Federated learning-oriented localized differential privacy protection method and system, computer equipment and storage medium

The invention provides a federated learning-oriented localized differential privacy protection method and system, computer equipment and a storage medium. The method comprises the following steps of sending a first model and a first model parameter to a client by a server; enabling the client to use local data to train the first model, and updating the first model parameter to a second model parameter; adopting a localized differential privacy algorithm, and adding disturbance to the second model parameter by the client to obtain a third model parameter; enabling the client to randomly select part of the third model parameters, exchange the third model parameters with the third model parameters at the corresponding positions of the other client, generate fourth model parameters and send the fourth model parameters to the server; according to the method, a powerful privacy protection effect is provided for sensitive data of the user, the privacy budget is saved, and meanwhile, the service quality of the model