Robust-PFedproto:基于个性化层的鲁棒联邦原型学习方法

徐明迪; 李正潇; 王梓行; 靳朝阳

doi:10.11959/j.issn.2096-109x.2025032

您当前的位置：

首页 >

文章列表页 >

Robust-PFedproto:基于个性化层的鲁棒联邦原型学习方法

学术论文 | 更新时间：2025-06-30

- Robust-PFedproto:基于个性化层的鲁棒联邦原型学习方法
- Robust-PFedproto: robust federated prototype learning based on personalized layers
- 网络与信息安全学报 2025年11卷第3期页码：67-80
- 作者机构：
  
  1.武汉数字工程研究所，湖北武汉 430074
  2.公安部第三研究所，上海 201204
- 作者简介：
  
  [ "徐明迪（1980- ），男，湖北随州人，武汉数字工程研究所研究员，主要研究方向为网络安全、可信计算。" ]
  [ "李正潇（1999- ），男，湖北武汉人，武汉数字工程研究所硕士生，主要研究方向为网络安全、隐私保护。" ]
  [ "王梓行（1999- ），男，江苏宿迁人，主要研究方向为数据安全、密码应用。" ]
  [ "靳朝阳（1993- ），女，河北邢台人，武汉数字工程研究所工程师，主要研究方向为网络安全、云安全。" ]
- 基金信息：
  
  国防科研基金项目(JCKY2022211C007;JCKY2021206B104)
- DOI：10.11959/j.issn.2096-109x.2025032
  中图分类号： TP393
- 收稿：2024-12-16，
  
  修回：2025-05-05，
  
  纸质出版：2025-06-25
- 稿件说明：
移动端阅览
徐明迪, 李正潇, 王梓行, 等. Robust-PFedproto:基于个性化层的鲁棒联邦原型学习方法[J]. 网络与信息安全学报, 2025, 11(3): 67-80.

Format: XU M D, LI Z X, WANG Z H, et al. Robust-PFedproto: robust federated prototype learning based on personalized layers[J]. Chinese Journal of Network and Information Security, 2025, 11(3): 67-80.
徐明迪, 李正潇, 王梓行, 等. Robust-PFedproto:基于个性化层的鲁棒联邦原型学习方法[J]. 网络与信息安全学报, 2025, 11(3): 67-80. DOI： 10.11959/j.issn.2096-109x.2025032.

Format: XU M D, LI Z X, WANG Z H, et al. Robust-PFedproto: robust federated prototype learning based on personalized layers[J]. Chinese Journal of Network and Information Security, 2025, 11(3): 67-80. DOI： 10.11959/j.issn.2096-109x.2025032.

摘要

联邦学习（federated learning，FL）是一种高效的分布式机器学习框架，其核心优势在于训练数据可保留在网络中的远程客户端。然而，该方法面临两大关键挑战。其一，不同客户端的数据分布通常呈现异构性，这种数据分布的差异性会显著降低整体训练效率；其二，由于中央服务器无法获取远程客户端的真实数据，恶意客户端可能上传恶意模型更新，进而对全局训练的准确性和效率产生负面影响。为解决数据异构性问题，提出了PFedproto框架，该框架在原型学习的基础上引入了个性化层。个性化层位于每个客户端的决策层之前，其参数可通过学习优化，使客户端模型能够更好地适配本地数据集，从而提升模型在本地任务中的预测准确率。在此基础上，进一步构建了3阶段防御方案Robust-PFedproto，旨在增强PFedproto框架的鲁棒性，以有效抵御数据投毒攻击。在4个真实图像分类数据集上开展了实验，结果表明PFedproto框架能够更高效地缓解数据异构性的影响，Robust-PFedproto方案在有效应对数据异构性的同时，对数据投毒攻击展现出良好的鲁棒性。

Abstract

Federated learning (FL)

a distributed machine learning framework

was recognized for retaining training data on remote clients. However

two critical challenges were identified. First

heterogeneous data distributions were commonly observed across clients

which significantly degraded overall training efficiency. Second

the central server could’t access authentic remote client data

allowing adversarial clients to upload malicious model updates

thus negatively affecting global training accuracy and efficiency. To address data heterogeneity

the PFedproto framework was proposed

incorporating personalized layers based on prototype learning. These layers

positioned before each client’s decision layer

were optimized to enhance client model adaptation to local datasets and improve localized task prediction accuracy. Building on this framework

a three-stage defense scheme named RobustPFedproto was developed to strengthen the PFedproto framework’s robustness against data poisoning attacks. Experiments were conducted on four real-world image classification datasets. Results demonstrate that the PFedproto framework effectively mitigates data heterogeneity impacts. Additionally

the Robust-PFedproto scheme not only efficiently handles data heterogeneity but also shows strong robustness against data poisoning attacks.

关键词

Keywords

references

WANG K I , ZHOU X K , LIANG W , et al . Federated transfer learning based cross-domain prediction for smart manufacturing [J ] . IEEE Transactions on Industrial Informatics , 2022 , 18 ( 6 ): 4088 - 4096 .

