Unlock instant, AI-driven research and patent intelligence for your innovation.

Target domain database construction, sample recognition method, terminal and storage medium

A construction method and technology of target domain, applied in target domain database construction, terminal and storage medium, sample identification method field, can solve the problem of high cost

Active Publication Date: 2021-12-24
NUBIA TECHNOLOGY CO LTD
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a new target domain database construction scheme to solve the problem in the prior art that it costs a lot to create a target domain database for a large user group. Aiming at this technical problem, a Target domain database construction, sample identification method, terminal and storage medium

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Target domain database construction, sample recognition method, terminal and storage medium
  • Target domain database construction, sample recognition method, terminal and storage medium
  • Target domain database construction, sample recognition method, terminal and storage medium

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0053] In view of the problem in the existing solutions that the target domain database can only be constructed by collecting and marking as many source domain samples as possible, and the construction cost is high, this embodiment provides a method for constructing the target domain database. Please refer to figure 1 Provided target domain database construction method flow chart:

[0054] S102. Determine the current configuration parameters of the target domain, and perform multiple iterations according to the configuration parameters to obtain the final configuration parameters of the target domain.

[0055] The so-called composition parameters refer to the parameters used to form the target domain database, such as the connection weight w between the visible layer unit v and the hidden layer unit h, the visible layer bias b, and the hidden layer bias c. It should be understood that since the current composition parameters may be adjusted during the iterative processing, the...

no. 2 example

[0081] This embodiment will further introduce the target domain database construction method and sample identification method in the foregoing embodiments with reference to specific examples:

[0082] Similar to the first embodiment, the executors implementing the target domain database construction method and the sample identification method in this embodiment include but are not limited to terminals and servers. Subsequent introductions of this embodiment take the terminal executing the target domain database construction method and the sample identification method as examples for introduction.

[0083] The samples referred to in the foregoing embodiments may be voice samples, and the source domain database may include voice samples of people in a certain region, and the target domain construction method is used to create a target domain voice database capable of recognizing voices in various regions. Correspondingly, the sample recognition method is used to recognize the sp...

no. 3 example

[0124] This embodiment firstly provides a storage medium, where one or more computer programs are stored. In an example of this embodiment, a target domain database construction program is stored in the storage medium, and the program can be read, compiled and executed by a processor, so as to realize the process of the target domain database construction method in the foregoing embodiments. In other examples of this embodiment, a sample identification program is further stored in the storage medium, and the sample identification program is executed by the processor to implement the sample identification method in the foregoing embodiments.

[0125] In addition, this embodiment also provides a terminal, please refer to Figure 8 , the terminal 80 includes a processing 81 , a memory 82 and a communication bus 83 . Wherein the communication bus 83 is used to realize the connection and communication between the processor 81 and the memory 82, the memory 82 is a kind of computer-...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a target domain database construction, a sample identification method, a terminal and a storage medium. Aiming at the problems of high cost of time and money in existing solutions, a method for constructing a target domain database is provided. By determining the current composition of the target domain Parameters, and perform multiple iterative processes to obtain the final composition parameters according to the composition parameters, and finally build the target domain database. In the iterative process, the maximum average difference between the source domain features and the target domain features is constrained by conditions, so that the target domain database constructed based on the final composition parameters obtained by multiple adjustments of the iterative process has a recognition error rate that meets the requirements. The present invention also provides a sample identification method, a terminal and a storage medium. Since a mixed domain database suitable for a large user group can be obtained through the training data expansion of at least one data set, the recognition rate of the target domain database is reduced while ensuring the recognition rate of the target domain database. Collect the cost of large sample collection in traditional protocols.

Description

technical field [0001] The present invention relates to the computer field, and more specifically, relates to a target domain database construction, a sample identification method, a terminal and a storage medium. Background technique [0002] In traditional machine learning, whether it is for voice recognition or character recognition, it is required that the processed data belong to the same distribution, that is, the training data and test data come from the same data set. For the recognition of handwritten digits in character recognition, people have gradually applied various methods to the recognition of handwritten digits, such as using support vector machines, BP (Back Propagation) neural networks, convolutional neural networks, etc. to classify handwritten digits. At the same time, practical applications have verified the excellent performance of these traditional machine learning algorithms in dealing with handwritten digit recognition, but the application scenarios...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/21
CPCG06F16/21
Inventor 刘赣
Owner NUBIA TECHNOLOGY CO LTD