Denoising Graph Inference Network for Document-Level Relation Extraction
Relation Extraction (RE) is to obtain a predefined relation type of two entities mentioned in a piece of text, e.g., a sentence-level or a document-level text. Most existing studies suffer from the noise in the text, and necessary pruning is of great importance. The conventional sentence-level RE ta...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Tsinghua University Press
2023-06-01
|
Series: | Big Data Mining and Analytics |
Subjects: | |
Online Access: | https://www.sciopen.com/article/10.26599/BDMA.2022.9020051 |
_version_ | 1797902876033417216 |
---|---|
author | Hailin Wang Ke Qin Guiduo Duan Guangchun Luo |
author_facet | Hailin Wang Ke Qin Guiduo Duan Guangchun Luo |
author_sort | Hailin Wang |
collection | DOAJ |
description | Relation Extraction (RE) is to obtain a predefined relation type of two entities mentioned in a piece of text, e.g., a sentence-level or a document-level text. Most existing studies suffer from the noise in the text, and necessary pruning is of great importance. The conventional sentence-level RE task addresses this issue by a denoising method using the shortest dependency path to build a long-range semantic dependency between entity pairs. However, this kind of denoising method is scarce in document-level RE. In this work, we explicitly model a denoised document-level graph based on linguistic knowledge to capture various long-range semantic dependencies among entities. We first formalize a Syntactic Dependency Tree forest (SDT-forest) by introducing the syntax and discourse dependency relation. Then, the Steiner tree algorithm extracts a mention-level denoised graph, Steiner Graph (SG), removing linguistically irrelevant words from the SDT-forest. We then devise a slide residual attention to highlight word-level evidence on text and SG. Finally, the classification is established on the SG to infer the relations of entity pairs. We conduct extensive experiments on three public datasets. The results evidence that our method is beneficial to establish long-range semantic dependency and can improve the classification performance with longer texts. |
first_indexed | 2024-04-10T09:24:10Z |
format | Article |
id | doaj.art-40d4b8484f354074b650b0adcab82842 |
institution | Directory Open Access Journal |
issn | 2096-0654 |
language | English |
last_indexed | 2024-04-10T09:24:10Z |
publishDate | 2023-06-01 |
publisher | Tsinghua University Press |
record_format | Article |
series | Big Data Mining and Analytics |
spelling | doaj.art-40d4b8484f354074b650b0adcab828422023-02-20T07:01:55ZengTsinghua University PressBig Data Mining and Analytics2096-06542023-06-016224826210.26599/BDMA.2022.9020051Denoising Graph Inference Network for Document-Level Relation ExtractionHailin Wang0Ke Qin1Guiduo Duan2Guangchun Luo3School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaSchool of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaRelation Extraction (RE) is to obtain a predefined relation type of two entities mentioned in a piece of text, e.g., a sentence-level or a document-level text. Most existing studies suffer from the noise in the text, and necessary pruning is of great importance. The conventional sentence-level RE task addresses this issue by a denoising method using the shortest dependency path to build a long-range semantic dependency between entity pairs. However, this kind of denoising method is scarce in document-level RE. In this work, we explicitly model a denoised document-level graph based on linguistic knowledge to capture various long-range semantic dependencies among entities. We first formalize a Syntactic Dependency Tree forest (SDT-forest) by introducing the syntax and discourse dependency relation. Then, the Steiner tree algorithm extracts a mention-level denoised graph, Steiner Graph (SG), removing linguistically irrelevant words from the SDT-forest. We then devise a slide residual attention to highlight word-level evidence on text and SG. Finally, the classification is established on the SG to infer the relations of entity pairs. We conduct extensive experiments on three public datasets. The results evidence that our method is beneficial to establish long-range semantic dependency and can improve the classification performance with longer texts.https://www.sciopen.com/article/10.26599/BDMA.2022.9020051relation eextraction (re)document-leveldenoisinglinguistic knowledgeattention mechanism |
spellingShingle | Hailin Wang Ke Qin Guiduo Duan Guangchun Luo Denoising Graph Inference Network for Document-Level Relation Extraction Big Data Mining and Analytics relation eextraction (re) document-level denoising linguistic knowledge attention mechanism |
title | Denoising Graph Inference Network for Document-Level Relation Extraction |
title_full | Denoising Graph Inference Network for Document-Level Relation Extraction |
title_fullStr | Denoising Graph Inference Network for Document-Level Relation Extraction |
title_full_unstemmed | Denoising Graph Inference Network for Document-Level Relation Extraction |
title_short | Denoising Graph Inference Network for Document-Level Relation Extraction |
title_sort | denoising graph inference network for document level relation extraction |
topic | relation eextraction (re) document-level denoising linguistic knowledge attention mechanism |
url | https://www.sciopen.com/article/10.26599/BDMA.2022.9020051 |
work_keys_str_mv | AT hailinwang denoisinggraphinferencenetworkfordocumentlevelrelationextraction AT keqin denoisinggraphinferencenetworkfordocumentlevelrelationextraction AT guiduoduan denoisinggraphinferencenetworkfordocumentlevelrelationextraction AT guangchunluo denoisinggraphinferencenetworkfordocumentlevelrelationextraction |