System for realizing data set construction and processing based on federated learning and its construction and generation method

Abstract

The invention relates to a system for realizing data set construction and processing based on federated learning, including a data set management module connected with historical data sets for exporting and managing data sets; a data set generation module and the data set management module It is used to connect the participating nodes to complete the generation of the data set of the node according to the data generation rules, and feed back the status to the central node. The invention also relates to a generation method for implementing data set construction based on federated learning. The federated learning-based data set construction processing system of the present invention and its construction and generation method are adopted. Unlike the previous federated learning, which needs to rely on its own sample set for training and use, the federated learning data set construction system provides multi-rule and multi-mode The simulation data generation mode provides the required data sets for the learning and training of each node of federated learning, which can effectively expand the data usage link and cover a variety of usage scenarios.

Description

technical field [0001] The present invention relates to the field of data processing, in particular to the field of artificial intelligence, and specifically refers to a system for realizing data set construction and processing based on federated learning and a construction and generation method thereof. Background technique [0002] Federated Learning (Federated Learning) is an emerging artificial intelligence basic technology, which was first proposed by Google in 2016. Under the premise of regularization, efficient machine learning can be carried out among multi-participants or multi-computing nodes. Federated learning requires a large amount of data sets in the training and model evaluation stages, but in actual use, data sets are often scarce resources, especially in the system acceptance stage, the existing data sets often cannot cover various training and use of algorithms Scenes. To this end, a federated learning dataset construction scheme is proposed to facilitat...

AI Technical Summary

Problems solved by technology

Federated learning requires a large amount of data sets in the training and model evaluation stages, but in actual use, data sets are often scarce resources, especially in the system acceptance stage, the existing data sets often cannot cover various training and use of algorithms Scenes

Images