Entity relationship extracting method based on deep neural network

A deep neural network and entity relationship technology, applied in the field of machine learning and natural language processing, can solve the problems of complex relationship extraction in Chinese, complex Chinese language features, and uncommon relationship type system, so as to improve the convergence speed and accuracy, improve the Accuracy and performance, the effect of simplifying manual workload

Inactive Publication Date: 2016-12-07
WUHAN UNIV OF TECH
View PDF8 Cites 36 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] At present, the construction of Chinese knowledge base still mainly has problems such as low utilization rate of unstructured text and low coverage of entities in specific fields. Complex lang

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
  • Entity relationship extracting method based on deep neural network
  • Entity relationship extracting method based on deep neural network
  • Entity relationship extracting method based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0031] Such as figure 1 As shown, an entity relationship extraction method based on deep neural network, the neural network structure diagram of this method is shown in figure 2 As shown, the method includes:

[0032] Step 1, each word or category keyword of the sentence is mapped to a word vector or category vector;

[0033] Using unlabeled corpus training to obtain word vectors with semantic information, that is, using a dictionary of size D, the sentence sequence {c 1 , c 2 ,... c n} After this layer of neural network will be mapped to a real number vector sequence {W 1 , W 2 ,...W n}, where W n is the word c n The word vector, such as image 3 shown. In this embodiment, pre-trained word vectors obtained based on Hikolov's Word2Vec tool are used.

[0034] Using unlabeled corpus training to obtain category keyword vectors with semantic in...

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 an entity relationship extracting method based on a deep neural network. The entity relationship extracting method comprises the following steps of: mapping each word or class keyword of a sentence to a word vector or a class vector respectively; carrying out characteristic extraction on the sentence according to the word vector and the class vector; and connecting extracted characteristics end to end and inputting into a whole-connection classification layer to obtain an extracted result. An entity relationship of a text is extracted by utilizing a common neural network and a convolutional neural network in a machine learning process, the accuracy and performance of the extraction of the entity relationship are improved and the artificial workload in the extraction of the entity relationship is simplified. The pre-trained word vector is utilized and the convergence rate and accuracy of the neural network are improved; and sentence characteristics and class characteristics are introduced and the convolutional neural network and the common neural network are used for extracting, so that the problem of long and short sentences is solved and the performance of the extraction of the entity relationship is improved.

Description

technical field [0001] The invention relates to machine learning and natural language processing, in particular to an entity relationship extraction method based on a deep neural network. Background technique [0002] Human-like intelligence is developing rapidly, and corresponding products have been used in education, medical and other industries, such as the Japanese Todai Robot project for primary education examination questions and the American IBH Watson project dedicated to intelligence question answering and expanding to the medical field. In China, the basic educational resources represented by "Wenzong" contain a wealth of knowledge, and a complete and high-quality knowledge base determines the intelligence level of a human-like intelligent question answering system. [0003] Deep learning technology has made major breakthroughs in the field of images and speech recognition, and has also been in full swing in the field of natural language. Among these deep learning...

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): G06F17/27
CPCG06F40/205G06F40/216G06F40/279G06F40/284G06F40/30
Inventor 熊盛武陈振东段鹏飞缪少豪王娜毛晶晶
Owner WUHAN UNIV OF TECH
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap