With the increasing demand for data privacy protection,Federated Learning could conduct the model training without centralizing data as a Distributed Machine Learning method.However,Federated Learning still faces the risk of participant data leakage in the process of model training.To address this issue,this paper proposes a privacy protection technique for Federated Learning based on Fully Homomorphic Encryption.By introducing FHE technology into the model parameter aggregation phase of Federated Learning,the parameter computation and updates are performed in an encrypted state,so as to ensure the privacy security of the participant data.At the same time,this paper designs an efficient privacy protection framework and conducts experimental validation on multiple public datasets.The experimental results show that the proposed Federated Learning framework not only ensures the accuracy of the model,but also improves the security and computing efficiency of data privacy protection.
维普
万方数据
