logo
科言猫学术研究的AI总结
language
手机端手机端
  • Phone Feature微信小程序
简体中文简体中文
notice消息通知
博客博客
delete
arrow
返回

LOGO-CL: Accelerating semi-supervised federated learning in edge computing

delete2025-07-04
delete0
PRE
AI
Y
Yang Xu
Q
Q. Y. Wang
H
Hongli Xu
Y
Yunming Liao
L
Liusheng Huang
X
Xin Hang
DOI：10.1016/j.comnet.2025.111496delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
联邦学习（FL）在边缘计算中跨客户端训练模型方面获得了重要关注。标注数据的稀缺性为联邦学习在专业领域（如医学诊断）带来了关键挑战，这促使联邦半监督学习（FSSL）的发展以利用未标注数据。尽管先前的工作能够充分利用未标注数据，但它们不可避免地面临两大固有挑战：通信资源有限和数据异质性。现有的FSSL方法通常在训练过程中为全局更新频率和/或本地更新频率分配固定值，并忽略了未标注数据分布的影响，这严重阻碍了收敛稳定性和训练效率。为解决这些问题，我们提出了一种新型FSSL框架，称为LOGO-CL。具体而言，我们通过分析这两个超参数对训练过程的耦合影响，联合优化本地和全局更新频率。我们开发了一种基于多臂赌博机（MAB）的在线算法，以自适应地确定多样化的本地更新频率以及适当的全局更新频率，从而提高训练效率。此外，LOGO-CL利用对比学习技术，通过在全局和本地训练过程中引入正则化项，缓解未标注数据异质性的影响，确保在数据分布不平衡时仍具有鲁棒性。大量实验表明，与基线方法相比，当达到相同目标精度时，LOGO-CL实现了2.3倍的模型训练加速，并将通信成本降低了44%。此外，在具有中等噪声的数据集上，LOGO-CL将模型精度提高了高达4.8%。

期刊

Computer Networks 封面图
Computer Networks
IF:
4.6
论文数:
1.7K
被引数:
1.6W

机构

暂无机构信息
引用论文

引用论文

Overcoming Noisy Labels and Non-IID Data in Edge Federated Learning
errIEEE TRANSACTIONS ON MOBILE COMPUTING
IF9.2
err2024-12-01
err0
PREAI
errXu, Yang; Liao, Yunming; Wang, Lun; Xu, Hongli; Jiang, Zhida; Zhang, Wuyang
err分享
err收藏
Distillation-Based Semi-Supervised Federated Learning for Communication-Efficient Collaborative Training With Non-IID Private Data
errIEEE TRANSACTIONS ON MOBILE COMPUTING
IF9.2
err2023-01-01
err112
errOAAI
errItahara, Sohei; Nishio, Takayuki; Koda, Yusuke; Morikura, Masahiro; Yamamoto, Koji
err分享
err收藏
Internet of things and data mining: An application oriented survey
errJournal of King Saud University - Computer and Information Sciences
IF0
err2022-06-01
err0
errOAAI
errPriyank Sunhare; Rameez R. Chowdhary; Manju K. Chattopadhyay
err分享
err收藏
Federated Learning With Non-IID Data: A Survey非IID数据的联邦学习: 一项调查
errIEEE INTERNET OF THINGS JOURNAL
IF8.9
err2024-06-01
err44
PREAI
errLu, Zili; Pan, Heng; Dai, Yueyue; Si, Xueming; Zhang, Yan
err分享
err收藏
Tackling Noisy Clients in Federated Learning with End-to-end Label Correction联邦学习中端到端标签修正以解决噪声客户端问题
errProceedings of the 33rd ACM International Conference on Information and Knowledge Management
IF0
err2024-10-21
err0
errOAAI
errXuefeng Jiang; Sheng Sun; Jia Li; Jingjing Xue; Runhan Li; Zhiyuan Wu; Gang Xu; Yuwei Wang; Min Liu
err分享
err收藏
Federated learning review: Fundamentals, enabling technologies, and future applications联合学习评论: 基础，使能技术和未来的应用
errINFORMATION PROCESSING & MANAGEMENT
IF6.9
err2022-11-01
err148
PREAI
errBanabilah, Syreen; Aloqaily, Moayad; Alsayed, Eitaa; Malik, Nida; Jararweh, Yaser
err分享
err收藏
Adaptive Federated Learning in Resource Constrained Edge Computing Systems资源受限边缘计算系统中的自适应联合学习
errIEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
IF17.2
err2019-06-01
err1.4K
errOAAI
errWang, Shiqiang; Tuor, Tiffany; Salonidis, Theodoros; Leung, Kin K.; Makaya, Christian; He, Ting; Chan, Kevin
err分享
err收藏
Can irrelevant data help semi-supervised learning, why and how?
errProceedings of the 20th ACM international conference on Information and knowledge management
IF0
err2011-10-24
err0
PREAI
errHaiqin Yang; Shenghuo Zhu; Irwin King; Michael R. Lyu
err分享
err收藏
Deep learning and alternative learning strategies for retrospective real-world clinical data回顾性真实世界临床数据的深度学习和替代学习策略
errNPJ DIGITAL MEDICINE
IF15.1
err2019-05-30
err149
errOAAI
errChen, David; Liu, Sijia; Kingsbury, Paul; Sohn, Sunghwan; Storlie, Curtis B.; Habermann, Elizabeth B.; Naessens, James M.; Larson, David W.; Liu, Hongfang
err分享
err收藏
Class Balanced Adaptive Pseudo Labeling for Federated Semi-Supervised Learning
err2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
IF0
err2023-06-01
err0
PREAI
errMing Li; Qingli Li; Yan Wang
err分享
err收藏
学者 查看更多内容