Artificial intelligence training model method and device based on block chain, and medium

Abstract

The invention discloses an artificial intelligence training model method and device based on a blockchain, and a medium. The method comprises the steps of determining a blockchain platform created inadvance based on a blockchain framework; receiving training data which is provided by each data providing node and is used for training a model; obtaining a model generated by training the training node through the training data; and writing the model into a blockchain platform. The whole process from training data providing to training completion is stored in the block chain platform. As the datain the blockchain platform is public and transparent, each node can inquire the training data and the model through a corresponding way at any time. As the blockchain platform is in distributed storage, the data tampering of a single node does not take effect, and the truth and credibility of the data on the blockchain platform are ensured. Therefore, the training data is difficult to be tampered, and the accuracy of the model is ensured. The model is written into the blockchain platform, so that the model is difficult to tamper.

Description

technical field [0001] This application relates to the field of artificial intelligence, in particular to a blockchain-based artificial intelligence training model method, equipment and media. Background technique [0002] In recent years, with the development of science and technology, artificial intelligence technology has developed rapidly, such as smart cities, unmanned driving, AI robots, etc., which have brought many conveniences to our lives. [0003] Usually, when using artificial intelligence technology, it is necessary to train a generative model according to the corresponding data. However, in the prior art, the data used for training the model needs to be collected, and during the collection process, the data may be tampered due to various reasons, thereby affecting the training of the model. And after the training is completed, in the process of saving and transmitting the model, there is also the possibility that the model parameters are maliciously tampered w...

AI Technical Summary

Problems solved by technology

However, in the existing technology, the data used for training the model needs to be collected, and during the collection process, due to various reasons, the data may be tampered with, thereby affecting the training of the model

And after the training is completed, in the process of saving and transmitting the model, there is also the possibility that the model parameters are maliciously tampered with

Images