Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Dependency syntax-based similarity calculation model and system and system building method

A technology of similarity calculation and dependent syntax, applied in the field of question answering systems involving knowledge graphs, can solve problems such as poor question answering performance of knowledge graphs, and achieve the effect of improving performance

Inactive Publication Date: 2021-02-12
SUZHOU UNIV
View PDF3 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] For this reason, the technical problem to be solved by the present invention is to overcome the problem of poor performance of knowledge graph question answering in the prior art, thereby providing a similarity calculation model and system based on dependency syntax that is conducive to improving the knowledge graph question answering performance and building system Methods

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
  • Dependency syntax-based similarity calculation model and system and system building method
  • Dependency syntax-based similarity calculation model and system and system building method
  • Dependency syntax-based similarity calculation model and system and system building method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0024] like figure 2 As shown, this embodiment provides a similarity calculation model based on dependency syntax, including: question semantic coding, which includes shortest dependency path syntax coding, syntax tree-based expression, and pre-trained word vector semantics Encoding; the semantic encoding of the candidate query graph corresponding to the question sentence is used to obtain the semantic encoding of the query graph through semantic encoding of the answer query graph; the semantic encoding of the pre-trained word vector, the syntax encoding of the shortest dependency path, and the expression based on the syntax tree are performed Splicing to obtain the semantic code of the question sentence; the mutual attention mechanism is used for the semantic code of the query graph and the semantic code of the question sentence, information interaction is carried out, and the semantic similarity is obtained through similarity calculation.

[0025] The similarity calculation...

Embodiment 2

[0045] This embodiment provides a system, including the dependency syntax-based similarity calculation model described in Embodiment 1.

[0046] Since the system described in the present embodiment includes the dependency syntax-based similarity calculation model described in the first embodiment, all the advantages of the first embodiment also have all the advantages of the present embodiment.

Embodiment 3

[0048] This embodiment provides a method for building a system, which is used to build the system described in Embodiment 2, including: Step S1: After preprocessing the questions, identify the subject entity of the questions from the questions; Step S2: According to the identified subject entity, construct the one-hop triplet and two-hop triplet of the subject entity; Step S3: Extract the time information and order information of the question sentence, and add the time information and order information to a In jump triplets and double jump triplets, complete the construction of each candidate query graph of the question sentence; step S4: input the query sentence and all query graphs corresponding to the question sentence into the knowledge graph question answering system, and complete the question sentence Calculate the similarity score of its corresponding query graph, select the query graph with the highest score as the optimal query graph corresponding to the question sen...

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 relates to a dependency syntax-based similarity calculation model and system and a system building method, and the method comprises the steps of coding question semantics which comprisesshortest dependency path syntax coding, syntax tree-based expression and pre-trained word vector semantic coding; performing semantic coding on candidate query graphs corresponding to the questions so that answer query graphs are coded by pre-training word vectors to obtain query graph semantic codes; splicing pre-trained word vector semantic codes, shortest dependency path syntax codes and expression based on a syntax tree to obtain a question dependency syntax; and performing a mutual attention mechanism on the query graph semantic code and the question dependency syntax, performing information interaction, and performing similarity calculation to obtain semantic similarity. The performance of the system can be improved.

Description

technical field [0001] The present invention relates to the technical field of knowledge graph question answering systems, in particular to a similarity calculation model and system based on dependency syntax and a method for building the system. Background technique [0002] The implementation methods of Knowledge Graph Question Answering System (KBQA) are mainly divided into two categories: retrieval-based methods and semantic parsing-based methods. [0003] The retrieval-based method is as follows: firstly, entity recognition and entity link processing are performed on the question sentences, and sequence labeling models are mostly used in this part at the present stage. Through these two steps, the subject entity candidates of the questions will be obtained, and then the candidate predicates corresponding to the subject entities will be determined. The similarity model and the question classification model are used more in this part. However, due to the complexity of Ch...

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): G06F16/33G06F16/332G06F16/335G06F40/279G06N3/08
CPCG06N3/08G06F16/3329G06F16/3344G06F16/335G06F40/279
Inventor 陈文亮张鹏举
Owner SUZHOU UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com