Remote supervision-oriented double-encoder and hierarchical relation information relation extraction method

The invention discloses a remote supervision-oriented double-encoder and hierarchical relation information relation extraction method. According to the method, sentence embedding, knowledge graph embedding and hierarchical structure relation embedding are firstly obtained, then the sentence embeddin...

Full description

Saved in:
Bibliographic Details
Main Authors CHEN JIANXIA, ZHANG CHENGLIN, CHENG YU, KIM, WUK, ZHANG MAOHUAN, WANG DEFA
Format Patent
LanguageChinese
English
Published 23.07.2024
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a remote supervision-oriented double-encoder and hierarchical relation information relation extraction method. According to the method, sentence embedding, knowledge graph embedding and hierarchical structure relation embedding are firstly obtained, then the sentence embedding, the knowledge graph embedding and the hierarchical structure relation embedding are fused together to obtain enhanced sentence representation, and the relation of entity pairs in sentences is obtained through the enhanced sentence representation. According to the method, embedding of each sentence is enhanced by correct relation embedding through constraint between hierarchies, so that the model achieves better performance, and finally, a classifier is used for classifying enhanced sentence representation, so that the relation of entity pairs in the sentences is obtained. According to the method, the complementary capability of the text encoder and the knowledge graph encoder in remote supervision relation extra
Bibliography:Application Number: CN202410468042