《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (3): 830-838.DOI: 10.11772/j.issn.1001-9081.2025040437

• 网络空间安全 • 上一篇    下一篇

基于多层联邦学习的终端数据隐私保护方案

平欢, 夏战国(), 刘思诚, 刘奇翰, 李春磊   

  1. 中国矿业大学 计算机科学与技术学院,江苏 徐州 221116
  • 收稿日期:2025-04-25 修回日期:2025-06-26 接受日期:2025-06-30 发布日期:2025-07-03 出版日期:2026-03-10
  • 通讯作者: 夏战国
  • 作者简介:平欢(1999—),女,河北沧州人,硕士研究生,主要研究方向:机器学习、联邦学习、隐私保护
    刘思诚(2001—),男,江苏徐州人,硕士研究生,主要研究方向:高光谱图像、激光雷达、机器学习、数据挖掘、多模态分类
    刘奇翰(2001—),男,江苏徐州人,硕士研究生,主要研究方向:深度学习、计算机视觉、行人轨迹预测
    李春磊(2001—),男,山东济南人,硕士研究生,主要研究方向:机器学习、深度学习、时间序列预测。
  • 基金资助:
    国家自然科学基金资助项目(62272461)

Terminal data privacy-preserving scheme based on hierarchical federated learning

Huan PING, Zhanguo XIA(), Sicheng LIU, Qihan LIU, Chunlei LI   

  1. School of Computer Science and Technology,China University of Mining and Technology,Xuzhou Jiangsu 221116,China
  • Received:2025-04-25 Revised:2025-06-26 Accepted:2025-06-30 Online:2025-07-03 Published:2026-03-10
  • Contact: Zhanguo XIA
  • About author:PING Huan, born in 1999, M. S. candidate. Her research interests include machine learning, federated learning, privacy protection.
    LIU Sicheng, born in 2001, M. S. candidate. His research interests include hyperspectral image, LiDAR, machine learning, data mining, multimodal classification.
    LIU Qihan, born in 2001, M. S. candidate. His research interests include deep learning, computer vision, pedestrian trajectory prediction.
    LI Chunlei, born in 2001, M. S. candidate. His research interests include machine learning, deep learning, time series prediction.
  • Supported by:
    National Natural Science Foundation of China(62272461)

摘要:

作为一种保护数据隐私的分布式机器学习框架,联邦学习面临着传统云-边-端三层结构难以满足日益严格的数据安全法规和指数级增长数据的挑战,同时边缘设备轻量化趋势导致计算能力受限。为解决上述问题,提出一种面向多级监管和异构资源环境的可选多层联邦学习(OHFL)方案。首先,采用树形多层结构实现边缘层数与每层服务器数量的灵活配置,从而适配不同应用场景下的监管层级和资源分布;其次,在保证通信效率的基础上,引入差分隐私机制,同时对部分神经网络层采用同态加密,并将加密任务下沉至最邻近的边缘服务器以分散计算压力。实验结果表明,OHFL方案在MNIST独立同分布(IID)数据集上达到了97.82%的分类准确率,且新增加密机制分别使用卷积神经网络(CNN)和ResNet18带来的时间损失仅为每轮6.24 s和6.80 s,说明OHFL方案显著提升了系统的计算与通信效率。可见,所提方案不仅在理论上提升了多层级联邦学习架构下的安全性与适应性,而且为实际应用中满足多级监管和高效安全计算需求提供了可行路径,特别适用于数据敏感且资源异构的物联网(IoT)与医疗等场景。

关键词: 联邦学习, 分层架构, 差分隐私, 同态加密, 隐私保护

Abstract:

Federated learning, as a distributed machine learning framework for data privacy protection, faces several challenges such as that the traditional cloud-edge-device structure cannot meet the increasingly strict data security regulations and the exponential growth of data. At the same time, the lightweight trend of edge devices leads to limited computing capabilities. To address these issues, an Optional Hierarchical Federated Learning (OHFL) scheme facing multi-level supervision and heterogeneous resource environments was proposed. Firstly, a tree-structured multi-layer architecture was adopted to configure the number of edge layers and the number of servers per layer flexibly, so as to adapt to supervision levels and resource distribution in different scenarios. Secondly, a differential privacy mechanism was introduced on the basis of ensuring communication efficiency, homomorphic encryption was applied to certain neural network layers, and the encryption tasks were offloaded to the nearest edge servers to distribute the computational burden. Experimental results show that the OHFL scheme achieves a classification accuracy of 97.82% on the MNIST Independent Identically Distributed (IID) dataset, with the additional encryption mechanism using Convolutional Neural Network (CNN) and ResNet18 incurring only a time overhead of 6.24 s and 6.80 s per round, verifying that the OHFL scheme improves the system efficiency of computation and communication significantly. It can be seen that the proposed scheme improves the security and adaptability of hierarchical federated learning architecture theoretically, and provides a feasible solution for meeting multi-level supervision and efficient secure computation in practical applications, especially suitable for scenarios such as Internet of Things (IoT) and healthcare with data sensitivity and resource heterogeneity.

Key words: federated learning, hierarchical architecture, differential privacy, homomorphic encryption, privacy protection

中图分类号: