Personalized Federated Learning With Contrastive Momentum
IEEE Transactions on Big Data · 2025-10 · 11 (5) · 2184-2194
DOI: 10.1109/tbdata.2024.3403387 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Big Data, 2025-10. Verfasst von Sen Fu und 3 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 48 Literaturangaben.
Abstract (Originalsprache)
In this paper, we propose pFedMo, a personalized federated learning algorithm with contrastive momentum. In pFedMo, we design a score function to personalize worker models by distilling knowledge from the aggregator's representation model so as to address the non-i.i.d. issue. To accelerate the convergence, we leverage the momentum acceleration on both the worker side and the aggregator side. However, the typical momentum without personalization does not suit well for the worker models with personalization, influencing convergence performance. To address this, we develop a personalized/contrastive momentum method for efficient momentum acceleration. We provide mathematical proof for the convergence of pFedMo on non-i.i.d. data. Extensive experiments based on real-world datasets and IoT system are conducted, verifying that pFedMo outperforms existing mainstream benchmarks, and achieves up to 35.90% accuracy increase and 3.64x training time speedup under a wide range of settings.
Zitieren
APA
Fu, S., Yang, Z., Hu, C., & Bao, W. (2025). Personalized Federated Learning With Contrastive Momentum. IEEE Transactions on Big Data, 11(5), 2184-2194. https://doi.org/10.1109/tbdata.2024.3403387
Vancouver
Fu S, Yang Z, Hu C, Bao W. Personalized Federated Learning With Contrastive Momentum. IEEE Transactions on Big Data. 2025;11(5):2184-2194. doi:10.1109/tbdata.2024.3403387
BibTeX
@article{Fu2025,
author = {Fu, Sen and Yang, Zhengjie and Hu, Chuang and Bao, Wei},
title = {Personalized Federated Learning With Contrastive Momentum},
journal = {IEEE Transactions on Big Data},
year = {2025},
volume = {11},
number = {5},
pages = {2184--2194},
doi = {10.1109/tbdata.2024.3403387}
}
EUVIMED