[1]张海超,赖金山,刘 东,等.边缘计算下的轻量级联邦学习隐私保护方案[J].计算机技术与发展,2023,33(09):161-167.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 024]
　ZHANG Hai-chao,LAI Jin-shan,LIU Dong,et al.Lightweight Federated Learning Privacy Protection Scheme under Edge Computing[J].,2023,33(09):161-167.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 024]

点击复制

边缘计算下的轻量级联邦学习隐私保护方案()

分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年09期

页码:

161-167

栏目:

人工智能

出版日期:

2023-09-10

文章信息/Info

Title:

Lightweight Federated Learning Privacy Protection Scheme under Edge Computing

文章编号:

1673-629X(2023)09-0161-07

作者:

张海超¹ ; 赖金山² ; 刘 东³ ; 张凤荔²

1. 四川公安厅科技信息化总队,四川 成都 610015;

2. 电子科技大学 信息与软件工程学院,四川 成都 610054;

3. 电子科技大学 计算机科学与工程学院,四川 成都 611731

Author(s):

ZHANG Hai-chao1 ; LAI Jin-shan2 ; LIU Dong3 ; ZHANG Feng-li2

1. Science and Technology Informatization Corps of Sichuan Public Security Department,Chengdu 610015,China;

2. School of Information and Software Engineering,University of Electronic Science and Technology of China,Chengdu 610054,China;

3. School of Computer Science and Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China

关键词:

联邦学习; 边缘计算; 同态加密; 差分隐私; 隐私保护

Keywords:

federated learning; edge computing; homomorphic encryption; differential privacy; privacy protection

分类号:

TP301

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 09. 024

摘要:

随着物联网和大数据技术的高速发展,以传统云计算模式为代表的集中式学习效率低下,且易受到单点攻击、共谋攻击、中间人攻击等一系列攻击手段,造成数据安全的

隐患。 边缘计算模式使得分布式联邦学习成为了可能,然而,联邦学习虽然能够保证数据在本地的安全和隐私,但是也面临众多安全威胁,如梯度泄露攻击,此外,效率问题

也是联邦学习的痛点所在。 为了保障边缘计算场景下的模型训练安全,提出了一种边缘计算下的轻量级联邦学习隐私保护方案(Lightweight Federated Learning Privacy Protection Scheme Under Edge Computing,LFLPP) 。 首先,提出一种云-边-端分层的联邦学习框架;其次,对不同层进行隐私保护;最后,提出一种周期性更新策略,极大地

提高了收敛速度。 使用乳腺癌肿瘤数据集和 CIFAR10 数据集在 LR 模型和 Resnet18 残差模型上进行训练和测试,同时使用 CIFAR10 数据集与 FedAvg 和PPFLEC

( Privacy-Preserving Federated Learning for Internet of Medical Things under Edge Computing) 两种方案进行对比实验,得出准确率和训练效率的差距,并进行准确率、效率以及安全性等方面的分析,该方案在 CIFAR-10 数据集上能达到84. 63% 的准确率。

Abstract:

With the rapid development of the Internet of Things and big data technology,centralized learning represented by the traditionalcloud computing model is inefficient?

and vulnerable to a series of attacks such as single point attack,collusion attack,man in the middleattack,resulting in hidden dangers of data security. The edge computing model makes distributed federated learning possible. However,although federated learning can ensure the security and privacy of data locally, it also faces many security threats, such as gradientdisclosure attacks. In addition,the efficiency is also the pain point of federated learning. In order to ensure the security of model trainingin the edge computing scenario,a lightweight federated learning privacy protection scheme under edge computing ( LFLPP) is proposed.Firstly,a cloud - edge - end layered federated learning framework is proposed. Secondly,privacy protection for different layers. Finally,a periodic updating strategy is proposed,which greatly improves the convergence speed. The breast cancer tumor data set and CIFAR10data set were used for training and testing on LR model and Resnet18 residual model. At the same time,CIFAR10 data set was used toconduct comparative experiments with FedAvg and PPFLEC ( Privacy Preserving Federated Learning for Internet of Medical Things underEdge Computing) ,to find out the gap between accuracy and training efficiency,and to conduct accuracy analysis,efficiency analysis andsecurity analysis, This scheme can achieve 84. 63% accuracy on CIFAR-10 dataset.

常用功能