面向异构数据的具有隐私保护性能的公平联邦学习

吴炬; 苗银宾; 张炜斌; 刘淑英; 马建峰

doi:10.11959/j.issn.2096-109x.AQ26019

您当前的位置：

首页 >

文章列表页 >

面向异构数据的具有隐私保护性能的公平联邦学习

学术论文 | 更新时间：2026-07-15

- 面向异构数据的具有隐私保护性能的公平联邦学习
- Fair federated learning with privacy protection for heterogeneous data
- 网络与信息安全学报 2026年12卷第3期页码：78-92
- 作者机构：
  
  1.西安电子科技大学网络与信息安全学院，陕西西安 710126
  2.西北大学计算机学院，陕西西安 710127
  3.咸阳师范学院计算机学院，陕西咸阳 712000
- 作者简介：
  
  吴炬（2002- ），男，西安电子科技大学网络与信息安全学院硕士生，主要研究方向为联邦学习、隐私保护。
  苗银宾（1988− ），男，西安电子科技大学网络与信息安全学院教授，主要研究方向为隐私计算、人工智能安全。
  张炜斌（2000− ），男，西安电子科技大学网络与信息安全学院博士生，主要研究方向为可信人工智能、无线感知安全。
  刘淑英（1982− ），女，咸阳师范学院计算机学院教授，主要研究方向为机器学习、网络数据安全。
  马建峰（1963− ），男，西安电子科技大学网络与信息安全学院教授，主要研究方向为应用密码学、无线网络安全、数据安全、移动智能系统安全。
- 基金信息：
  
  大学生创新创业训练计划(S202510722063)
- DOI：10.11959/j.issn.2096-109x.AQ26019
  中图分类号： TP309
- 收稿：2026-01-21，
  
  修回：2026-04-07，
  
  纸质出版：2026-06-25
- 稿件说明：
移动端阅览
吴炬, 苗银宾, 张炜斌, 等. 面向异构数据的具有隐私保护性能的公平联邦学习[J]. 网络与信息安全学报, 2026, 12(3): 78-92.

Wu Ju, Miao Yinbin, Zhang Weibin, et al. Fair federated learning with privacy protection for heterogeneous data[J]. Chinese Journal of Network and Information Security, 2026, 12(3): 78-92.
吴炬, 苗银宾, 张炜斌, 等. 面向异构数据的具有隐私保护性能的公平联邦学习[J]. 网络与信息安全学报, 2026, 12(3): 78-92. DOI： 10.11959/j.issn.2096-109x.AQ26019.

Wu Ju, Miao Yinbin, Zhang Weibin, et al. Fair federated learning with privacy protection for heterogeneous data[J]. Chinese Journal of Network and Information Security, 2026, 12(3): 78-92. DOI： 10.11959/j.issn.2096-109x.AQ26019.

摘要

联邦学习在非独立同分布数据场景下易因梯度冲突导致公平性失衡，且模型参数传输存在隐私泄露风险。现有公平性联邦学习算法虽有效缓解了梯度冲突，但复杂异质数据下局部与全局模型的一致性缺乏针对性优化，导致局部训练偏离全局最优，进而削弱公平性并限制整体模型性能。针对此问题，提出FedALF及LDP-FedALF算法。使用公平驱动目标和分层优化缓解梯度冲突；引入交替方向乘子法进行优化，在客户端局部目标函数中设计对偶变量项与二次惩罚项，构建局部与全局模型一致性约束，进一步提升公平性与模型性能；同时在模型传输中加入自适应本地差分隐私技术保护数据隐私。通过大量实验验证得出，FedALF相较于现有公平联邦学习算法，测试准确率平均提升10.82%，公平性平均提升19.80%；LDP-FedALF相较于现有公平联邦学习算法，测试准确率平均提升7.53%，公平性平均提升15.36%。

Abstract

Federated learning suffers from fairness imbalance due to gradient conflicts in non-independent and indentically distributed (Non-IID) data scenarios

with model parameter transmission risking privacy leakage. Existing fair federated learning algorithms mitigate gradient conflicts effectively but lack targeted optimization for local-global model consistency under complex heterogeneous data

leading to deviated local training from global optima

impaired fairness and limited overall performance. To address this

the FedALF algorithm and LDP-FedALF algorithm were proposed: it alleviates gradient conflicts via fairness-driven objectives and hierarchical optimization; adopts ADMM for optimization

designing dual variables and quadratic penalty terms in clients' local objective functions to enforce local-global model consistency constraints

further enhancing fairness and performance; and integrates adaptive local differential privacy into model transmission for privacy protection. Through a large number of experiments

FedALF achieved an average improvement of 10.82% in test accuracy and 19.80% in fairness performance compared to existing fair federated learning algorithms. LDP-FedALF achieved an average improvement of 7.53% in test accuracy and 15.36% in fairness performance compared to existing fair federated learning algorithms.

