ERPPFL : An Efficient and Robust Privacy‐Preserving Federated Learning Scheme
Concurrency and Computation: Practice and Experience · 2026-07 · 38 (13)
DOI: 10.1002/cpe.70799 ↗Kurz gefasst
Fachartikel in der Zeitschrift Concurrency and Computation: Practice and Experience, online erschienen am 09.07.2026. Verfasst von Gongli Li und 3 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben. Gefördert von: Natural Science Foundation of Henan Province.
Abstract (Originalsprache)
To address the high computational and communication overhead of existing privacy‐preserving schemes for federated learning, a lightweight secure encryption scheme is proposed. This scheme employs a negotiation‐free masking‐based blinding technique to protect local model parameters. Meanwhile, to improve the efficiency of secure model aggregation, a dual secure aggregation strategy is proposed. This strategy enables the server to directly perform efficient dual aggregation over ciphertexts without accessing any clients' local model parameters or mask information and to accurately update the global model within the ciphertext domain. In addition, to cope with client dropouts caused by network instability, a fault‐tolerant mechanism is incorporated, which significantly enhances the robustness of the proposed scheme. Experimental results demonstrate that, while ensuring data privacy, the proposed scheme maintains high model training accuracy. Moreover, compared with the fastest existing scheme, the average runtime is reduced by 37%, and the average communication overhead is reduced by 53%.
Förderung
- Natural Science Foundation of Henan Province
Zitieren
APA
Li, G., Liu, Y., Du, R., & Lei, H. (2026). ERPPFL : An Efficient and Robust Privacy‐Preserving Federated Learning Scheme. Concurrency and Computation: Practice and Experience, 38(13). https://doi.org/10.1002/cpe.70799
Vancouver
Li G, Liu Y, Du R, Lei H. ERPPFL : An Efficient and Robust Privacy‐Preserving Federated Learning Scheme. Concurrency and Computation: Practice and Experience. 2026;38(13). doi:10.1002/cpe.70799
BibTeX
@article{Li2026,
author = {Li, Gongli and Liu, Yihe and Du, Ruiying and Lei, Hongzhi},
title = {ERPPFL : An Efficient and Robust Privacy‐Preserving Federated Learning Scheme},
journal = {Concurrency and Computation: Practice and Experience},
year = {2026},
volume = {38},
number = {13},
doi = {10.1002/cpe.70799}
}
EUVIMED