Federal learning-based machine learning method

Abstract

The invention discloses a federal learning-based machine learning method. The method comprises the following steps that training nodes participating in federated learning input training data; the training node performs feature processing on the training data to obtain feature data; the training node adopts the feature data to perform Poisson regression algorithm-based model training of a machine learning model; in a primary iteration process, each training node participating in training sends gradient information to the forwarding node, then obtains gradient information of other nodes from the forwarding node, and updates and calculates local gradient information; the training node updates the model weight of the local node through the updated gradient information; the training node judges whether the machine learning model converges or not, and if yes, iteration is quitted. The federated learning-based machine learning method provided by the invention has the beneficial effect of effectively enabling each training node to interact with the intermediate data through a node forwarding mode.

Description

technical field [0001] This application relates to the field of data processing, in particular, to a machine learning method based on federated learning. Background technique [0002] In the near future, the medical industry will incorporate more high-tech technologies such as artificial intelligence and sensor technology, so that medical services will become truly intelligent and promote the prosperity and development of medical care. Under the background of China's new medical reform, smart medical care is entering the lives of ordinary people. Data in the medical industry needs privacy protection. Therefore, when artificial intelligence is applied to the medical field for research, model training, and data prediction, it often requires multiple medical institutions to carry out networking and data collaboration. [0003] In the prior art, when training a machine learning model based on federated learning, the data generated during the training cannot interact well, resul...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, when training a machine learning model based on federated learning, the data generated during the training cannot interact well, which leads to the inability of the model to converge, which in turn affects the efficiency of platform model training

Images