关键词

Keywords

references

McMahan B , Moore E , Ramage D , et al . Communication-efficient learning of deep networks from decentralized data [C ] // Artificial Intelligence and Statistics . Fort Lauderdale : PMLR , 2017 : 1273 - 1282 .

周传鑫 , 孙奕 , 汪德刚 , 等 . 联邦学习研究综述 [J ] . 网络与信息安全学报 , 2021 , 7 ( 5 ): 77 - 92 .

Zhou C X , Sun Y , Wang D G , et al . Survey of federated learning research [J ] . Chinese Journal of Network and Information Security , 2021 , 7 ( 5 ): 77 - 92 .

Hu B T , Lu Z D , Li H , et al . Convolutional neural network architectures for matching natural language sentences [C ] // Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 2 . New York : ACM , 2014 : 2042 - 2050 .

Wan F , Wei P X , Han Z J , et al . Min-entropy latent model for weakly supervised object detection [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2019 , 41 ( 10 ): 2395 - 2409 .

Lu J W , Liong V E , Zhou X Z , et al . Learning compact binary face descriptor for face recognition [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2015 , 37 ( 10 ): 2041 - 2056 .

Li T , Sahu A K , Zaheer M , et al . Federated optimization in heterogeneous networks [C ] // Proceedings of Machine Learning and Systems 2020 . Austin : mlsys.org , 2020 : 429 - 450 .

Shi Y X , Yu H , Leung C . Towards fairness-aware federated learning [J ] . IEEE Transactions on Neural Networks and Learning Systems , 2024 , 35 ( 9 ): 11922 - 11938 .

Li T , Hu S , Beirami A , et al . Ditto: Fair and robust federated learning through personalization [C ] // International conference on machine learning . PMLR , 2021 : 6357 - 6368 .

Mohri M , Sivek G , Suresh A T . Agnostic federated learning [C ] // International Conference on Machine Learning , ICML 2019 . Long Beach : PMLR , 2019: 4615 - 4625 .

Wang Z , Fan X L , Qi J Z , et al . Federated learning with fair averaging [PP ] . arXiv preprint , 2021 , arXiv: 2104.14937 .

Hu Z O , Shaloudegi K , Zhang G J , et al . Federated learning meets multi-objective optimization [J ] . IEEE Transactions on Network Science and Engineering , 2022 , 9 ( 4 ): 2039 - 2051 .

Pan Z B , Wang S Y , Li C , et al . FedMDFG: federated learning with multi-gradient descent and fair guidance [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2023 , 37 ( 8 ): 9364 - 9371 .

Yu F , Wang D Q , Shelhamer E , et al . Deep layer aggregation [C ] // Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE Press , 2018 : 2403 - 2412 .

Ma X S , Zhang J , Guo S , et al . Layer-wised model aggregation for personalized federated learning [C ] // Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE Press , 2022 : 10082 - 10091 .

Lee S , Zhang T , Avestimehr S . Layer-wise adaptive model aggregation for scalable federated learning [C ] // Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence and Thirteenth Symposium on Educational Advances in Artificial Intelligence . New York : ACM , 2023 : 8491 - 8499 .

Pan Z B , Li C , Yu F C , et al . FedLF: layer-wise fair federated learning [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2024 , 38 ( 13 ): 14527 - 14535 .

迟欢欢 , 熊平 , 刘恒竹 , 等 . 激励驱动的联邦学习研究综述: 隐私与安全 [J ] . 网络与信息安全学报 , 2025 , 11 ( 3 ): 19 - 44 .

Chi H H , Xiong P , Liu H Z , et al . Survey on incentive-driven federated learning: privacy and security [J ] . Chinese Journal of Network and Information Security , 2025 , 11 ( 3 ): 19 - 44 .

Zhu L , Liu Z , Han S . Deep leakage from gradients [C ] //Advances in Neural Information Processing Systems: Vol. 32 . Vancouver : Curran Associates , 2019 : 14774 - 14784 .

