群体智能中的联邦学习算法综述  

A survey on federated learning in crowd intelligence

作　　者：杨强 童咏昕[3] 王晏晟[3] 范力欣 王薇 陈雷[2] 王魏[4] 康焱[1] YANG Qiang;TONG Yongxin;WANG Yansheng;FAN Lixin;WANG Wei;CHEN Lei;WANG Wei;KANG Yan(Qianhai WeBank Co.,Ltd.,Shenzhen 518063,China;The Hong Kong University of Science and Technology,Hong Kong 999077,China;Beihang University,Beijing 100191,China;Nanjing University,Nanjing 210033,China)

机构地区：[1]深圳前海微众银行股份有限公司,广东深圳518063 [2]香港科技大学,中国香港999077 [3]北京航空航天大学,北京100191 [4]南京大学,江苏南京210033

出　　处：《智能科学与技术学报》2022年第1期29-44,共16页Chinese Journal of Intelligent Science and Technology

基　　金：国家重点研发计划基金资助项目(No.2018AAA0101100);国家自然科学基金资助项目(No.U21A20516,No.61822201,No.U1811463,No.62076017);微众学者计划。

摘　　要：群体智能是在互联网高速普及下诞生的人工智能新范式。然而,数据孤岛与数据隐私保护问题导致群体间数据共享困难,群体智能应用难以构建。联邦学习是一类新兴的打破数据孤岛、联合构建群智模型的重要方法。首先,介绍了联邦学习的基础概念以及其与群体智能的关系;其次,基于群体智能视角对联邦学习算法框架进行了分类,从隐私、精度与效率3个角度讨论了联邦学习算法优化技术;而后,阐述了基于线性模型、树模型与神经网络模型的联邦学习算法模型;最后,介绍了联邦学习代表性开源平台与典型应用,并对联邦学习研究进行总结展望。Crowd intelligence is emerging as a new artificial intelligence paradigm owing to the rapid development of the Internet.However,the data isolation and data privacy preservation problems make it difficult to share data among the crowd and to build crowd intelligent applications.Federated learning is a novel solution that aims to collaboratively build models by breaking the data barriers in crowd.Firstly,the basic ideas of federated learning and a comparison with crowd intelligence were introduced.Secondly,federated learning algorithms were divided into three categories according to the crowd organization,and further optimization techniques on privacy,accuracy and efficiency were discussed.Thirdly,fe-derated learning operators based on linear models,tree models and neural network models were presented respectively.Finally,mainstream federated learningopensource platforms and typical applications were introduced,followed by the conclusion.

关 键 词：群体智能 联邦学习 隐私保护 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]