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

Training method and training device based on semantic recognition and terminal equipment

A technology of semantic recognition and training method, applied in the field of semantic recognition, can solve the problem of low efficiency of voice interaction, and achieve the effect of improving processing efficiency and accuracy, enhancing ease of use and practicability, and expanding the scope of recognition

Pending Publication Date: 2020-06-30
UBTECH ROBOTICS CORP LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present invention provides a semantic recognition-based training method, training device and terminal equipment to solve the problems in the prior art that voice matching is limited by the operation instruction library and the efficiency of voice interaction is very low when there are too many operation instructions. question

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
  • Training method and training device based on semantic recognition and terminal equipment
  • Training method and training device based on semantic recognition and terminal equipment
  • Training method and training device based on semantic recognition and terminal equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0067] see figure 1 , is a schematic diagram of the implementation flow of the semantic recognition-based training method provided by the embodiment of the present invention. The method is applied to intelligent terminal devices capable of voice interaction, such as robots, mobile phones, computers, tablet computers, or smart home products. The smart terminal device judges the operation instruction generated by voice recognition, obtains the intention of the operation instruction, and executes the action of the corresponding operation instruction.

[0068] As shown, the method includes the following steps:

[0069] Step S101, preprocessing the preset operation instruction corpus to obtain a training set based on the first operation instruction text.

[0070] In this embodiment, the preset operating instruction corpus may be pre-configured corpus of smart terminal equipment, including colloquial operating instruction corpus without keywords. The operating instruction corpus ...

Embodiment 2

[0102] refer to figure 2 , is a schematic diagram of the implementation flow of the semantic recognition-based training method provided by another embodiment of the present invention. This embodiment further verifies the prediction model of the first embodiment for the operation instruction text after speech recognition.

[0103] As shown, the method includes:

[0104] Step S201, receiving operation instruction information generated through voice recognition.

[0105] In this embodiment, the input voice is recognized by the voice recognition device, and the corresponding operation instruction information is acquired. The operation instruction information generated according to the input voice includes content such as stop words or time generated by a pause in the middle.

[0106] Step S202, performing formatting processing on the operation instruction information, and acquiring a second operation instruction text corresponding to the operation instruction information.

[0...

Embodiment 3

[0118] see image 3 , is a schematic diagram of a training device based on semantic recognition provided in Embodiment 3 of the present invention. For ease of description, only parts related to the embodiment of the present invention are shown.

[0119] The training device 300 based on semantic recognition includes:

[0120] The processing unit 31 is configured to preprocess the preset operation instruction corpus, and obtain a training set based on the first operation instruction text;

[0121] The training unit 32 is used to perform one-level training on the preprocessed operation instruction corpus, and obtain an intermediate vector model based on the operation instruction corpus;

[0122] Establishing a mapping unit 33, configured to establish a mapping relationship between the intermediate vector model and the training set;

[0123] The model building unit 34 is configured to perform secondary training on the intermediate vector model corresponding to the training set a...

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 is suitable for the technical field of semantic recognition, and provides a training method and a training device based on semantic recognition and terminal equipment, and the training method comprises the steps: carrying out the preprocessing of a preset operation instruction corpus, and obtaining a training set based on a first operation instruction text; performing primary training on the preprocessed operation instruction corpus to obtain an intermediate vector model based on the operation instruction corpus; establishing a mapping relationship between an intermediate vectormodel and the training set; according to the mapping relationship, carrying out secondary training on an intermediate vector model corresponding to the training set to obtain a prediction model basedon an operation instruction intention. According to the method, the acquisition of the operation instruction intention is trained through deep learning, and the characteristics of different operationinstructions can be acquired, so that the terminal equipment executing the operation instructions can understand the operation instruction intention more easily, the range of identifying the operationinstructions by the terminal equipment is expanded, and the processing efficiency and accuracy of the operation instructions are improved.

Description

technical field [0001] The invention belongs to the technical field of semantic recognition, and in particular relates to a training method, training device and terminal equipment for semantic recognition. Background technique [0002] With the development of voice recognition technology, voice interaction with various smart terminal devices is used in more and more scenarios; for example, through voice interaction with robots, the robot is given operating instructions to control the robot to perform corresponding operations. [0003] At present, to obtain the corresponding text after speech recognition, it is necessary to configure a large number of relevant operation instructions on the smart terminal device, and to match the text after speech recognition through keywords or regular expressions to obtain the operation instructions corresponding to the speech; However, for colloquial speech, there is a difference in the text or the order of the text in the preset operation ...

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
IPC IPC(8): G10L15/22G10L15/18G10L15/06
CPCG10L15/22G10L15/1822G10L15/063G10L2015/223G10L2015/0631
Inventor 熊友军罗沛鹏廖洪涛
Owner UBTECH ROBOTICS CORP LTD