赵瑞 , 段小文 , 刘新 , 等 . 非独立同分布数据下层次化Sinkhorn距离聚类联邦学习算法 [J ] . 计算机研究与发展 , 2025 , 62 ( 12 ): 3002 - 3012 .

Zhao R , Duan X W , Liu X , et al . Hierarchical sinkhorn distance-based clustering federated learning algorithm for Non-IID data [J ] . Journal of Computer Research and Development , 2025 , 62 ( 12 ): 3002 - 3012 .

李宁 , 徐家磊 , 李菲 , 等 . 自适应聚合与元学习优化的联邦学习方法研究 [J ] . 计算机技术与发展 , 2026 , 36 ( 3 ): 178 - 185 .

Li N , Xu J L , Li F , et al . Federated learning with adaptive aggregation and meta-learning optimization [J ] . Computer Technology and Development , 2026 , 36 ( 3 ): 178 - 185 .

Mao W J , Yu B , Zhang C , et al . FedKT: federated learning with knowledge transfer for Non-IID data [J ] . Pattern Recognition , 2025 , 159 : 111143 .

Sun B L , Song X , Tu Y C , et al . FedAgent: federated learning on non-IID data via reinforcement learning and knowledge distillation [J ] . Expert Systems with Applications , 2025 , 285 : 127973 .

Li T , Sanjabi M , Beirami A , et al . Fair resource allocation in federated learning [PP ] . arXiv preprint , 2020 , arXiv: 1905.10497 .

Ling X P , Fu J , Wang K C , et al . FedFDP: fairness-aware federated learning with differential privacy [PP ] . arXiv preprint , 2026 , arXiv: 2402.16028 .

Sabah F , Chen Y W , Yang Z , et al . FairDPFL-SCS: fair dynamic personalized federated learning with strategic client selection for improved accuracy and fairness [J ] . Information Fusion , 2025 , 115 : 102756 .

Wei E M , Ozdaglar A . Distributed alternating direction method of multipliers [C ] // Proceedings of the 2012 IEEE 51st IEEE Conference on Decision and Control (CDC) . Piscataway : IEEE Press , 2013 : 5445 - 5450 .

Hong M Y , Luo Z Q , Razaviyayn M . Convergence analysis of alternating direction method of multipliers for a family of nonconvex problems [C ] // Proceedings of the 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . Piscataway : IEEE Press , 2015 : 3836 - 3840 .

Abadi M , Chu A , Goodfellow I , et al . Deep learning with differential privacy [C ] // Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security . New York : ACM , 2016 : 308 - 318 .

Miao Y B , Xie R P , Li X H , et al . Compressed federated learning based on adaptive local differential privacy [C ] // Proceedings of the 38th Annual Computer Security Applications Conference . New York : ACM , 2022 : 159 - 170 .

Sun L C , Qian J W , Chen X . LDP-FL: practical private aggregation in federated learning with local differential privacy [PP ] . arXiv preprint , 2021 , arXiv: 2007. 15789 .

Fliege J , Svaiter B F . Steepest descent methods for multicriteria optimization [J ] . Mathematical Methods of Operations Research , 2000 , 51 ( 3 ): 479 - 494 .

Gebken B , Peitz S , Dellnitz M . On the hierarchical structure of Pareto critical sets [J ] . Journal of Global Optimization , 2019 , 73 ( 4 ): 891 - 913 .

Gong Y H , Li Y C , Freris N M . FedADMM: a robust federated deep learning framework with adaptivity to system heterogeneity [C ] // Proceedings of the 2022 IEEE 38th International Conference on Data Engineering (ICDE) . Piscataway : IEEE Press , 2022 : 2575 - 2587 .

Xiao H , Rasul K , Vollgraf R . Fashion-MNIST: a novel image dataset for benchmarking machine learning algorithms [PP ] . arXiv preprint , 2017 , arXiv: 1708.07747 .

Krizhevsky A , Hinton G . Learning multiple layers of features from tiny images [R ] . Toronto : University of Toronto , 2009 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

支持多数不规则用户的隐私保护联邦学习框架

DP-FedDRC：缓解维度坍塌的差分隐私个性化联邦学习

基于动态聚类和信誉评估的抗攻击联邦框架

基于风险权值聚合联邦学习的金融欺诈检测模型

FL-IoV：基于联邦学习的综合车联网入侵检测方法