Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks
BackgroundEvent extraction is essential for natural language processing. In the biomedical field, the nested event phenomenon (event A as a participating role of event B) makes extracting this event more difficult than extracting a single event. Therefore, the performance of...
Main Authors: | , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
JMIR Publications
2022-06-01
|
Series: | JMIR Medical Informatics |
Online Access: | https://medinform.jmir.org/2022/6/e37804 |
_version_ | 1797734975581192192 |
---|---|
author | Yan Wang Jian Wang Huiyi Lu Bing Xu Yijia Zhang Santosh Kumar Banbhrani Hongfei Lin |
author_facet | Yan Wang Jian Wang Huiyi Lu Bing Xu Yijia Zhang Santosh Kumar Banbhrani Hongfei Lin |
author_sort | Yan Wang |
collection | DOAJ |
description |
BackgroundEvent extraction is essential for natural language processing. In the biomedical field, the nested event phenomenon (event A as a participating role of event B) makes extracting this event more difficult than extracting a single event. Therefore, the performance of nested biomedical events is always underwhelming. In addition, previous works relied on a pipeline to build an event extraction model, which ignored the dependence between trigger recognition and event argument detection tasks and produced significant cascading errors.
ObjectiveThis study aims to design a unified framework to jointly train biomedical event triggers and arguments and improve the performance of extracting nested biomedical events.
MethodsWe proposed an end-to-end joint extraction model that considers the probability distribution of triggers to alleviate cascading errors. Moreover, we integrated the syntactic structure into an attention-based gate graph convolutional network to capture potential interrelations between triggers and related entities, which improved the performance of extracting nested biomedical events.
ResultsThe experimental results demonstrated that our proposed method achieved the best F1 score on the multilevel event extraction biomedical event extraction corpus and achieved a favorable performance on the biomedical natural language processing shared task 2011 Genia event corpus.
ConclusionsOur conditional probability joint extraction model is good at extracting nested biomedical events because of the joint extraction mechanism and the syntax graph structure. Moreover, as our model did not rely on external knowledge and specific feature engineering, it had a particular generalization performance. |
first_indexed | 2024-03-12T12:52:16Z |
format | Article |
id | doaj.art-2cb74b5058a945e2a3b28914a550f127 |
institution | Directory Open Access Journal |
issn | 2291-9694 |
language | English |
last_indexed | 2024-03-12T12:52:16Z |
publishDate | 2022-06-01 |
publisher | JMIR Publications |
record_format | Article |
series | JMIR Medical Informatics |
spelling | doaj.art-2cb74b5058a945e2a3b28914a550f1272023-08-28T22:14:13ZengJMIR PublicationsJMIR Medical Informatics2291-96942022-06-01106e3780410.2196/37804Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural NetworksYan Wanghttps://orcid.org/0000-0003-1036-9365Jian Wanghttps://orcid.org/0000-0002-4656-7446Huiyi Luhttps://orcid.org/0000-0003-4384-4636Bing Xuhttps://orcid.org/0000-0002-1217-4169Yijia Zhanghttps://orcid.org/0000-0002-5843-4675Santosh Kumar Banbhranihttps://orcid.org/0000-0003-2573-9162Hongfei Linhttps://orcid.org/0000-0003-0872-7688 BackgroundEvent extraction is essential for natural language processing. In the biomedical field, the nested event phenomenon (event A as a participating role of event B) makes extracting this event more difficult than extracting a single event. Therefore, the performance of nested biomedical events is always underwhelming. In addition, previous works relied on a pipeline to build an event extraction model, which ignored the dependence between trigger recognition and event argument detection tasks and produced significant cascading errors. ObjectiveThis study aims to design a unified framework to jointly train biomedical event triggers and arguments and improve the performance of extracting nested biomedical events. MethodsWe proposed an end-to-end joint extraction model that considers the probability distribution of triggers to alleviate cascading errors. Moreover, we integrated the syntactic structure into an attention-based gate graph convolutional network to capture potential interrelations between triggers and related entities, which improved the performance of extracting nested biomedical events. ResultsThe experimental results demonstrated that our proposed method achieved the best F1 score on the multilevel event extraction biomedical event extraction corpus and achieved a favorable performance on the biomedical natural language processing shared task 2011 Genia event corpus. ConclusionsOur conditional probability joint extraction model is good at extracting nested biomedical events because of the joint extraction mechanism and the syntax graph structure. Moreover, as our model did not rely on external knowledge and specific feature engineering, it had a particular generalization performance.https://medinform.jmir.org/2022/6/e37804 |
spellingShingle | Yan Wang Jian Wang Huiyi Lu Bing Xu Yijia Zhang Santosh Kumar Banbhrani Hongfei Lin Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks JMIR Medical Informatics |
title | Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks |
title_full | Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks |
title_fullStr | Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks |
title_full_unstemmed | Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks |
title_short | Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks |
title_sort | conditional probability joint extraction of nested biomedical events design of a unified extraction framework based on neural networks |
url | https://medinform.jmir.org/2022/6/e37804 |
work_keys_str_mv | AT yanwang conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT jianwang conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT huiyilu conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT bingxu conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT yijiazhang conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT santoshkumarbanbhrani conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks AT hongfeilin conditionalprobabilityjointextractionofnestedbiomedicaleventsdesignofaunifiedextractionframeworkbasedonneuralnetworks |