计算机与现代化 ›› 2025, Vol. 0 ›› Issue (08): 63-69.doi: 10.3969/j.issn.1006-2475.2025.08.009

• 信息安全 • 上一篇    下一篇

基于联邦学习的数据隐私保护方案

  


  1. (临沂大学信息科学与工程学院,山东 临沂 276000)
  • 出版日期:2025-08-27 发布日期:2025-08-28
  • 作者简介: 作者简介:程钰雯(2000 —),女,山东临沂人,硕士研究生,研究方向:隐私保护,E-mail: cywlyu@163.com; 景义君(2000 —),女,山东济南人,硕士研究生,研究方向:数据安全,E-mail: 13864092331@163.com; 时自成(1999—),男,山东枣庄人,硕士研究生,研究方向:物联网安全,E-mail: shi5108992@163.com; 荆长强(1985—),男,山东临沂人,副教授,博士,研究方向:物联网安全,E-mail: jingchangqiang@lyu.edu.cn; 郭锋(1979—),男,山东临沂人,副教授,博士,研究方向:工业物联网安全,E-mail: gflytu@163.com; 武传坤(1964—),男,山东临沂人,教授,博士,研究方向:密码学,物联网安全,E-mail: chuankunwu@lyu.edu.cn。
  • 基金资助:
    基金项目:国家自然科学基金青年基金资助项目(61901206)
      

A Data Privacy Protection Scheme Based on Federated Learning


  1. (School of Information Science and Engineering, Linyi University, Linyi 276000, China)
  • Online:2025-08-27 Published:2025-08-28

摘要: 摘要:当前医疗数据领域面临着医疗数据孤岛问题,制约了数据在不同机构之间的流通和分享,不利于患者的跨机构诊疗。为了解决这一问题,本文提出一种基于联邦学习的数据隐私保护方案(Federated Learning with Schnorr Zero-knowledge Based Identity Authentication and Differential Privacy Protection, FL-SZIDP)。首先,建立一种基于联邦学习的数据共享框架。其次,为了防御敌手通过反向攻击窃取原始数据信息,对每个参与者上传的模型参数添加差分隐私噪声进行扰动;为了防止联邦学习中恶意参与方的加入,对参与方完成基于Schnorr零知识证明的身份验证,确保训练参与方的身份可信性。最后,通过MNIST数据集验证了本文算法的有效性。实验结果表明,FL-SZIDP方案在保护隐私的同时确保了准确性。


关键词: 关键词:联邦学习; 隐私保护; 差分隐私; 数据安全
 ,
  

Abstract:
Abstract: The current healthcare data domain faces the issue of data silos, which restricts the flow and sharing of data among different institutions and hinders cross-institutional treatment for patients. To address this problem, this paper proposes a privacy protection scheme based on federated learning (Federated Learning with Schnorr Zero-knowledge Based Identity Authentication and Differential Privacy Protection, FL-SZIDP). Firstly, a data-sharing framework based on federated learning is established. Secondly, to defend against adversaries attempting to steal original data through reverse attacks, differential privacy noise is added to the model parameters uploaded by each participant. To prevent malicious participants from joining the federated learning process, identity authentication based on Schnorr zero-knowledge proof is performed, ensuring the credibility of the participants’ identities. Finally, the effectiveness of the proposed algorithm is verified using the MNIST data set. The experimental results show that the scheme FL-SZIDP ensures accuracy while protecting privacy.

Key words: Key words: federal learning, privacy protection, differential privacy, data security

中图分类号: