MODEL TRAINING METHOD, DATA PROCESSING METHOD, AND APPARATUS: abstract simplified (18327952)

• This abstract for appeared for patent application number 18327952 Titled 'MODEL TRAINING METHOD, DATA PROCESSING METHOD, AND APPARATUS'

Simplified Explanation

This application describes a method for training a machine learning system that consists of a server and multiple client devices. The method involves the following steps: 1. The server sends a shared model to a client device. 2. The client device uses the shared model to make predictions on a dataset and calculates a loss value based on these predictions. 3. The client device also uses its own private model to make predictions on the same dataset and calculates a second loss value. 4. The client device combines the first and second loss values to obtain a third loss value. 5. The third loss value is used to update the client device's private model.

Original Abstract Submitted

This application discloses a model training method, and relates to the field of artificial intelligence. The method provided in this application is applicable to a machine learning system. The machine learning system includes a server and at least two client side devices. The method includes: A first client side device receives a first shared model sent by the server; outputs a first prediction result for a data set through the first shared model; obtains a first loss value based on the first prediction result; outputs a second prediction result for the data set through a first private model of the first client side device; obtains a second loss value based on the second prediction result; and performs second combination processing on the first loss value and the second loss value to obtain a third loss value, where the third loss value is used to update the first private model.