Joint modeling method based on federated learning

Abstract

The invention relates to a joint modeling method based on federated learning, and belongs to the technical field of machine learning. The method includes that a master control node customizes a federated learning strategy according to the target task and controls modeling steps and combination; each participation node performs joint data preprocessing; joint training is performed on each participation node to obtain a final joint model; the master control node distributes the joint model to each participation node to finish synchronization of each node model; and each participation node jointly evaluates the performance of the joint model. According to the invention, the characteristic intermediate indexes of the cooperative terminals are aggregated through the main control terminal, and the characteristics of the overall data are utilized, so that the problems that the feature processing cannot understand the full view of the data and cannot utilize the complete data characteristics are solved; according to the invention, the performance of the joint model under multiple data partitioning modes is obtained through cross validation evaluation, the variance is reduced by averaging the results of the multiple models, and the problem that the model performance is sensitive to the data set partitioning modes is solved.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a joint modeling method based on federated learning. Background technique [0002] Federated learning is an emerging technology based on machine learning, which has received extensive attention from all walks of life in recent years. The so-called federated learning refers to the joint training of machine learning models by multiple participants without exposing local data, which solves the problem of data islands and ensures data privacy and security. [0003] In the existing federated learning technology, the feature processing method is to perform feature processing on each client separately. Since there is no raw data interaction between clients, the feature processing of the federated model cannot understand the whole picture of the data and utilize the complete data characteristics; For the model evaluation part, the solution is that each participant uses local tr...

AI Technical Summary

Problems solved by technology

That is, it needs to manually run federated learning many times, so there are still problems of difficult model optimization and low efficiency

Images