欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2026, Vol. 43 ›› Issue (4): 553-565.DOI: 10.7523/j.ucas.2024.051

• 电子信息与计算机科学 • 上一篇    下一篇

FLShadow:基于影子模型的拜占庭鲁棒联邦聚合方法

徐辰晨1(), 汪旭童2,3, 王涛春1, 陈付龙1, 刘奇旭2,3   

  1. 1.安徽师范大学计算机与信息学院,安徽 芜湖 241000
    2.中国科学院信息工程研究所,北京 100085
    3.中国科学院大学网络空间安全学院,北京 100049
  • 收稿日期:2023-10-08 修回日期:2025-05-17 发布日期:2024-06-11
  • 通讯作者: 徐辰晨
  • 基金资助:
    国家自然科学基金(61972438);安徽师范大学科研项目(2021xjxm037);安徽省重点研究与开发计划项目(2022a05020049);安徽省自然科学基金(2108085MF219)

FLShadow:Byzantine-robust federated aggregation based on a trusted shadow model

Chenchen XU1(), Xutong WANG2,3, Taochun WANG1, Fulong CHEN1, Qixu LIU2,3   

  1. 1.School of Computer Information,Anhui Normal University,Wuhu 241000,Anhui,China
    2.Institute of Information Engineering,Chinese Academy of Sciences,Beijing 100085,China
    3.School of Cyber Security,University of Chinese Academy of Sciences,Beijing 100049,China
  • Received:2023-10-08 Revised:2025-05-17 Published:2024-06-11
  • Contact: Chenchen XU

摘要:

在联邦学习中,拜占庭节点可能精心操纵客户端模型更新,导致聚合后中心模型准确率下降、无法收敛或通信轮数增加。由于中心模型无法仅根据客户端更新进行鲁棒性聚合,因此需要可信任的参考梯度。为此,提出基于影子模型拜占庭鲁棒的联邦聚合方法。首先,中心服务器预先收集小规模可信影子数据集并训练得到影子模型,对比客户端模型与可信影子模型更新方向,累计恶意得分,从而标记恶意客户端;然后,中心服务器剔除恶意客户端更新并修剪得到可信梯度;最后,进行聚合,保证模型性能和效率。在多种模型架构和真实数据集中进行测试。结果表明,本方法在3个数据集中能够有效防御6种拜占庭节点攻击。

关键词: 联邦学习, 拜占庭攻击, 影子模型, 影子数据集, 聚合规则, 隐私安全, 对抗攻击

Abstract:

In federated learning, Byzantine nodes can carefully manipulate the model updates of clients. As a result, the central model accuracy decreases or fails to converge after aggregation. The number of communication rounds also increases. Without trusted reference gradients, the central model can not be properly aggregated solely from updates provided by untrustworthy clients. In this paper, we introduce a trusted reference and we regard it as a shadow model. To overcome this challenge, we propose a novel Byzantine robust federated aggregation method. The central server collects the shadow dataset in advance and trains a model called the shadow model. The central server compares the update direction between the client model and the trusted shadow model. Accordingly, the central server computes malicious score and marks the malicious clients. Then, the central model deletes or prunes the updates from the malicious clients. Finally, the central model aggregates the corrected gradients to ensure good convergence and maintain accuracy. The proposed method has been evaluated on a variety of model architectures and real datasets. The results show that the proposed method can effectively defend against six different Byzantine node attacks on three datasets.

Key words: federated learning, Byzantine attack, shadow model, shadow dataset, aggregation rule, privacy security, adversarial attack

中图分类号: