Link prediction method and system for knowledge graph

A knowledge map and link prediction technology, applied in neural learning methods, unstructured text data retrieval, biological neural network models, etc., can solve the problem of limited interaction capture ability, low link prediction accuracy, and inability of ConvE to interact, etc. question

Active Publication Date: 2021-07-30
QILU UNIV OF TECH
View PDF9 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

They aim to increase the number of interactions between embedded components, but ConvE cannot maximize the interaction between entities and relations
[0006] In summary, the inventors found that in the field of link prediction in knowl...

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
  • Link prediction method and system for knowledge graph
  • Link prediction method and system for knowledge graph
  • Link prediction method and system for knowledge graph

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0037] Such as figure 1 and figure 2 As shown, this embodiment provides a link prediction method for a knowledge graph, which specifically includes the following steps:

[0038] Step S101: Obtain entity vectors and relational embedding vectors in the knowledge graph.

[0039] The data of the method in this embodiment are taken from three knowledge map data sets applied to FB15K-237, WN18RR and YAGO3-10. The entity vector and relation embedding vector in the knowledge graph are e s and e r .

[0040] Step S102: generating a random permutation of the entity vectors and relation embedding vectors.

[0041] The order of input is not fixed, but multiple permutations are generated by random permutation to capture more interactions.

[0042] First generate e s and e r A random permutation of , we denote as In most cases, for different i The included interaction sets are disjoint. Through the knowledge of permutations and combinations, we can know that in all permutation...

Embodiment 2

[0086] This embodiment provides a link prediction system for knowledge graphs, which specifically includes:

[0087] A vector acquisition module, which is used to acquire entity vectors and relational embedding vectors in the knowledge graph;

[0088] a random permutation module for generating a random permutation of the entity vectors and relational embedding vectors;

[0089] A reshaping module, which is used to reshape randomly arranged entity vectors and relational embedding vectors into a matrix form;

[0090] A batch normalization processing module, which is used for batch normalization processing the reshaped entity vector and relational embedding vector in the matrix;

[0091] Link prediction module, which is used to convolve the normalized entity vector and relationship embedding vector using circular convolution, and the convolution vectorization output is fed back to the fully connected layer to obtain the entity embedding matrix and predict the knowledge map link...

Embodiment 3

[0094] This embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps in the above-mentioned link prediction method for a knowledge graph are implemented.

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 belongs to the field of knowledge graph link prediction, and provides a link prediction method and system for a knowledge graph. The method comprises the following steps: acquiring an entity vector and a relation embedding vector in a knowledge graph; generating a random arrangement of the entity vector and the relation embedding vector; shaping the randomly arranged entity vectors and relation embedding vectors into a matrix form; carrying out batch normalization processing on the shaped entity vectors and relation embedding vectors in the matrix; and carrying out convolution on the entity vector and the relation embedding vector after normalization processing by using cyclic convolution, outputting convolution vectorization and feeding back to a full connection layer to obtain an entity embedding matrix, and predicting a link of the knowledge graph.

Description

technical field [0001] The invention belongs to the field of link prediction of knowledge graphs, in particular to a method and system for link prediction of knowledge graphs. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] Knowledge graphs (KGs) are graph-structured knowledge bases, where nodes represent entities and edges represent relationships between entities. These relations are represented in the form of (s,r,o) triples (eg: entity s=Yao Ming, relation r=nationality, entity o=China). KGs are used in many aspects, such as information retrieval, natural language understanding, question answering system, recommendation system, etc. [0004] Although knowledge graphs contain millions of entities and triples, they are far from complete compared with existing facts and newly added knowledge in the real world. Therefore, the completion ...

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): G06F16/36G06N3/04G06N3/08
CPCG06F16/367G06N3/08G06N3/045
Inventor 李爱民李稼川刘腾刘笑含
Owner QILU 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