Differentially Private Federated Learning With an Adaptive Noise Mechanism

被引:1
|
作者
Xue, Rui [1 ]
Xue, Kaiping [1 ,2 ,3 ]
Zhu, Bin [1 ]
Luo, Xinyi [1 ]
Zhang, Tianwei [4 ]
Sun, Qibin [1 ]
Lu, Jun [1 ,2 ,3 ]
机构
[1] Univ Sci & Technol China, Sch Cyber Sci & Technol, Hefei 230027, Anhui, Peoples R China
[2] Key Lab Med Elect & Digital Hlth Zhejiang Prov, Jiaxing 314001, Peoples R China
[3] Engn Res Ctr Intelligent Human Hlth Situat Awarene, Jiaxing 314001, Zhejiang, Peoples R China
[4] Nanyang Technol Univ, Sch Comp Sci & Engn, Singapore 639798, Singapore
基金
中国国家自然科学基金;
关键词
Federated learning; differential privacy; adaptive noise;
D O I
10.1109/TIFS.2023.3318944
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Federated Learning (FL) enables multiple distributed clients to collaboratively train a model with owned datasets. To avoid the potential privacy threat in FL, researchers propose the DP-FL strategy, which utilizes differential privacy (DP) to add elaborate noise to the exchanged parameters to hide privacy information. DP-FL guarantees the privacy of FL at the cost of model performance degradation. To balance the trade-off between model accuracy and security, we propose a differentially private federated learning scheme with an adaptive noise mechanism. This is challenging, as the distributed nature of FL makes it difficult to appropriately estimate sensitivity, where sensitivity is a concept in DP that determines the scale of noise. To resolve this, we design a generic method for sensitivity estimates based on local and global historical information. We also provide instances on four commonly used optimizers to verify its effectiveness. The experiments on MNIST, FMNIST and CIFAR-10 convincingly prove that our proposed scheme achieves higher accuracy while keeping high-level privacy protection compared to prior works.
引用
收藏
页码:74 / 87
页数:14
相关论文
共 50 条
  • [1] Adap DP-FL: Differentially Private Federated Learning with Adaptive Noise
    Fu, Jie
    Chen, Zhili
    Han, Xiao
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON TRUST, SECURITY AND PRIVACY IN COMPUTING AND COMMUNICATIONS, TRUSTCOM, 2022, : 656 - 663
  • [2] Differentially Private Federated Learning with Functional Mechanism
    Cao, Shi-Xiang
    Chen, Chao-Meng
    Tang, Peng
    Su, Sen
    [J]. Jisuanji Xuebao/Chinese Journal of Computers, 2023, 46 (10): : 2178 - 2195
  • [3] The Skellam Mechanism for Differentially Private Federated Learning
    Agarwal, Naman
    Kairouz, Peter
    Liu, Ziyu
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021,
  • [4] Differentially private federated learning framework with adaptive clipping
    Wang, Fangwei
    Xie, Meiyun
    Li, Qingru
    Wang, Changguang
    [J]. Xi'an Dianzi Keji Daxue Xuebao/Journal of Xidian University, 2023, 50 (04): : 111 - 112
  • [5] AdaDpFed: A Differentially Private Federated Learning Algorithm With Adaptive Noise on Non-IID Data
    Zhao, Zirun
    Sun, Yi
    Bashir, Ali Kashif
    Lin, Zhaowen
    [J]. IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, 2024, 70 (01) : 2536 - 2545
  • [6] PADP-FedMeta: A personalized and adaptive differentially private federated meta learning mechanism for AIoT
    Dong, Fang
    Ge, Xinghua
    Li, Qinya
    Zhang, Jinghui
    Shen, Dian
    Liu, Siqi
    Liu, Xiao
    Li, Gang
    Wu, Fan
    Luo, Junzhou
    [J]. JOURNAL OF SYSTEMS ARCHITECTURE, 2023, 134
  • [7] An Optimized Sparse Response Mechanism for Differentially Private Federated Learning
    Ma, Jiating
    Zhou, Yipeng
    Cui, Laizhong
    Guo, Song
    [J]. IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2024, 21 (04) : 2285 - 2295
  • [8] Differentially Private Federated Learning in Edge Networks: The Perspective of Noise Reduction
    Li, Yiwei
    Wang, Shuai
    Chi, Chong-Yung
    Quek, Tony Q. S.
    [J]. IEEE NETWORK, 2022, 36 (05): : 167 - 172
  • [9] CSRA: Robust Incentive Mechanism Design for Differentially Private Federated Learning
    Yang, Yunchao
    Hu, Miao
    Zhou, Yipeng
    Liu, Xuezheng
    Wu, Di
    [J]. IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2024, 19 : 892 - 906
  • [10] Incentive Mechanism for Differentially Private Federated Learning in Industrial Internet of Things
    Xu, Yin
    Xiao, Mingjun
    Tan, Haisheng
    Liu, An
    Gao, Guoju
    Yan, Zhaoyang
    [J]. IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2022, 18 (10) : 6927 - 6939