ZHANG W , WANG Z W , LI X . Blockchain-based decentralized federated transfer learning methodology for collaborative machinery fault diagnosis [J ] . Reliability Engineering & System Safety , 2023 , 229 : 108885 .

CHEN J B , LI J P , HUANG R Y , et al . Federated transfer learning for bearing fault diagnosis with discrepancy-based weighted federated averaging [J ] . IEEE Transactions on Instrumentation and Measurement , 2022 , 71 : 3514911 .

YUN L , ZHEN Y , FANDONG M , et al . An empirical study of catastrophic forgetting in large language models during continual fine-tuning [J ] . CoRR , 2023 , abs/2308.08747.

GUO Z J , LI X D , HAN L , et al . Robust inference for federated meta-learning [J ] . Journal of the American Statistical Association , 2025 : 1 - 27 .

REN H Y , ANICIC D , RUNKLER T A . TinyReptile: TinyML with federated meta-learning [C ] // Proceedings of the 2023 International Joint Conference on Neural Networks (IJCNN) . Piscataway : IEEE Press , 2023 : 1 - 9 .

DONG F , GE X H , LI Q Y , et al . PADP-FedMeta: a personalized and adaptive differentially private federated meta learning mechanism for AIoT [J ] . Journal of Systems Architecture , 2023 , 134 : 102754 .

FALLAH A , MOKHTARI A , OZDAGLAR A . Personalized federated learning: a meta-learning approach [J ] . CoRR , 2020 .

WU C H , WU F Z , LYU L J , et al . Communication-efficient federated learning via knowledge distillation [J ] . Nature Communications , 2022 , 13 : 2032 .

YUHANG M , ZHONGLE X , JUE W , et al . Continual federated learning based on knowledge distillation [C ] // Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence . 2022 : 2182 - 2188 .

VENKATESWARAN P , ISAHAGIAN V , MUTHUSAMY V , et al . FedGen: generalizable federated learning for sequential data [C ] // Proceedings of the 2023 IEEE 16th International Conference on Cloud Computing (CLOUD) . Piscataway : IEEE Press , 2023 : 308 - 318 .

ARIVAZHAGAN M G , AGGARWAL V , SINGH A K , et al . Federated learning with personalization layers [J ] . arXiv preprint arXiv: 1912.00818 , 2019 .

PILLUTLA K , MALIK K , MOHAMED A R , et al . Federated learning with partial model personalization [C ] // Proceedings of International Conference on Machine Learning(ICML 2022) . New York : PMLR , 2022 : 17716 - 17758 .

BAEK J , JEONG W , JIN J , et al . Personalized subgraph federated learning [J ] . ICML , 2023 .

TAN Y , LONG G D , LIU L , et al . FedProto: federated prototype learning across heterogeneous clients [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2022 , 36 ( 8 ): 8432 - 8440 .

FENG H , YI H W , LI X H , et al . Review of privacy-preserving research in recommendation systems [J ] . Journal of Frontiers of Computer Science and Technology , 2023 , 17 ( 8 ): 1814 - 1832 .

KIM H , KWAK Y , JUNG M , et al . ProtoFL: unsupervised federated learning via prototypical distillation [C ] // Proceedings of the 2023 IEEE/CVF International Conference on Computer Vision (ICCV) . Piscataway : IEEE Press , 2023 : 6447 - 6456 .

MICHIELI U , OZAY M . Prototype guided federated learning of visual feature representations [J ] . arXiv preprint arXiv: 2105.08982 , 2021 .

YANG Y C , HUI B , YUAN H L , et al . PrivateFL: accurate, differentially private federated learning via personalized data transformation [C ] // Proceedings of the 32nd USENIX Security Symposium . Berkeley : USENIX Association , 2023 : 1595 - 1611 .

ZHANG J L , CHEN B , CHENG X , et al . PoisonGAN: generative poisoning attacks against federated learning in edge computing systems [J ] . IEEE Internet of Things Journal , 2021 , 8 ( 5 ): 3310 - 3322 .

CAO X Y , GONG N Z . MPAF: model poisoning attacks to federated learning based on fake clients [C ] // Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) . Piscataway : IEEE Press , 2022 : 3395 - 3403 .

PEVA B , EL M E M , RACHID G , et al . Machine learning with adversaries: Byzantine tolerant gradient descent [J ] . Advances In neural Information Processing Systems , 2017 , 30 : 118 - 128 .

XIE C , KOYEJO O , GUPTA I . Generalized byzantine-tolerant SGD [J ] . arXiv preprint arXiv: 1802.10116 , 2018 .

CHEN Y D , SU L L , XU J M . Distributed statistical machine learning in adversarial settings: Byzantine gradient descent [C ] // Proceedings of the Abstracts of the 2018 ACM International Conference on Measurement and Modeling of Computer Systems . New York : ACM , 2018 : 96 - 96 .

YIN D , CHEN Y D , RAMCHANDRAN K , et al . Byzantine-robust distributed learning: towards optimal statistical rates [J ] . arXiv preprint arXiv: 1803.01498 , 2018 .

