Privacy Enhanced Federated Learning via Privacy Masks and Additive Homomorphic Encryption
Authors:
- 1 August 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 471-478
Abstract
Federated learning has become an effective method to realize collaborative learning among multiple data centers. However, the problem of privacy leakage has not been perfectly solved. We proposed a scheme based on homomorphic encryption with random privacy masks to cope with the problem of private leakage caused by honest but curious participants. Through combining Paillier encryption scheme with privacy mask, security of the learning process is enhanced with a trivial computation and communication cost. Experimental results show that the decryption time is reduced by more than half after using CRT optimization, without degrading the training precision.Keywords
Funding Information
- National Natural Science Foundation of China (62102452,62172436)
- Natural Science Foundation of Shaanxi Province (2023-JC-YB-584)
This publication has 38 references indexed in Scilit:
- Machine learning: Trends, perspectives, and prospectsScience, 2015
- What Is Machine Learning?2015
- Efficient distance metric learning by adaptive sampling and mini-batch stochastic gradient descent (SGD)Machine Learning, 2014
- (Leveled) Fully Homomorphic Encryption without BootstrappingACM Transactions on Computation Theory, 2014
- Homomorphic Encryption from Learning with Errors: Conceptually-Simpler, Asymptotically-Faster, Attribute-Based2013
- Fully Homomorphic Encryption over the Integers2010
- Fully homomorphic encryption using ideal lattices2009
- Paillier Encryption and Signature Schemes2006
- Public-Key Cryptosystems Based on Composite Degree Residuosity Classes1999
- Efficient Public-Key Cryptosystems Provably Secure Against Active Adversaries1999