EUVIMEDEUVIMED entdecken
Crossref

Decentralized Federated Learning With Period Gradient Tracking Over Time-Varying Networks

Fang Shi, Yuehong Chen, Qiong Huang, Tiansheng Huang, Guozhi Liu, Li Shen

IEEE Transactions on Parallel and Distributed Systems · 2026-07 · 37 (7) · 1530-1540

DOI: 10.1109/tpds.2026.3685244 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Parallel and Distributed Systems, 2026-07. Verfasst von Fang Shi und 5 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Gefördert von: Science and Technology Program of Guangzhou, 2025 Guangdong Provincial Key Research Platforms of Universities, National Key R&D Projects.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

To address the communication challenges associated with Federated Learning (FL), Decentralized Federated Learning (DFL) eliminates the central server and trains the model with decentralized method, enabling each client to only communicate with its neighbors. However, per our analysis, model trained with DFL experiences performance degradation because of data-heterogeneity and time-varying topologies. To address these issues, we propose a Dynamic K-step Gradient Tracking (DKGT) method to enhance the performance of DFL over time varying networks. Specifically, DKGT employs K-step local updates and gradient tracking to reduce the communication cost and the variance from heterogeneous data distribution, and we use dynamic gradient tracking parameter to correct gradient over time varying graph. Theoretically, we derive a universal convergence rate for smooth and non-convex problem at the rate of $\mathcal {O}(\frac{(f(\mathbf{x}_{0})-f(\mathbf{x}^{*}))}{\sqrt{T}(L\sqrt{KN})^{-1}-\tau (pKL\sqrt{TKN})^{-1}}+\frac{\sigma ^{2}}{KTN\tau (pK-\tau )})$ O ( ( f ( x 0 ) - f ( x * ) ) T ( L K N ) - 1 - τ ( p K L T K N ) - 1 + σ 2 K T N τ ( p K - τ ) ) , that $\tau$ τ and $p$ p respectively represent the time window length and the connectivity of time-varying networks. Experimentally, we illustrate the robustness and effectiveness of this heterogeneity correction on extensive non-convex neural network training tasks over different topologies and dynamic network settings.

Förderung

  • Science and Technology Program of Guangzhou
  • 2025 Guangdong Provincial Key Research Platforms of Universities
  • National Key R&D Projects
  • National Natural Science Foundation of China
  • Basic and Applied Basic Research Foundation of Guangdong Province
  • Shenzhen Municipal Fundamental Research Program
  • Natural Science Foundation
  • CCF-Tencent Rhino-Bird Open Research Fund

Zitieren

APA

Shi, F., Chen, Y., Huang, Q., Huang, T., Liu, G., & Shen, L. (2026). Decentralized Federated Learning With Period Gradient Tracking Over Time-Varying Networks. IEEE Transactions on Parallel and Distributed Systems, 37(7), 1530-1540. https://doi.org/10.1109/tpds.2026.3685244

Vancouver

Shi F, Chen Y, Huang Q, Huang T, Liu G, Shen L. Decentralized Federated Learning With Period Gradient Tracking Over Time-Varying Networks. IEEE Transactions on Parallel and Distributed Systems. 2026;37(7):1530-1540. doi:10.1109/tpds.2026.3685244

BibTeX
@article{Shi2026,
  author = {Shi, Fang and Chen, Yuehong and Huang, Qiong and Huang, Tiansheng and Liu, Guozhi and Shen, Li},
  title = {Decentralized Federated Learning With Period Gradient Tracking Over Time-Varying Networks},
  journal = {IEEE Transactions on Parallel and Distributed Systems},
  year = {2026},
  volume = {37},
  number = {7},
  pages = {1530--1540},
  doi = {10.1109/tpds.2026.3685244}
}
Kostenlos starten →