LI L P , XU W , CHEN T Y , et al . RSA: Byzantine-robust stochastic aggregation methods for distributed learning from heterogeneous datasets [C ] // Proceedings of the Thirty-First Innovative Applications of Artificial Intelligence Conference . 2019 : 1544 - 1551 .

ELKORDY A R , AVESTIMEHR A S . HeteroSAg: secure aggregation with heterogeneous quantization in federated learning [J ] . IEEE Transactions on Communications , 2022 , 70 ( 4 ): 2372 - 2386 .

WU Z X , LING Q , CHEN T Y , et al . Federated variance-reduced stochastic gradient descent with robustness to Byzantine attacks [J ] . IEEE Transactions on Signal Processing , 2020 , 68 : 4583 - 4596 .

ZHAO L C , JIANG J L , FENG B , et al . SEAR: secure and efficient aggregation for Byzantine-robust federated learning [J ] . IEEE Transactions on Dependable and Secure Computing , 2022 , 19 ( 5 ): 3329 - 3342 .

LIU X Y , LI H W , XU G W , et al . Privacy-enhanced federated learning against poisoning adversaries [J ] . IEEE Transactions on Information Forensics and Security , 2021 , 16 : 4574 - 4588 .

YANG Z P , SHI Y M , ZHOU Y , et al . Trustworthy federated learning via blockchain [J ] . IEEE Internet of Things Journal , 2023 , 10 ( 1 ): 92 - 109 .

CUI L Z , SU X X , MING Z X , et al . CREAT: blockchain-assisted compression algorithm of federated learning for content caching in edge computing [J ] . IEEE Internet of Things Journal , 2022 , 9 ( 16 ): 14151 - 14161 .

刘天瑞 , 王连海 , 王启正 , 等 . 算力网络环境下基于区块链的自适应联邦学习 [J ] . 网络与信息安全学报 , 2024 , 10 ( 3 ): 130 - 142 .

LIU T R , WANG L H , WANG Q Z , et al . Blockchain based adaptive federated learning in computational power network [J ] . Chinese Journal of Network and Information Security , 2024 , 10 ( 3 ): 130 - 142 .

LI Y R , ZHANG S B , CHANG Y , et al . Privacy-preserving and poisoning-defending federated learning in fog computing [J ] . IEEE Internet of Things Journal , 2024 , 11 ( 3 ): 5063 - 5077 .

XU H W , FENG Y , CHEN J , et al . Unsupervised anomaly detection via variational auto-encoder for seasonal KPIs in Web applications [C ] // Proceedings of the 2018 World Wide Web Conference . New York : ACM , 2018 : 187 - 196 .

KIEU T , YANG B , GUO C J , et al . Outlier detection for time series with recurrent autoencoder ensembles [C ] // Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence . California : IJCAI , 2019 : 2725 - 2732 .

MA Z R , MA J F , MIAO Y B , et al . ShieldFL: mitigating model poisoning attacks in privacy-preserving federated learning [J ] . IEEE Transactions on Information Forensics and Security , 2022 , 17 : 1639 - 1654 .

FUNG C , YOON C J M , BESCHASTNIKH I . The limitations of federated learning in Sybil settings [C ] // Proceedings of the 23rd International Symposium on Research in Attacks, Intrusions and Defenses (RAID 2020) . Berkeley : USENIX Association , 2020 : 301 - 316 .

MODAK S . A new measure for assessment of clustering based on kernel density estimation [J ] . Communications in Statistics - Theory and Methods , 2023 , 52 ( 17 ): 5942 - 5951 .

STEINWARR I , SRIPERUMBUDUR B K , THOMANN P . Adaptive clustering using kernel density estimators [J ] . Journal of Machine Learning Research , 2023 , 24 ( 275 ): 1 - 56 .

LI X Y , QU Z , ZHAO S Q , et al . LoMar: a local defense against poisoning attack on federated learning [J ] . IEEE Transactions on Dependable and Secure Computing , 2023 , 20 ( 1 ): 437 - 450 .

ALORAN I , SAMET S . Defending federated learning against model poisoning attacks [C ] // Proceedings of the 2023 IEEE International Conference on Big Data (BigData) . Piscataway : IEEE Press , 2023 : 3584 - 3589 .

FENG X , CHENG W H , CAO C J , et al . DPFLA: defending private federated learning against poisoning attacks [J ] . IEEE Transactions on Services Computing , 2024 , 17 ( 4 ): 1480 - 1491 .

JEBREEL N M , DOMINGO-FERRER J . FL-Defender: combating targeted attacks in federated learning [J ] . Knowledge-Based Systems , 2023 , 260 : 110178 .

YUAN H L , PAUL W , AUCOTT J , et al . PFedEdit: personalized federated learning via automated model editing [C ] // Computer Vision - ECCV 2024 . Berlin : Springer , 2025 : 91 - 107 .

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

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

FLWD:基于联邦学习的Webshell检测方法