Model processing method and system, terminal equipment and server

Abstract

The invention discloses a model processing method and system, terminal equipment and a server. According to the method, the technical concept that a to-be-adjusted model (the model is constructed by use of model adjustment information sent by the server by use of at least one set of terminal equipment) acquired from the server is adjusted by use of data of a terminal on the terminal equipment, model adjustment information corresponding to the adjusted model is sent to the server, so that the server adjusts the model on the basis of the model adjustment information from the terminal is proposed. By use of the scheme, model training by use of terminal data can be realized, the trained model has a better model effect with no need of uploading terminal data to the server by the terminal equipment, and the privacy of user data in the terminal equipment is guaranteed.

Description

Technical field [0001] The invention belongs to the technical field of deep learning, and particularly relates to a model processing method, system, terminal device and server. Background technique [0002] With the continuous advancement of computer and terminal technology, it is becoming increasingly important to use big data and deep learning methods to train neural network models, and to use mobile smart terminals for image, voice, natural language and other reasoning applications to improve user experience. . [0003] At present, there are mainly two ways to train neural network models. One is that the server uses its own data to train the model and adapts the trained model to the mobile terminal to directly perform inference applications. This method generally uses some public As the server’s own data, the data set is quite different from the user’s real data, which will lead to the relatively poor effect of the trained model; the other is the user’s real interaction data (s...

AI Technical Summary

Problems solved by technology

[0003] At present, there are mainly two ways to train the neural network model. One is to use the server to use its own data to train the model, and adapt the trained model to the mobile terminal for direct reasoning and application. This method generally uses some public As the server's own data, the data set is quite different from the user's real data, which will lead to a relatively poor effect of the trained model; the other is the user's real interaction data (such as voice, image , text and other data) to the server, and the server trains the model based on the user’s real interaction data, and then the terminal device downloads the model for use. The obvious defect of this method is that the privacy of user data is difficult to guarantee

Images