ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展

• •    下一篇

联邦学习开源框架综述

林伟伟,石方,曾岚,李董东,许银海,刘波   

  1. (华南理工大学计算机科学与工程学院 广州 510006)(吉林大学数学学院 长春 130012)(华南师范大学计算机学院 广州 510631)(linww@scut.edu.cn)
  • 出版日期: 2022-08-24

A Survey of Federated Learning Open-Source Frameworks

Lin Weiwei, Shi Fang, Zeng Lan, Li Dongdong, Xu Yinhai, Liu Bo   

  1. (School of Computer Science & Engineering, South China University of Technology, Guangzhou 510006)(School of Mathematics, Jilin University, Changchun 130012)(School of Computer Science, South China Normal University, Guangzhou 510631)
  • Online: 2022-08-24

摘要: 近年来,联邦学习作为破解数据共享壁垒的有效解决方案被广泛关注,并被逐步应用于医疗、金融和智慧城市等领域.联邦学习框架是联邦学习学术研究和工业应用的基石.虽然Google、OpenMined、微众银行和百度等企业开源了各自的联邦学习框架和系统,然而,目前缺少对这些联邦学习开源框架的技术原理、适用场景、存在问题等的深入研究和比较.为此,根据各开源框架在业界的受众程度,选取了目前应用较广和影响较大的联邦学习开源框架进行深入研究.针对不同类型的联邦学习框架,首先分别从系统架构和系统功能2个层次对各框架进行剖析;其次从隐私机制、机器学习算法、计算范式、学习类型、训练架构、通信协议、可视化等多个维度对各框架进行深入对比分析.而且,为了帮助读者更好地选择和使用开源框架实现联邦学习应用,给出了面向2个不同应用场景的联邦学习实验.最后,基于目前框架存在的开放性问题,从隐私安全、激励机制、跨框架交互等方面讨论了未来可能的研究发展方向,旨在为开源框架的开发创新、架构优化、安全改进以及算法优化等提供参考和思路.

关键词: 联邦学习, 开源框架, 模型训练, 机器学习, 大数据

Abstract: In recent years, FL (federated learning) has gained widespread attention as an effective solution to breaking down the barrier to data sharing and is being progressively applied in areas such as healthcare, finance, and smart cities. FL frameworks are the cornerstone of academic research and industrial applications. Although companies such as Google, OpenMined, WeBank, and Baidu have open-sourced their own FL frameworks and systems, there is a lack of in-depth research and comparison of the technical principles, applicability scenarios, and problems of these FL open-source frameworks. For this reason, according to the preference level of each open-source framework in the industry, we select the widely used open-source frameworks to analyze. For the different types of FL frameworks, firstly, the system architecture and system function are analyzed, respectively. Secondly, we compare and analyze each framework from the aspects of privacy mechanism, machine learning algorithm, computing paradigm, learning type, training architecture, communication protocol, visualization, etc. Moreover, this paper presents two FL experiments for different application scenarios to help the readers choose and use the open-source framework to implement federated learning applications. Finally, based on the openness of the current framework, the paper discusses the possible future research directions from the aspects of privacy security, incentive mechanism, cross-framework interaction, etc. This paper aims to provide references and ideas for developing and innovating an open-source framework, architecture optimization, security improvement, and algorithm optimization.

Key words: Key words federated learning, open-source framework, model training, machine learning, big data

中图分类号: