math></inline-formula> (dimension-independent) communication and <inline-formula> <tex-math notation="LaTeX">O(N+ d) </tex-math></inline-formula> computation for verification in each epoch, where <inline-formula> <tex-math notation="LaTeX">N </tex-math></inline-formula> is the number of clients and <inline-formula> <tex-math notation="LaTeX">d </tex-math></inline-formula> is the dimension of gradient vectors. Since <inline-formula> <tex-math notation="LaTeX">d </tex-math></inline-formula> can be large in some real-world FL applications (e.g., 100K), our dimension-independent communication is especially desirable for clients with limited bandwidth and high-dimensional gradients. In addition, the proposed protocol can be used in the FL setting where secure aggregation is needed or there is a subset of clients dropping out of protocol execution. Experimental results indicate that our protocol is efficient in these settings.