A Comparative Review of Privacy Protection Techniques in Federated Learning

Jiayi Li

doi:10.47297/taposatWSP2633-456910.20250611

PDF
导出
分享
收藏
专辑

A Comparative Review of Privacy Protection Techniques in Federated Learning

6卷第11期

浏览量: 0 Downloads: 0

A Comparative Review of Privacy Protection Techniques in Federated Learning
A Comparative Review of Privacy Protection Techniques in Federated Learning
School of Data Science and Artificial Intelligence, Wenzhou Institute of Technology, Wenzhou, Zhejiang, 325000, China
2025年6卷第11期页码：49-54
纸质出版：2025-11-30
DOI：10.47297/taposatWSP2633-456910.20250611
Accepted：

Scan QR Code

Jiayi Li. (2025). A Comparative Review of Privacy Protection Techniques in Federated Learning. 科技理论与实践（英文版）, 6(11), 49-54.

Li Jiayi.A Comparative Review of Privacy Protection Techniques in Federated Learning[J].Theory and Practice of Science and Technology,2025,06(11):49-54.
Jiayi Li. (2025). A Comparative Review of Privacy Protection Techniques in Federated Learning. 科技理论与实践（英文版）, 6(11), 49-54. DOI： 10.47297/taposatWSP2633-456910.20250611.

Li Jiayi.A Comparative Review of Privacy Protection Techniques in Federated Learning[J].Theory and Practice of Science and Technology,2025,06(11):49-54. DOI： 10.47297/taposatWSP2633-456910.20250611.

摘要

Abstract

This study addresses privacy protection challenges in federated learning by establishing a six-dimensional comparative framework encompassing "dataset type

accuracy loss

computational overhead

communication cost

privacy strength

and application domain." It systematically reviews representative works on differential privacy (DP)

homomorphic encryption (HE)

secure multi-party computation (SMPC)

and their engineering variants MPHE-1

MHESA

and QSFL. Establishing a unified evaluation metric

we conclude: DP incurs minimal communication overhead but sacrifices accuracy; HE achieves near-cleartext accuracy with substantial computational overhead; SMPC offers strong security but is constrained by interaction rounds and network conditions; the three variants respectively address low-round scenarios

dropout scenarios

and edge efficiency scenarios. This paper provides contextualized selection recommendations and a reproducibility checklist

while proposing future research directions for large models and real-world network environments. It offers methodological and engineering references for compliant and implementable privacy-preserving federated learning.

关键词

Keywords

references

Dwork, C. (2006). Differential privacy. In Automata, languages and programming: 33rd international colloquium, ICALP 2006 (Vol. 4052, pp. 1–12). Springer.

Liu YX, Chen H, L iu YH, Li CP. Privacy-preserving Techniques in Federated Learning. Journal of Software, 2022, 33(3): 1057-1092 (in Chinese) .

Cheon J H, Kim A, Kim M, Song Y. Homomorphic encryption for arithmetic of approximate numbers. In: Takagi T, Peyrin T, eds. Advances in Cryptology – ASIACRYPT 2017. Lecture Notes in Computer Science, vol. 10624 . Cham: Springer, 409–437, 2017.

Teng WANG, Zheng HUO, Yaxin HUANG, Yilin FAN. Review on privacy-preserving technologies in federated learning[J]. Journal of Computer Applications, 2023, 43(2): 437-449.

Yao, A. C. (1982). Protocols for secure computations. In Proceedings of the 23rd Annual Symposium on Foundations of Computer Science (FOCS '82) (pp. 160–164). IEEE.

Goldreich, O., Micali, S., & Wigderson, A. (1987). How to play any mental game—A completeness theorem for protocols with honest majority. In Proceedings of the 19th Annual ACM Symposium on Theory of Computing (STOC '87) (pp. 218–229). ACM.

Ben-Or, M., Goldwasser, S., & Wigderson, A. (1988). Completeness theorems for non-cryptographic fault-tolerant distributed computation. In Proceedings of the 20th Annual ACM Symposium on Theory of Computing (STOC '88) (pp. 1–10). ACM.

Hosseini, E., Chen, S., & Khisti, A. (2024). Secure Aggregation in Federated Learning using Multiparty Homomorphic Encryption. arXiv preprint.

Gao, Q., Sun, Y., Chen, X., Yang, F., & Wang, Y. (2024). An efficient multi-party secure aggregation method based on multi-homomorphic attributes. Electronics, 13(4), Article 671.

Tran, A. T., Luong, T. D., & Pham, X. S. (2024). Privacy-preserving deep learning model with integer quantization and secure multi-party computation. Annals of Operations Research.

Abadi, M., Chu, A., Goodfellow, I., McMahan, H. B., Mironov, I., Talwar, K., & Zhang, L. (2016). Deep learning with differential privacy. Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security (CCS '16) (pp. 308–318). ACM.

Hu, C., & Li, B. (2024). MASKCRYPT: Federated learning with selective homomorphic encryption. IEEE Transactions on Dependable and Secure Computing. Advance online publication.

Koti, N., Ranellucci, S., Srinivasan, A., Rosulek, M., Trieu, N., Mandal, A., & Wang, X. (2022). Tetrad: Actively secure 4PC for secure training and inference. Proceedings of the Network and Distributed System Security Symposium (NDSS 2022).

Wagh, S., Gupta, D., & Chandran, N. (2019). SecureNN: 3-party secure computation for neural network training. Proceedings on Privacy Enhancing Technologies, 2019(3), 26–49.

文章被引用时，请邮件提醒。

Submit

A Survey of Privacy-Preserving Techniques in Federated Graph Neural Networks

A Review of the Principles and Typical Applications of Differential Privacy Technology

Privacy Protection of Clustering High-Dimensional Data Based on African Vultures

Related Author

Lingyu Wang School of Arts and Sciences, China University of Petroleum-Beijing at Karamay

Ketong Liu Reading College, Nanjing University of Information Science and Technology

Chenglin Zhang School of Information & Electronics Technology, Jiamusi University

Lili He School of Information & Electronics Technology, Jiamusi University

Related Institution

School of Arts and Sciences, China University of Petroleum-Beijing at Karamay

Reading College, Nanjing University of Information Science and Technology

School of Information & Electronics Technology, Jiamusi University

⁰