Integrating graph convolutional networks to enhance prompt learning for biomedical relation extraction.

Graph convolutional networks Pre-trained language models Prompt learning Relation extraction

Journal

Journal of biomedical informatics
ISSN: 1532-0480
Titre abrégé: J Biomed Inform
Pays: United States
ID NLM: 100970413

Informations de publication

Date de publication:
28 Aug 2024
Historique:
received: 22 05 2024
revised: 14 08 2024
accepted: 25 08 2024
medline: 31 8 2024
pubmed: 31 8 2024
entrez: 29 8 2024
Statut: aheadofprint

Résumé

Biomedical relation extraction aims to reveal the relation between entities in medical texts. Currently, the relation extraction models that have attracted much attention are mainly to fine-tune the pre-trained language models (PLMs) or add template prompt learning, which also limits the ability of the model to deal with grammatical dependencies. Graph convolutional networks (GCNs) can play an important role in processing syntactic dependencies in biomedical texts. In this work, we propose a biomedical relation extraction model that fuses GCNs enhanced prompt learning to handle limitations in syntactic dependencies and achieve good performance. Specifically, we propose a model that combines prompt learning with GCNs for relation extraction, by integrating the syntactic dependency information analyzed by GCNs into the prompt learning model, by predicting the correspondence with [MASK] tokens labels for relation extraction. Our model achieved F1 scores of 85.57%, 80.15%, 95.10%, and 84.11% in the biomedical relation extraction datasets GAD, ChemProt, PGR, and DDI, respectively, all of which outperform some existing baseline models. In this paper, we propose enhancing prompt learning through GCNs, integrating syntactic information into biomedical relation extraction tasks. Experimental results show that our proposed method achieves excellent performance in the biomedical relation extraction task.

Sections du résumé

BACKGROUND AND OBJECTIVE OBJECTIVE
Biomedical relation extraction aims to reveal the relation between entities in medical texts. Currently, the relation extraction models that have attracted much attention are mainly to fine-tune the pre-trained language models (PLMs) or add template prompt learning, which also limits the ability of the model to deal with grammatical dependencies. Graph convolutional networks (GCNs) can play an important role in processing syntactic dependencies in biomedical texts.
METHODS METHODS
In this work, we propose a biomedical relation extraction model that fuses GCNs enhanced prompt learning to handle limitations in syntactic dependencies and achieve good performance. Specifically, we propose a model that combines prompt learning with GCNs for relation extraction, by integrating the syntactic dependency information analyzed by GCNs into the prompt learning model, by predicting the correspondence with [MASK] tokens labels for relation extraction.
RESULTS RESULTS
Our model achieved F1 scores of 85.57%, 80.15%, 95.10%, and 84.11% in the biomedical relation extraction datasets GAD, ChemProt, PGR, and DDI, respectively, all of which outperform some existing baseline models.
CONCLUSIONS CONCLUSIONS
In this paper, we propose enhancing prompt learning through GCNs, integrating syntactic information into biomedical relation extraction tasks. Experimental results show that our proposed method achieves excellent performance in the biomedical relation extraction task.

Identifiants

pubmed: 39209087
pii: S1532-0464(24)00135-7
doi: 10.1016/j.jbi.2024.104717
pii:
doi:

Types de publication

Journal Article

Langues

eng

Sous-ensembles de citation

IM

Pagination

104717

Informations de copyright

Copyright © 2024 Elsevier Inc. All rights reserved.

Déclaration de conflit d'intérêts

Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Di Zhao reports financial support was provided by Liaoning Provincial Natural Science Foundation. If there are other authors they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Auteurs

Bocheng Guo (B)

School of Computer Science and Engineering, Dalian Minzu University, Dalian, 116650, Liaoning, China.

Jiana Meng (J)

School of Computer Science and Engineering, Dalian Minzu University, Dalian, 116650, Liaoning, China.

Di Zhao (D)

School of Computer Science and Engineering, Dalian Minzu University, Dalian, 116650, Liaoning, China; School of Computer Science and Technology, Dalian University of Technology, Dalian, 116024, Liaoning, China; Postdoctoral workstation of Dalian Yongjia Electronic Technology Co., Ltd, Liaoning, China. Electronic address: zhaodi@dlnu.edu.cn.

Xiangxing Jia (X)

School of Computer Science and Engineering, Dalian Minzu University, Dalian, 116650, Liaoning, China.

Yonghe Chu (Y)

College of Information Science and Engineering, Henan University of Technology, Zhengzhou, 450001, Henan, China.

Hongfei Lin (H)

School of Computer Science and Technology, Dalian University of Technology, Dalian, 116024, Liaoning, China.

Classifications MeSH