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

Method of text entailment relation recognition based on deep learning

A deep learning and relation recognition technology, applied in the field of deep learning-based text implication relation recognition, to achieve the effect of better understanding tasks, realization of understanding tasks, efficiency and accuracy improvement

Active Publication Date: 2017-10-10
ZHEJIANG UNIV
View PDF2 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Although the amount of data has been accumulated enough and the computing power has been greatly improved, there are still many problems in the relevant intelligent algorithms. There are gaps in natural language understanding. How to make the machine understand the current text and make further reasoning , has become a difficult point in the current natural language understanding

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
  • Method of text entailment relation recognition based on deep learning
  • Method of text entailment relation recognition based on deep learning
  • Method of text entailment relation recognition based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0074] In this embodiment, a textual implication relationship recognition algorithm based on deep learning is specifically performed as follows:

[0075] Get two texts, one of them as the "premise" and the other as the "hypothesis":

[0076] Premise: Two women having drinks and smoking cigarettes at the bar.

[0077] Hypothesis: Two women are at a bar.

[0078] The preset maximum sentence length is 15.

[0079] Step 1: Preprocess the "premise" and "hypothesis" to generate the "premise" string S p =["Two","women","having","drinks","and","smoking","cigarettes","at","the","bar","."] and "hypothesis" String S h =["Two","women","are","at","a","bar","."];

[0080] Step 2: Change the "premise" character S p And the "hypothetical" string S h , Using Str2Matrix method to generate n×l max The "premise" fundamental matrix M p And n×l max The "hypothetical" fundamental matrix M h ;

[0081] Where WV word Represents the n-dimensional word vector of word.

[0082] Step 3: Calculate the distance rela...

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 method of text entailment relation recognition based on deep learning, and belongs to the field of text recognition. The method comprises the steps of: (1) obtaining two texts, wherein one as a 'premise' and the other one as a 'hypothesis', conducting preprocess, and generating a 'premise' string and a 'hypothesis' string; (2) generating a 'premise' basic matrix and a 'hypothesis' basic matrix; (3) calculating the distance relation matrix of 'premise' and 'hypothesis'; (4) generating a joint matrix Up and a joint matrix Uh; (5) sending the joint matrix into a deep learning model separately, generating m dimension double precision 'premise' sentence vector and 'hypothesis' sentence vector; (6) generating classification vectors by combining the sentence vectors of 'premise' and 'hypothesis' according to the sentence vector mixing method; (7) sending the classification vector to a classifier, and the classifier outputting relation between 'premise' and 'hypothesis'.

Description

Technical field [0001] The invention belongs to the field of natural language understanding, and is a method for recognizing the relationship between text implications based on deep learning. Background technique [0002] With the advent of the era of big data, the speed of data growth is getting faster and faster. At the same time, these data are also full of useless and redundant information. It is becoming more and more important for computers to "understand" the meaning of text, and to collect and obtain valuable information from big data. Big data has four characteristics: large amount, high speed, variety, and value. These characteristics make it more and more difficult for computers to quickly obtain valuable information from the Internet. However, once the computer has a deep understanding of the semantics of the text, the computer can automatically collect and organize valuable information on the Internet, which greatly improves labor productivity. [0003] Text entailm...

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): G06F17/30
CPCG06F16/35
Inventor 刘思阳张森林樊臻刘妹琴
Owner ZHEJIANG 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