EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

VFL-Cafe: Communication-Efficient Vertical Federated Learning via Dynamic Caching and Feature Selection

Jiahui Zhou, Han Liang, Tian Wu, Xiaoxi Zhang, Yu Jiang, Chee Wei Tan iD

Entropy · 14.01.2025 · 27 (1) · 66

DOI: 10.3390/e27010066 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Entropy, online erschienen am 14.01.2025. Verfasst von Jiahui Zhou und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer and Science and Engineering, Sun Yat-sen University, Guangzhou 510275, China. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NSFC, Guangdong Basic and 493 Applied Basic Research Foundation, Guangdong Basic and Applied Basic Research Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Vertical Federated Learning (VFL) is a promising category of Federated Learning that enables collaborative model training among distributed parties with data privacy protection. Due to its unique training architecture, a key challenge of VFL is high communication cost due to transmitting intermediate results between the Active Party and Passive Parties. Current communication-efficient VFL methods rely on using stale results without meticulous selection, which can impair model accuracy, particularly in noisy data environments. To address these limitations, this work proposes VFL-Cafe, a new VFL training method that leverages dynamic caching and feature selection to boost communication efficiency and model accuracy. In each communication round, the employed caching scheme allows multiple batches of intermediate results to be cached and strategically reused by different parties, reducing the communication overhead while maintaining model accuracy. Additionally, to eliminate the negative impact of noisy features that may undermine the effectiveness of using stale results to reduce communication rounds and incur significant model degradation, a feature selection strategy is integrated into each round of local updates. Theoretical analysis is then conducted to provide guidance on cache configuration, optimizing performance. Finally, extensive experimental results validate VFL-Cafe’s efficacy, demonstrating remarkable improvements in communication efficiency and model accuracy.

Schlagwörter

Stichwörter der Autor:innen

communication efficientdynamic cachingfeature selectionvertical federated learning

Autor:innen und Institutionen

  • School of Computer and Science and Engineering, Sun Yat-sen University, Guangzhou 510275, China.
  • College of Computing and Data Science, Nanyang Technological University in Singapore, Singapore 639798, Singapore.

ORCID-Profile: Chee Wei Tan

Förderung

  • NSFC
  • Guangdong Basic and 493 Applied Basic Research Foundation
  • Guangdong Basic and Applied Basic Research Foundation
  • Young Outstanding Award under the Zhujiang Talent Plan of Guangdong Province
  • Singapore Ministry of Education Academic Fund
  • NTU startup

Zitieren

APA

Zhou, J., Liang, H., Wu, T., Zhang, X., Jiang, Y., & Tan, C. W. (2025). VFL-Cafe: Communication-Efficient Vertical Federated Learning via Dynamic Caching and Feature Selection. Entropy, 27(1), 66. https://doi.org/10.3390/e27010066

Vancouver

Zhou J, Liang H, Wu T, Zhang X, Jiang Y, Tan CW. VFL-Cafe: Communication-Efficient Vertical Federated Learning via Dynamic Caching and Feature Selection. Entropy. 2025;27(1):66. doi:10.3390/e27010066

BibTeX
@article{Zhou2025,
  author = {Zhou, Jiahui and Liang, Han and Wu, Tian and Zhang, Xiaoxi and Jiang, Yu and Tan, Chee Wei},
  title = {VFL-Cafe: Communication-Efficient Vertical Federated Learning via Dynamic Caching and Feature Selection},
  journal = {Entropy},
  year = {2025},
  volume = {27},
  number = {1},
  pages = {66},
  doi = {10.3390/e27010066}
}
Kostenlos starten →