基于自适应差分隐私与客户选择优化的联邦学习方法  

Federated learning method based on adaptive differential privacy and client selection optimization

作　　者：徐超 张淑芬 陈海田 彭璐璐 张帅华 XU Chao;ZHANG Shufen;CHEN Haitian;PENG Lulu;ZHANG Shuaihua(College of Science,North China University of Science and Technology,Tangshan Hebei 063210,China;Hebei Key Laboratory of Data Science and Application(North China University of Science and Technology),Tangshan Hebei 063210,China;Tangshan Key Laboratory of Big Data Security and Intelligent Computing,Beijing Jiaotong University,Tangshan Hebei 063210,China;Tangshan Key Laboratory of Data Science(North China University of Science and Technology),Tangshan Hebei 063210,China)

机构地区：[1]华北理工大学理学院,河北唐山063210 [2]河北省数据科学与应用重点实验室(华北理工大学),河北唐山063210 [3]北京交通大学唐山市大数据安全与智能计算重点实验室,河北唐山063210 [4]唐山市数据科学重点实验室(华北理工大学),河北唐山063210

出　　处：《计算机应用》2025年第2期482-489,共8页journal of Computer Applications

基　　金：国家自然科学基金资助项目(U20A20179)。

摘　　要：将差分隐私应用于联邦学习的方法是保护训练数据隐私的关键技术之一。针对之前多数工作未考虑参数的异质性,对训练参数均匀裁剪使每轮加入的噪声都是均匀的,从而影响模型收敛和训练参数质量的问题,提出一种基于梯度裁剪的自适应噪声添加方案。考虑梯度的异质性,在不同轮次为不同客户端执行自适应的梯度裁剪,从而使噪声大小自适应调整;同时,为进一步提升模型性能,对比传统的客户端随机采样方式,提出一种结合轮盘赌与精英保留的客户端采样方法。结合上述2种方法,提出一种结合客户端选择的自适应差分隐私联邦学习(CS&AGC DP_FL)方法。实验结果表明,在隐私预算为0.5时,相较于自适应差分隐私的联邦学习方法(Adapt DP_FL),所提方法能在相同级别的隐私约束下使最终的模型分类准确率提升4.9个百分点,并且在收敛速度方面,所提方法相较于对比方法进入收敛状态所需的轮次减少了4~10轮。The method of applying differential privacy to federated learning has been one of the key techniques for protecting the privacy of training data.Addressing the issue that most previous works do not consider the heterogeneity of parameters,resulting in pruning training parameters uniformly,leading to uniform noise addition in each round,thus affecting model convergence and the quality of training parameters,an adaptive noise addition scheme based on gradient clipping was proposed.Considering the heterogeneity of gradients,adaptive gradient clipping was executed for different clients in different rounds,thereby allowing for the adaptive adjustment of noise magnitude.At the same time,to further improve model performance,different from traditional client random sampling methods,a client sampling method that combines roulette and elite preservation was proposed.Combining the aforementioned two methods,a Client Selection and Adaptive Gradient Clipping Differential Privacy_Federated Learning(CS&AGC DP_FL) was proposed.Experimental results demonstrate that,when the privacy budget is 0.5,compared to the Federated Learning method based on Adaptive Differential Privacy(Adapt DP_FL),the proposed method improves the final model's classification accuracy by 4.9percentage points under the same level of privacy constraints.Additionally,in terms of convergence speed,the proposed method requires 4 to 10 fewer rounds to reach convergence compared to the methods to be compared.

关 键 词：联邦学习 差分隐私 自适应噪声 轮盘赌 精英保留 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]