Federal learning method and device, electronic equipment and computer readable storage medium

Abstract

The embodiment of the invention discloses a federal learning method and device, electronic equipment and a computer readable storage medium. According to the embodiment of the invention, after the local models are trained and the first model information of the first local model after training is sent to the server, the aggregated first federated model returned by the server is received, the first federated model is verified, the first verification information is subjected to disturbance processing, and the first model information of the first local model after training is obtained. Updating a preset perturbed first verification information set based on the perturbed first verification information, screening out target perturbed first verification information from the updated perturbed first verification information set, and constructing a model index corresponding to the target perturbed first verification information; sending the first verification information after target disturbance and the model index to a server, so that the server generates a trained federal model; according to the scheme, the training efficiency of federal learning can be improved, and the embodiment of the invention can be applied to various scenes such as cloud technology, artificial intelligence, intelligent traffic and auxiliary driving.

Description

technical field [0001] The present invention relates to the field of communication technologies, and in particular, to a federated learning method, apparatus, electronic device and computer-readable storage medium. Background technique [0002] With the development of computer computing and the advancement of artificial intelligence technology, federated learning has gradually become a hot topic. Federated learning completes the model of machine learning and deep learning models through multi-party cooperation. Federated learning also includes horizontal federated learning. In horizontal federated learning, multiple computing devices use their own sample data to jointly train the same model. In order to ensure the security of sample data, cryptographic security mechanisms are usually used to protect the training data of the participants. . [0003] In the process of research and practice of the prior art, the inventor of the present invention found that the security mechani...

AI Technical Summary

Problems solved by technology

[0003] During the research and practice of the existing technology, the inventors of the present invention found that the security mechanism based on cryptography requires encryption of the information exchanged between the server and the participants, which increases the communication overhead, and requires the server and participants to Multiple interactions between parties have high requirements for network communication bandwidth and stability. Therefore, the training efficiency of federated learning is low.

Images