A triple joint extraction method combining hybrid embedding and relational label embedding

The purpose of triple extraction is to obtain relationships between entities from unstructured text and apply them to downstream tasks.The embedding mechanism has a great impact on the performance of the triple extraction model, and the embedding vector should contain rich semantic information that...

Full description

Saved in:
Bibliographic Details
Main Authors: Jianfeng DAI, Xingyu CHEN, Ligang DONG, Xian JIANG
Format: Article
Language:zho
Published: Beijing Xintong Media Co., Ltd 2023-02-01
Series:Dianxin kexue
Subjects:
Online Access:http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2023021/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841530800335486976
author Jianfeng DAI
Xingyu CHEN
Ligang DONG
Xian JIANG
author_facet Jianfeng DAI
Xingyu CHEN
Ligang DONG
Xian JIANG
author_sort Jianfeng DAI
collection DOAJ
description The purpose of triple extraction is to obtain relationships between entities from unstructured text and apply them to downstream tasks.The embedding mechanism has a great impact on the performance of the triple extraction model, and the embedding vector should contain rich semantic information that is closely related to the relationship extraction task.In Chinese datasets, the information contained between words is very different, and in order to avoid the loss of semantic information problems generated by word separation errors, a triple joint extraction method combining hybrid embedding and relational label embedding (HEPA) was designed, and a hybrid embedding means that combines letter embedding and word embedding was proposed to reduce the errors generated by word separation errors.A relational embedding mechanism that fuses text and relational labels was added, and an attention mechanism was used to distinguish the relevance of entities in a sentence with different relational labels, thus improving the matching accuracy.The method of matching entities with pointer annotation was used, which improved the extraction effect on relational overlapping triples.Comparative experiments are conducted on the publicly available DuIE dataset, and the F1 value of HEPA is improved by 2.8% compared to the best performing baseline model (CasRel).
format Article
id doaj-art-046ffbf3ab0749c58974ac19ead72ca9
institution Kabale University
issn 1000-0801
language zho
publishDate 2023-02-01
publisher Beijing Xintong Media Co., Ltd
record_format Article
series Dianxin kexue
spelling doaj-art-046ffbf3ab0749c58974ac19ead72ca92025-01-15T02:59:06ZzhoBeijing Xintong Media Co., LtdDianxin kexue1000-08012023-02-013913214459570845A triple joint extraction method combining hybrid embedding and relational label embeddingJianfeng DAIXingyu CHENLigang DONGXian JIANGThe purpose of triple extraction is to obtain relationships between entities from unstructured text and apply them to downstream tasks.The embedding mechanism has a great impact on the performance of the triple extraction model, and the embedding vector should contain rich semantic information that is closely related to the relationship extraction task.In Chinese datasets, the information contained between words is very different, and in order to avoid the loss of semantic information problems generated by word separation errors, a triple joint extraction method combining hybrid embedding and relational label embedding (HEPA) was designed, and a hybrid embedding means that combines letter embedding and word embedding was proposed to reduce the errors generated by word separation errors.A relational embedding mechanism that fuses text and relational labels was added, and an attention mechanism was used to distinguish the relevance of entities in a sentence with different relational labels, thus improving the matching accuracy.The method of matching entities with pointer annotation was used, which improved the extraction effect on relational overlapping triples.Comparative experiments are conducted on the publicly available DuIE dataset, and the F1 value of HEPA is improved by 2.8% compared to the best performing baseline model (CasRel).http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2023021/triple extractionrelational embeddingBERTattention mechanismpointer annotation
spellingShingle Jianfeng DAI
Xingyu CHEN
Ligang DONG
Xian JIANG
A triple joint extraction method combining hybrid embedding and relational label embedding
Dianxin kexue
triple extraction
relational embedding
BERT
attention mechanism
pointer annotation
title A triple joint extraction method combining hybrid embedding and relational label embedding
title_full A triple joint extraction method combining hybrid embedding and relational label embedding
title_fullStr A triple joint extraction method combining hybrid embedding and relational label embedding
title_full_unstemmed A triple joint extraction method combining hybrid embedding and relational label embedding
title_short A triple joint extraction method combining hybrid embedding and relational label embedding
title_sort triple joint extraction method combining hybrid embedding and relational label embedding
topic triple extraction
relational embedding
BERT
attention mechanism
pointer annotation
url http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2023021/
work_keys_str_mv AT jianfengdai atriplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT xingyuchen atriplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT ligangdong atriplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT xianjiang atriplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT jianfengdai triplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT xingyuchen triplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT ligangdong triplejointextractionmethodcombininghybridembeddingandrelationallabelembedding
AT xianjiang triplejointextractionmethodcombininghybridembeddingandrelationallabelembedding