Privacy Preserving Machine Learning with Homomorphic Encryption and Federated Learning

Privacy protection has been an important concern with the great success of machine learning. In this paper, it proposes a multi-party privacy preserving learning framework, named PFMLP, based on partially homomorphic encryption and federated The core idea is all parties just transmitting encrypted gradients by encryption. From experiments, model trained PFMLP almost same accuracy, deviation less than 1%. Considering computational overhead encryption, we use improved Paillier algorithm which can speed up training 25–28%. Moreover, comparisons key length, network structure, number clients, etc. are also discussed in detail paper.