AI algorithm model training platform and model training method fused with federated learning

Through the AI ​​algorithm model training platform that integrates federated learning, the problem of traditional platforms lacking flexibility and data security protection is solved, efficient and customized model training and powerful data security are achieved, and compliance and reliability of AI applications are ensured.

CN119721290APending Publication Date: 2025-03-28CLOUDCHAIN GRP CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411684407.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-22
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

Traditional AI development platforms lack flexibility and data security protection capabilities, making them difficult to meet customized needs and effectively ensure the security and privacy of enterprise data.

Method used

The AI ​​algorithm model training platform that integrates federated learning is adopted. The platform provides highly flexible and customized model training capabilities through data processing modules, model building modules, model training modules and model management modules, while deeply integrating federated learning modes to protect data privacy and enhance data security.

Benefits of technology

It realizes the high flexibility and customization of the AI ​​algorithm model training platform, improves the efficiency and effectiveness of model training, and significantly enhances the security and privacy protection capabilities of data, avoids the risks of data leakage and abuse, and ensures the compliance and reliability of AI applications.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides an AI algorithm model training platform and a model training method fused with federated learning, and the platform comprises a data processing module which is used for receiving original data for model training, carrying out the preprocessing of the original data, extracting feature data for model training from the original data, and transmitting the feature data to a database; constructing a training set, a verification set and a test set based on the preprocessed original data and the extracted feature data; the model construction module is used for constructing an initial model for training in a mode of preset model parameter adjustment, visual modeling or automatic modeling; the model training module is used for training based on a federated learning method to obtain an AI algorithm model; and the model management module is used for managing the trained AI algorithm model by using the verification set and the test set. The platform and the method have high flexibility and are easy to customize, and AI algorithm models can be trained in a customized and customized manner according to different requirements.
Need to check novelty before this filing date? Find Prior Art