FLEDGE: Ledger-based Federated Learning Resilient to Inference and Backdoor Attacks

被引:2
|
作者
Castillo, Jorge [3 ]
Rieger, Phillip [1 ]
Fereidooni, Hossein [1 ,2 ]
Chen, Qian [3 ]
Sadeghi, Ahmad-Reza [1 ]
机构
[1] Tech Univ Darmstadt, Darmstadt, Germany
[2] KOBIL GmbH, Worms, Germany
[3] Univ Texas San Antonio, San Antonio, TX 78249 USA
关键词
blockchain; federated learning; homomorphic encryption; security and privacy; FRAMEWORK; PRIVATE;
D O I
10.1145/3627106.3627194
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Federated learning (FL) is a distributed learning process that uses a trusted aggregation server to allow multiple parties (or clients) to collaboratively train a machine learning model without having them share their private data. Recent research, however, has demonstrated the effectiveness of inference and poisoning attacks on FL. Mitigating both attacks simultaneously is very challenging. Stateof-the-art solutions have proposed the use of poisoning defenses with Secure Multi-Party Computation (SMPC) and/or Differential Privacy (DP). However, these techniques are not efficient and fail to address the malicious intent behind the attacks, i.e., adversaries (curious servers and/or compromised clients) seek to exploit a system for monetization purposes. To overcome these limitations, we present a ledger-based FL framework known as FLEDGE that allows making parties accountable for their behavior and achieve reasonable efficiency for mitigating inference and poisoning attacks. Our solution leverages crypto-currency to increase party accountability by penalizing malicious behavior and rewarding benign conduct. We conduct an extensive evaluation on four public datasets: Reddit, MNIST, Fashion-MNIST, and CIFAR-10. Our experimental results demonstrate that (1) FLEDGE provides strong privacy guarantees for model updates without sacrificing model utility; (2) FLEDGE can successfully mitigate different poisoning attacks without degrading the performance of the global model; and (3) FLEDGE offers unique reward mechanisms to promote benign behavior during model training and/or model aggregation.
引用
收藏
页码:647 / 661
页数:15
相关论文
共 50 条
  • [41] XMAM:X-raying models with a matrix to reveal backdoor attacks for federated learning
    Zhang, Jianyi
    Zhang, Fangjiao
    Jin, Qichao
    Wang, Zhiqiang
    Lin, Xiaodong
    Hei, Xiali
    [J]. DIGITAL COMMUNICATIONS AND NETWORKS, 2024, 10 (04) : 1154 - 1167
  • [42] XMAM:X-raying models with a matrix to reveal backdoor attacks for federated learning
    Jianyi Zhang
    Fangjiao Zhang
    Qichao Jin
    Zhiqiang Wang
    Xiaodong Lin
    Xiali Hei
    [J]. Digital Communications and Networks., 2024, 10 (04) - 1167
  • [43] Privacy-Preserving Backdoor Attacks Mitigation in Federated Learning Using Functional Encryption
    Olagunju, Funminiyi
    Adom, Isaac
    Mahmoud, Nabil Mahmoud
    [J]. SOUTHEASTCON 2024, 2024, : 531 - 539
  • [44] Defending Federated Learning from Backdoor Attacks: Anomaly-Aware FedAVG with Layer-Based Aggregation
    Manzoor, Habib Ullah
    Khan, Ahsan Raza
    Sher, Tahir
    Ahmad, Wasim
    Zoha, Ahmed
    [J]. 2023 IEEE 34TH ANNUAL INTERNATIONAL SYMPOSIUM ON PERSONAL, INDOOR AND MOBILE RADIO COMMUNICATIONS, PIMRC, 2023,
  • [45] Comparative Analysis of Membership Inference Attacks in Federated and Centralized Learning
    Abbasi Tadi, Ali
    Dayal, Saroj
    Alhadidi, Dima
    Mohammed, Noman
    [J]. INFORMATION, 2023, 14 (11)
  • [46] FedGrad: Mitigating Backdoor Attacks in Federated Learning Through Local Ultimate Gradients Inspection
    Nguyen, Thuy Dung
    Nguyen, Anh Duy
    Nguyen, Thanh-Hung
    Wong, Kok-Seng
    Pham, Huy Hieu
    Nguyen, Truong Thao
    Le Nguyen, Phi
    [J]. 2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN, 2023,
  • [47] Backdoor Attacks and Defenses in Federated Learning: State-of-the-Art, Taxonomy, and Future Directions
    Gong, Xueluan
    Chen, Yanjiao
    Wang, Qian
    Kong, Weihan
    [J]. IEEE WIRELESS COMMUNICATIONS, 2023, 30 (02) : 114 - 121
  • [48] BadCleaner: Defending Backdoor Attacks in Federated Learning via Attention-Based Multi-Teacher Distillation
    Zhang, Jiale
    Zhu, Chengcheng
    Ge, Chunpeng
    Ma, Chuan
    Zhao, Yanchao
    Sun, Xiaobing
    Chen, Bing
    [J]. IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2024, 21 (05) : 4559 - 4573
  • [49] Multi-level membership inference attacks in federated Learning based on active GAN
    Hao Sui
    Xiaobing Sun
    Jiale Zhang
    Bing Chen
    Wenjuan Li
    [J]. Neural Computing and Applications, 2023, 35 : 17013 - 17027
  • [50] Multi-level membership inference attacks in federated Learning based on active GAN
    Sui, Hao
    Sun, Xiaobing
    Zhang, Jiale
    Chen, Bing
    Li, Wenjuan
    [J]. NEURAL COMPUTING & APPLICATIONS, 2023, 35 (23): : 17013 - 17027