EUVIMEDEUVIMED entdecken
Crossref

Communication-Efficient Personalized Federated Learning With Incentive-Driven Adaptive Model Pruning and Neighbor Selection

Ting Li, Huiting Mo, Tao Ouyang, Yinlong Liu, Kai Yang

IEEE Internet of Things Journal · 2026-08 · 13 (16) · 37084-37100

DOI: 10.1109/jiot.2026.3701375 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Internet of Things Journal, 2026-08. Verfasst von Ting Li und 4 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Gefördert von: Guangxi Young Talent Universal Support Policy Project, Natural Science Foundation of Guangxi Zhuang Autonomous Region.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Personalized federated learning (PFL) enables client-specific models to address data heterogeneity but suffers from high communication overhead and unstable participation in resource-constrained and self-interested environments. Existing mainstream approaches predominantly prioritize training process efficiency but lack explicit consideration of incentive mechanisms and rational client behaviors, which may lead to clients behaving conservatively, reducing participation, and limiting the effectiveness and scalability of PFL in real-world deployments. In this article, we propose IncenPNS, the first incentive-driven adaptive framework that jointly optimizes model pruning, neighbor selection, and incentive mechanisms for communication-efficient PFL. IncenPNS formulates the joint design as a unified optimization objective that balances personalization performance, communication efficiency, and incentive utility under dynamic and heterogeneous environments. To solve the resulting coupled and high-dimensional decision problem, we develop a multiagent soft actor–critic (MASAC)-based learning algorithm that enables clients to adapt pruning rates and collaboration decisions through online interaction. Moreover, a budget-balanced incentive mechanism is incorporated to operate under partial observability, aligning individual rationality (IR) with system-level objectives and ensuring reliable participation of self-interested clients. Extensive experiments on representative benchmarks show that IncenPNS achieves up to a 34.8% reduction in communication cost, 68.8% faster convergence, and 83.3% improvement in personalization accuracy compared with state-of-the-art baselines.

Förderung

  • Guangxi Young Talent Universal Support Policy Project
  • Natural Science Foundation of Guangxi Zhuang Autonomous Region

Zitieren

APA

Li, T., Mo, H., Ouyang, T., Liu, Y., & Yang, K. (2026). Communication-Efficient Personalized Federated Learning With Incentive-Driven Adaptive Model Pruning and Neighbor Selection. IEEE Internet of Things Journal, 13(16), 37084-37100. https://doi.org/10.1109/jiot.2026.3701375

Vancouver

Li T, Mo H, Ouyang T, Liu Y, Yang K. Communication-Efficient Personalized Federated Learning With Incentive-Driven Adaptive Model Pruning and Neighbor Selection. IEEE Internet of Things Journal. 2026;13(16):37084-37100. doi:10.1109/jiot.2026.3701375

BibTeX
@article{Li2026,
  author = {Li, Ting and Mo, Huiting and Ouyang, Tao and Liu, Yinlong and Yang, Kai},
  title = {Communication-Efficient Personalized Federated Learning With Incentive-Driven Adaptive Model Pruning and Neighbor Selection},
  journal = {IEEE Internet of Things Journal},
  year = {2026},
  volume = {13},
  number = {16},
  pages = {37084--37100},
  doi = {10.1109/jiot.2026.3701375}
}
Kostenlos starten →