Chinese dialogue knowledge retrieval method based on knowledge retrieval graph and pre-training model

A knowledge retrieval and pre-training technology, applied in neural learning methods, digital data information retrieval, biological neural network models, etc., can solve the problems of complex semantic information and poor dialogue effect, and achieve the effect of good dialogue effect.

Active Publication Date: 2021-03-19
SUN YAT SEN UNIV
View PDF7 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when dealing with topic switching between multiple rounds of dialogues, the existing technologies including the above-mentioned patents are not effective because the semantic information of dialogues and knowledge becomes more complex.

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
  • Chinese dialogue knowledge retrieval method based on knowledge retrieval graph and pre-training model
  • Chinese dialogue knowledge retrieval method based on knowledge retrieval graph and pre-training model
  • Chinese dialogue knowledge retrieval method based on knowledge retrieval graph and pre-training model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] Please refer to figure 1 , a Chinese dialogue knowledge retrieval method based on a knowledge retrieval graph and a pre-trained model, comprising the following steps:

[0054] S01, create a knowledge retrieval data set; the knowledge retrieval data set includes a sample pair data set and a knowledge retrieval graph; the sample pair data set includes a positive sample pair, and the positive sample pair consists of the sample dialogue content and the dialogue with the sample The knowledge triplet corresponding to the content is formed; the knowledge retrieval graph is constructed according to the knowledge triplet of the sample pair data set;

[0055] S02, constructing a pre-training model for obtaining feature representation by using the knowledge-enhanced semantic understanding model;

[0056] S03, by extracting positive sample pairs and randomly constructing negative sample pairs in the sample pair data set according to the preset batch size, constructing a training d...

Embodiment 2

[0095] A Chinese dialogue knowledge retrieval system based on knowledge retrieval graph and pre-trained model, see Figure 4 , including a knowledge retrieval data set creation module 1, a pre-training model construction module 2, a pre-training model training module 3, and a dialogue content acquisition retrieval module 4 to be processed; the pre-training model training module 3 is connected to the knowledge retrieval data set creation module 1 and a pre-training model building module 2, the dialogue content to be processed acquisition retrieval module 4 is connected to the retrieval library creation module 1 and the pre-training model training module 3, wherein:

[0096] The knowledge retrieval data set creation module 1 is used to create a knowledge retrieval data set; the knowledge retrieval data set includes a sample pair data set and a knowledge retrieval graph; the sample pair data set includes a positive sample pair, and the positive sample pair consists of The sample ...

Embodiment 3

[0101] A storage medium on which a computer program is stored, and when said computer program is executed by a processor, it realizes the steps of the Chinese dialogue knowledge retrieval method based on a knowledge retrieval graph and a pre-trained model in embodiment 1.

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 provides a Chinese dialogue knowledge retrieval method based on a knowledge retrieval graph and a pre-training model in order to solve the technical problem that in the prior art, the performance is poor when topic switching between multiple rounds of dialogues is processed. According to the method, text semantic information is modeled by utilizing a pre-training model, and the relationship between knowledge is modeled by utilizing knowledge retrieval graph structure information, so that knowledge triples related to subsequent replies of a current dialogue can be retrieved more accurately, and subsequently generated knowledge dialogues can better meet topic switching among multiple rounds of dialogues, and a better dialogue effect is obtained.

Description

technical field [0001] The present invention relates to the technical field of natural language processing, in particular to the application of natural language processing in man-machine dialogue, and more specifically, to a Chinese dialogue knowledge retrieval method, system, storage medium and computer equipment. Background technique [0002] Natural language processing technology has always been committed to enabling machines to communicate fluently and freely like humans. This is also the original intention of many voice assistants on the market, such as Siri, Xiaodu, etc. While the dialogue is smooth, the dialogue responses produced by it can be rich in knowledge, not just responding to the content of the dialogue. Therefore, when providing training corpus for the machine, in addition to providing the corresponding historical dialogue information, it is also necessary to provide the corresponding dialogue knowledge, so that the machine can generate knowledgeable dialog...

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 Applications(China)
IPC IPC(8): G06F16/332G06F16/33G06F16/36G06K9/62G06N3/04G06N3/08
CPCG06F16/3329G06F16/3344G06F16/367G06N3/08G06N3/045G06F18/214
Inventor 戴斯铭潘嵘毛明志
Owner SUN YAT SEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products