EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

PAG-FL: Personalization via Adaptive Gradient Filtering in Federated Learning

Yuri Seo, Hyeon-Ki Jo, Eui-Nam Huh

IEEE Access · 2026 · 14 · 25840-25853

DOI: 10.1109/access.2026.3664853 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Yuri Seo und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: IITP (Institute of Information & Coummunications Technology Planning & Evaluation)-ITRC (Information Technology Research Center) grant funded by the Korea gove…, Korea government (MSIT) Development of AI Semiconductor Cloud Platform Establishment and Optimization Technologies, 50%.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Personalized Federated Learning (PFL) has emerged as a key paradigm for delivering personalized AI services across diverse clients. While prior research has primarily focused on handling data heterogeneity, recent studies have additionally targeted noise and communication efficiency. However, most approaches exchange knowledge only at the logit or prototype level, which limits representation sharing and struggles to mitigate both data heterogeneity and noise. Moreover, communication overhead accumulated over hundreds of training rounds still imposes substantial bandwidth costs, and fixed personalization ratios hinder effective optimization to non-IID data. To address these challenges, we propose Personalization via Adaptive Gradient Filtering in Federated Learning (PAG-FL), a communication-efficient and robust PFL framework. PAG-FL reduces communication costs by collecting only class-wise average embeddings and local prediction heads from clients. It further alleviates embedding bias through $L_{2}$ normalization and a projection layer, while optimizing personalization ratios using thresholds derived from the interquartile range (IQR) of gradients. Extensive experiments on Fashion-MNIST, CIFAR-10, and CIFAR-100 demonstrate that PAG-FL consistently outperforms baseline methods with statistically significant margins. Specifically, under 20% label noise and non-IID conditions ( $\alpha =0.3, \beta =3.5$ ), PAG-FL achieves accuracy improvements of up to 19% and F1 Score gains of up to 17% (e.g., in Fashion-MNIST under a fixed users scenario) compared to the best baselines. In fixed and subset user scenarios, it maintains this robustness, yielding substantial accuracy gains of approximately 19%. Furthermore, PAG-FL significantly accelerates convergence, reducing communication rounds by up to 96% to reach a 50% accuracy threshold on Fashion-MNIST, and decreasing communication overhead by up to 1, $700\times $ on CIFAR-10 compared to FedAvg.

Förderung

  • IITP (Institute of Information & Coummunications Technology Planning & Evaluation)-ITRC (Information Technology Research Center) grant funded by the Korea gove…
  • Korea government (MSIT) Development of AI Semiconductor Cloud Platform Establishment and Optimization Technologies, 50%

Zitieren

APA

Seo, Y., Jo, H., & Huh, E. (2026). PAG-FL: Personalization via Adaptive Gradient Filtering in Federated Learning. IEEE Access, 14, 25840-25853. https://doi.org/10.1109/access.2026.3664853

Vancouver

Seo Y, Jo H, Huh E. PAG-FL: Personalization via Adaptive Gradient Filtering in Federated Learning. IEEE Access. 2026;14:25840-25853. doi:10.1109/access.2026.3664853

BibTeX
@article{Seo2026,
  author = {Seo, Yuri and Jo, Hyeon-Ki and Huh, Eui-Nam},
  title = {PAG-FL: Personalization via Adaptive Gradient Filtering in Federated Learning},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {25840--25853},
  doi = {10.1109/access.2026.3664853}
}
Kostenlos starten →