Robust and Sparse Personalized Federated Learning via Gamma-Divergence

Abstract

To address the challenges of statistical heterogeneity across clients and within-client data contamination in federated learning, we propose a robust and sparse personalized federated learning method based on the gamma-divergence, termed Gamma-PFL. Specifically, we combine a loss function based on the gamma-divergence with fusion and sparse regularization terms to exploit similarities in regression coefficients across clients while enabling robust estimation of personalized models and variable selection. Furthermore, to efficiently solve the estimation problem in the federated server--client system, we develop an optimization algorithm that combines the Alternating Direction Method of Multipliers with the Majorization--Minimization algorithm. Numerical simulations showed that Gamma-PFL maintained relatively stable estimation accuracy as the contamination proportion or outlier magnitude increased, demonstrating its robustness to within-client data contamination. In addition, a real data analysis showed that Gamma-PFL achieved predictive performance comparable to or better than that of an existing personalized federated learning method.

Personalized Federated Learning through Global Knowledge Distillation and Local Head Adaptation
Polycarpo Souza Neto, José Mairton Barros da Silva Júnior, Charles Casimiro Cavalcante
arXiv15 Sep 2026PreprintarXiv licenceFull textStatisticsarXiv
Geographically Regularized AUC-Maximizing Personalized Federated Learning
Mayu Hiraishi, Kensuke Tanioka, Toshio Shimokawa
arXiv8 Sep 2026PreprintarXiv licenceFull textStatisticsarXiv
Personalized and differentially private federated learning for mobility data mining via the TrajectoryGuard framework
Ali Alqazzaz
PLOS One29 Oct 2026Peer-reviewedCC BYFull textGeneral earth science StatisticsPLOSEurope PMCPMC
Distributionally Robust Federated Learning with Multi-Source Data
Yingzhu Liu, Zhongkui Li, Pengcheng You +1
arXiv17 Sep 2026PreprintarXiv licenceFull textOptimisation & controlarXiv
Byzantine-Robust Federated Representation Learning
Leonardo F. Toso, James Anderson, Rafael Pinot +1
arXiv29 Sep 2026PreprintarXiv licenceFull textOptimisation & controlarXiv

Cite

BibTeX RIS

BibTeX

@misc{akimaru2026robust,
  title = {{Robust and Sparse Personalized Federated Learning via Gamma-Divergence}},
  author = {Kotaro Akimaru and Kensuke Tanioka},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.27827},
  eprint = {2610.27827},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.27827},
}

RIS

TY  - UNPB
TI  - Robust and Sparse Personalized Federated Learning via Gamma-Divergence
AU  - Kotaro Akimaru
AU  - Kensuke Tanioka
PY  - 2026
DA  - 2026/10/20
PB  - arXiv
DO  - 10.48550/arxiv.2610.27827
UR  - https://arxiv.org/abs/2610.27827
AB  - To address the challenges of statistical heterogeneity across clients and within-client data contamination in federated learning, we propose a robust and sparse personalized federated learning method based on the gamma-divergence, termed Gamma-PFL. Specifically, we combine a loss function based on the gamma-divergence with fusion and sparse regularization terms to exploit similarities in regression coefficients across clients while enabling robust estimation of personalized models and variable selection. Furthermore, to efficiently solve the estimation problem in the federated server--client system, we develop an optimization algorithm that combines the Alternating Direction Method of Multipliers with the Majorization--Minimization algorithm. Numerical simulations showed that Gamma-PFL maintained relatively stable estimation accuracy as the contamination proportion or outlier magnitude increased, demonstrating its robustness to within-client data contamination. In addition, a real data analysis showed that Gamma-PFL achieved predictive performance comparable to or better than that of an existing personalized federated learning method.
ER  -

Abstract

Related

Cite