联邦学习安全综述
摘 要
由于信息安全的管控越来越严格,致使许多高校、企业或者相关部门之间的数据不能交换,而联邦学习的出现能够有效解决数据孤岛的问题。但是联邦学习存在许多漏洞,攻击者利用这些漏洞会对隐私安全造成威胁,例如攻击者可能会恶意破坏客户端设备上的训练数据或本地模型更新,再将这些数据发送到中央服务器,还可能拦截中央服务器和客户端设备之间交换的模型更新,用恶意模型去替换他们。本文介绍了联邦学习存在的一些威胁例如隐私推理攻击、中毒攻击等,除此之外,还介绍了当前预防这些攻击的方法的最新研究进展。
关键词:联邦学习;隐私安全;威胁;预防
01
绪论
图1联邦学习流程图
02
联邦学习存在的威胁
03
联邦学习安全的研究进展
图2去中心化的联邦学习
04
未来挑战与机遇
05
结束语
参考文献
[1] Mourad Benmalek,Mohamed Ali Benrekia & Yacine Challal.(2022).Security of Federated Learning: Attacks, Defensive Mechanisms, and Challenges. RIA(1). doi:10.18280/RIA.360106.
[2] K. Wei et al., "Federated Learning With Differential Privacy: Algorithms and Performance Analysis," in IEEE Transactions on Information Forensics and Security, vol. 15, pp. 3454-3469, 2020, doi: 10.1109/TIFS.2020.2988575.
[3] Y. Miao, Z. Liu, H. Li, K. -K. R. Choo and R. H. Deng, "Privacy-Preserving Byzantine-Robust Federated Learning via Blockchain Systems," in IEEE Transactions on Information Forensics and Security, vol. 17, pp. 2848-2861, 2022, doi: 10.1109/TIFS.2022.3196274.
[4] X. Liu, H. Li, G. Xu, Z. Chen, X. Huang and R. Lu, "Privacy-Enhanced Federated Learning Against Poisoning Adversaries," in IEEE Transactions on Information Forensics and Security, vol. 16, pp. 4574-4588, 2021, doi: 10.1109/TIFS.2021.3108434.
[5] G. Xu, H. Li, Y. Zhang, S. Xu, J. Ning and R. H. Deng, "Privacy-Preserving Federated Deep Learning With Irregular Users," in IEEE Transactions on Dependable and Secure Computing, vol. 19, no. 2, pp. 1364-1381, 1 March-April 2022, doi: 10.1109/TDSC.2020.3005909.
[6] G. Xu, H. Li, S. Liu, K. Yang and X. Lin, "VerifyNet: Secure and Verifiable Federated Learning," in IEEE Transactions on Information Forensics and Security, vol. 15, pp. 911-926, 2020, doi: 10.1109/TIFS.2019.2929409.
[7] J. Zhao, H. Zhu, F. Wang, R. Lu, Z. Liu and H. Li, "PVD-FL: A Privacy-Preserving and Verifiable Decentralized Federated Learning Framework," in IEEE Transactions on Information Forensics and Security, vol. 17, pp. 2059-2073, 2022, doi: 10.1109/TIFS.2022.3176191.
[8] L. Zhao, J. Jiang, B. Feng, Q. Wang, C. Shen and Q. Li, "SEAR: Secure and Efficient Aggregation for Byzantine-Robust Federated Learning," in IEEE Transactions on Dependable and Secure Computing, vol. 19, no. 5, pp. 3329-3342, 1 Sept.-Oct. 2022, doi: 10.1109/TDSC.2021.3093711.
[9] Nguyen, T.D., Rieger, P., Chen, H., Yalame, H., Mollering, H., Fereidooni, H., Marchal, S., Miettinen, M., Mirhoseini, A., Zeitouni, S., Koushanfar, F., Sadeghi, A., & Schneider, T. (2022). FLAME: Taming Backdoors in Federated Learning. USENIX Security Symposium.
[10] Xuhong Zhang,Shouling Ji,Jinyin Chen,Jingzheng Wu,Shanqing Guo,Jun Zhou,Alex X. Liu,Ting Wang.T. (2021).Label Inference Attacks Against Vertical Federated Learning.USENIX Security Symposium.
中国保密协会
科学技术分会
长按扫码关注我们
作者:秦康元
责编:高琪
2022年精彩文章TOP5回顾
近期精彩文章回顾