Federal model training method and device, certificate detection method and device, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence detection models, and provides a federal model training method and device, a certificate detection method and device, equipment and a medium, and the method comprises the steps: obtaining a participant complete set initial certificate federal model; screening out a qualified participant subset from the participant complete set according to a preset screening scheme; sending the initial certificate federation model to a participating terminal; and receiving returned model parameters and performance consumption parameters. The Fedavg algorithm is used for aggregating the model parameters to obtain model iteration parameters, and meanwhile the contribution degree and the excitation value are determined according to the model iteration parameters, the model parameters and the performance consumption parameters through the contribution excitation model; and when a preset convergence condition is not met through detection, the qualified participant subset is updated, and the initial certificate federated model is iteratively updated according to the model iteration parameters until convergence is performed to obtain a certificate detection model. According to the invention, the terminal is excited to participate in model training through the contribution degree and the excitation value, and the efficiency and accuracy of federal learning are improved.

Description

technical field [0001] The present invention relates to the technical field of detection models of artificial intelligence, in particular to a federated model training, certificate detection method, device, computer equipment and storage medium. Background technique [0002] With the increasing popularity of machine learning, intelligent applications driven by big data will soon be applied to all aspects of our daily life, including intelligent voice, medical treatment, transportation and so on. However, in the traditional machine learning method, the key to ensuring the accuracy of the training model is to collect a sufficient amount of data, and the data may contain private information about individuals, such as ID photos, etc., which has caused public concerns about personal privacy. concerns about damage. [0003] Recently, federated learning has been widely used due to its significant advantages in privacy protection. However, most existing federated learning systems o...

AI Technical Summary

Problems solved by technology

However, in the traditional machine learning method, the key to ensuring the accuracy of the training model is to collect a sufficient amount of data, and the data may contain private information about individuals, such as ID photos, etc., which has caused public concerns about personal privacy. damage concerns

[0003] Recently, federated learning has been widely used due to its significant advantages in privacy protection. However, most existing federated learning systems optimistically assume that there are enough users willing to participate in federated learning. In fact, due to the participation Terminals need to consume a lot of computing and communication resources, and need to contribute local data. Participating terminals lack the initiative or motivation to join the federated learning process, which is not conducive to the training and promotion of federated learning, the accuracy of model training is not high, and long training period

Images