《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (4): 1160-1168.DOI: 10.11772/j.issn.1001-9081.2022030337

• 网络空间安全 • 上一篇    

基于个性化差分隐私的联邦学习算法

尹春勇(), 屈锐   

  1. 南京信息工程大学 计算机学院、软件学院、网络空间安全学院,南京 210044
  • 收稿日期:2022-03-21 修回日期:2022-05-13 接受日期:2022-05-25 发布日期:2023-04-11 出版日期:2023-04-10
  • 通讯作者: 尹春勇
  • 作者简介:屈锐(1999—),男,江苏宿迁人,硕士研究生,主要研究方向:差分隐私、联邦学习。

Federated learning algorithm based on personalized differential privacy

Chunyong YIN(), Rui QU   

  1. School of Computer Science,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China
  • Received:2022-03-21 Revised:2022-05-13 Accepted:2022-05-25 Online:2023-04-11 Published:2023-04-10
  • Contact: Chunyong YIN
  • About author:QU Rui, born in 1999, M. S. candidate. His research interests include differential privacy, federated learning.

摘要:

联邦学习(FL)可以有效保护用户的个人数据不被攻击者获得,而差分隐私(DP)则可以实现FL的隐私增强,解决模型训练参数导致的隐私泄露问题。然而,现有的基于DP的FL方法只关注统一的隐私保护预算,而忽略了用户的个性化隐私需求。针对此问题,提出了一种两阶段的基于个性化差分隐私的联邦学习(PDP-FL)算法。在第一阶段,依据用户的隐私偏好对用户隐私进行分级,并添加满足用户隐私偏好的噪声,以实现个性化隐私保护,同时上传隐私偏好对应的隐私等级给中央聚合服务器;在第二阶段,为实现对全局数据的充分保护,采取本地和中心同时保护的策略,并根据用户上传的隐私等级,添加符合全局DP阈值的噪声,以量化全局的隐私保护水平。实验结果表明,在MNIST和CIFAR-10数据集上,PDP-FL算法的分类准确度分别为93.8%~94.5%和43.4%~45.2%,优于基于本地化差分隐私的联邦学习(LDP-Fed)和基于全局差分隐私的联邦学习(GDP-FL),同时满足了个性化隐私保护的需求。

关键词: 联邦学习, 差分隐私, 隐私偏好, 隐私分级, 个性化隐私保护

Abstract:

Federated Learning (FL) can effectively protect users' personal data from attackers. Differential Privacy (DP) is applied to enhance the privacy of FL, which can solve the problem of privacy disclose caused by parameters in the model training. However, existing FL methods based on DP on concentrate on the unified privacy protection budget and ignore the personalized privacy requirements of users. To solve this problem, a two-stage Federated Learning with Personalized Differential Privacy (PDP-FL) algorithm was proposed. In the first stage, the user's privacy was graded according to the user's privacy preference, and the noise meeting the user's privacy preference was added to achieve the purpose of personalized privacy protection. At the same time, the privacy level corresponding to the privacy preference was uploaded to the central aggregation server. In the second stage, in order to fully protect the global data, the simultaneous local and central protection strategy was adopted. And according to the privacy level uploaded by the user, the noise conforming to the global DP threshold was added to quantify the global privacy protection level. Experimental results show that on MNIST and CIFAR-10 datasets, the classification accuracy of PDP-FL algorithm reaches 93.8% to 94.5% and 43.4% to 45.2% respectively, which is better than those of Federated learning with Local Differential Privacy (LDP-Fed) algorithm and Federated Learning with Global Differential Privacy (GDP-FL) algorithm, PDP-FL algorithm meets the needs of personalized privacy protection.

Key words: Federated Learning (FL), Differential Privacy (DP), privacy preference, privacy rating, personalized privacy protection

中图分类号: