Method for training neural network model and apparatus

Patent metadata

Field	Value
Publication number	US-11966844-B2
Application number	US-202217981120-A
Country	US
Kind code	B2
Filing date	Nov 4, 2022
Priority date	Dec 29, 2017
Publication date	Apr 23, 2024
Grant date	Apr 23, 2024

Abstract

This application provides a method for training a neural network model and an apparatus. The method includes: obtaining annotation data that is of a service and that is generated by a terminal device in a specified period; training a second neural network model by using the annotation data that is of the service and that is generated in the specified period, to obtain a trained second neural network model; and updating a first neural network model based on the trained second neural network model. In the method, training is performed based on the annotation data generated by the terminal device, so that in an updated first neural network model compared with a universal model, an inference result has a higher confidence level, and a personalized requirement of a user can be better met.

First claim

What is claimed is:

1. A method for training a neural network model, comprising:
obtaining, by a terminal device, annotation data of a service generated by the terminal device in, wherein the service is to be processed by a first neural network model based on performing online inference and a second neural network model based on performing offline inference, and wherein precision of the first neural network model is lower than precision of the second neural network model;
training, by the terminal device, the second neural network model by using the annotation data to obtain a trained second neural network model; and
updating, by the terminal device, the first neural network model based on the trained second neural network model.

2. The method according to claim 1 , wherein the method further comprises:
before updating the first neural network model and in response to determining that the trained second neural network model is an initial update of a second neural network model corresponding to a first version, storing, by the terminal device, the first neural network model;
after updating the first neural network model, receiving, by the terminal device, a second version software package sent by a cloud server, wherein the second version software package comprises a first neural network model corresponding to a second version; and in response to determining that the first neural network model corresponding to the second version is different from the stored first neural network model, updating, by the terminal device, the first neural network model corresponding to the second version.

3. The method according to claim 2 , wherein updating the first neural network model corresponding to the second version comprises:
obtaining, by the terminal device, annotation data of the service generated in a plurality of historical preset periods;
training, by the terminal device by using the annotation data of the service generated in the plurality of historical preset periods, a second neural network model corresponding to the second version, to obtain a trained second neural network model corresponding to the second version; and
updating, by the terminal device based on the trained second neural network model corresponding to the second version, the first neural network model corresponding to the second version.

4. The method according to claim 1 , wherein the terminal device generates the annotation data of the service based on:
performing the online inference on first input data of the service by using the first neural network model, to obtain an online inference result, wherein
if a valid feedback of a user for the online inference result is received, generating, by the terminal device, the annotation data of the service based on the first input data and the valid feedback of the user for the online inference result; or
if no valid feedback of a user for the online inference result is received, generating, by the terminal device, the annotation data of the service based on the first input data and the online inference result after determining that a confidence level of the online inference result is greater than a first threshold.

5. The method according to claim 1 , wherein the terminal device generates the annotation data of the service based on:
performing the offline inference on second input data of the service by using a third neural network model, to obtain an offline inference result, wherein precision of the third neural network model is higher than the precision of the second neural network model, or the third neural network model is the second neural network model; and
in response to determining that a confidence level of the offline inference result is greater than a second threshold, generating, by the terminal device, the annotation data of the service based on the second input data and the offline inference result.

6. The method according to claim 1 , wherein training the second neural network model by using the annotation data comprises:
training the second neural network model when the terminal device is in a charging state.

7. A terminal device, comprising:
a memory storing instructions; and
one or more processors in communication with the memory for executing the instructions to:
obtain annotation data of a service generated by the terminal device in, wherein the service is to be processed by a first neural network model based on performing online inference and a second neural network model based on performing offline inference, and wherein precision of the first neural network model is lower than precision of the second neural network model;
train the second neural network model by using the annotation data to obtain a trained second neural network model; and
update the first neural network model based on the trained second neural network model.

8. The terminal device according to claim 7 , wherein the one or more processors are for executing the instructions to:
before updating the first neural network model and in response to determining that the trained second neural network model is an initial update of a second neural network model corresponding to a first version, store the first neural network model;
receive a second version software package sent by a cloud server, and the second version software package comprises a first neural network model corresponding to a second version; and
in response to determining that the first neural network model corresponding to the second version is different from the stored first neural network model, update, by the terminal device, the first neural network model corresponding to the second version.

9. The terminal device according to claim 8 , wherein the one or more processors are for executing the instructions to:
obtain annotation data that is of the service and that is generated in a plurality of historical preset periods;
train, by using the annotation data that is of the service and that is generated in the plurality of historical preset periods, a second neural network model corresponding to the second version, to obtain a trained second neural network model corresponding to the second version; and
update, based on the trained second neural network model corresponding to the second version, the first neural network model corresponding to the second version.

10. The terminal device according to claim 7 , wherein the one or more processors are for executing the instructions to:
perform the online inference on first input data of the service by using the first neural network model, to obtain an online inference result, wherein
if a valid feedback of a user for the online inference result is received, generate the annotation data of the service based on the first input data and the valid feedback of the user for the online inference result; or
if no valid feedback of a user for the online inference result is received, generate the annotation data of the service based on the first input data and the online inference result after determining that a confidence level of the online inference result is greater than a first threshold.

11. The terminal device according to claim 7 , wherein the one or more processors are for executing the instructions to:
perform the offline inference on second input data of the service by using a third neural network model, to obtain an offline inference result, wherein precision of the third neural network model is higher than the precision of the second neural network model, or the third neural network model is the second neural network model; and
in response to determining that a confidence level of the offline inference result is greater than a second threshold, generate the ann

Assignees

• Huawei Tech Co Ltd

Inventors

• Ma Tao
• Su Qing
• Jin Ying

Classifications

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/0495

  Quantised networks; Sparse networks; Compressed networks · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/091

  Active learning · CPC title