Privacy-Preserving and Byzantine-Robust Federated Learning
IEEE Transactions on Dependable and Secure Computing · 2024-03 · 21 (2) · 889-904
DOI: 10.1109/tdsc.2023.3264697 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Dependable and Secure Computing, 2024-03. Verfasst von Caiqin Dong und 6 weiteren Autor:innen. Die Arbeit wurde bisher 84-mal zitiert (Crossref) und nennt 49 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Science and Technology Major Project of Tibetan Autonomous Region of China, National Joint Engineering Research Center of Network Security Detection and Protection Technology.
Abstract (Originalsprache)
Federated learning (FL) trains a model over multiple datasets by collecting the local models rather than raw data, which can help facilitate distributed data analysis in many real-world applications. Since the model parameters can leak information about the training datasets, it is necessary to preserve the privacy of the FL participants’ local models. Furthermore, FL is vulnerable to poisoning attacks which can significantly decrease the model utility. To settle the above issues, we propose a privacy-preserving and Byzantine-robust FL scheme $\Pi _{\text{P2Brofl}}$ Π P2Brofl that maintains robustness in the presence of poisoning attacks and preserves the privacy of local models simultaneously. Specifically, $\Pi _{\text{P2Brofl}}$ Π P2Brofl leverages three-party computation (3 PC) to securely achieve a Byzantine-robust aggregation method. To improve the efficiency of privacy-preserving local model selection and aggregation, we propose a maliciously secure top- $k$ k protocol $\Pi _{\text{top}-k}$ Π top - k that has low communication overhead. Moreover, we present an efficient maliciously secure shuffling protocol $\Pi _{\text{shuffle}}$ Π shuffle since secure shuffling is necessary for our secure top- $k$ k protocol. The security proof of the scheme is given and experiments on real-world datasets are conducted in this paper. When the proportion of Byzantine participants is 50%, the error rate of the model only increases by 1.05% while it increases by 23.78% without using our protection.
Förderung
- National Natural Science Foundation of China
- Science and Technology Major Project of Tibetan Autonomous Region of China
- National Joint Engineering Research Center of Network Security Detection and Protection Technology
- Guangdong Key Laboratory of Data Security and Privacy Preserving
- Guangdong Hong Kong Joint Laboratory for Data Security and Privacy Protection
- Engineering Research Center of Trustworthy AI
- Natural Science Foundation of Guangdong Province
- Dongguan Social Development Science and Technology Key Project
Zitieren
APA
Dong, C., Weng, J., Li, M., Liu, J., Liu, Z., Cheng, Y., & Yu, S. (2024). Privacy-Preserving and Byzantine-Robust Federated Learning. IEEE Transactions on Dependable and Secure Computing, 21(2), 889-904. https://doi.org/10.1109/tdsc.2023.3264697
Vancouver
Dong C, Weng J, Li M, Liu J, Liu Z, Cheng Y, et al. Privacy-Preserving and Byzantine-Robust Federated Learning. IEEE Transactions on Dependable and Secure Computing. 2024;21(2):889-904. doi:10.1109/tdsc.2023.3264697
BibTeX
@article{Dong2024,
author = {Dong, Caiqin and Weng, Jian and Li, Ming and Liu, Jia-Nan and Liu, Zhiquan and Cheng, Yudan and Yu, Shui},
title = {Privacy-Preserving and Byzantine-Robust Federated Learning},
journal = {IEEE Transactions on Dependable and Secure Computing},
year = {2024},
volume = {21},
number = {2},
pages = {889--904},
doi = {10.1109/tdsc.2023.3264697}
}
EUVIMED