Extracting chemical-protein relations using attention-based neural networks

Sijia Liu, Feichen Shen, Ravikumar Komandur Elayavilli, Yanshan Wang, Majid Rastegar-Mojarad, Vipin Chaudhary, Hongfang D Liu

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Relation extraction is an important task in the field of natural language processing. In this paper, we describe our approach for the BioCreative VI Task 5: text mining chemical-protein interactions. We investigate multiple deep neural network (DNN) models, including convolutional neural networks, recurrent neural networks (RNNs) and attention-based (ATT-) RNNs (ATT-RNNs) to extract chemical-protein relations. Our experimental results indicate that ATT-RNN models outperform the same models without using attention and the ATT-gated recurrent unit (ATT-GRU) achieves the best performing micro average F1 score of 0.527 on the test set among the tested DNNs. In addition, the result of word-level attention weights also shows that attention mechanism is effective on selecting the most important trigger words when trained with semantic relation labels without the need of semantic parsing and feature engineering. The source code of this work is available at https://github.com/ohnlp/att-chemprot.

Original languageEnglish (US)
JournalDatabase : the journal of biological databases and curation
Volume2018
DOIs
StatePublished - Jan 1 2018

Fingerprint

Neural Networks (Computer)
Recurrent neural networks
Semantics
neural networks
Natural Language Processing
Neural networks
Proteins
Data Mining
proteins
Weights and Measures
Labels
Processing
engineering
extracts
testing

ASJC Scopus subject areas

  • Information Systems
  • Biochemistry, Genetics and Molecular Biology(all)
  • Agricultural and Biological Sciences(all)

Cite this

Extracting chemical-protein relations using attention-based neural networks. / Liu, Sijia; Shen, Feichen; Komandur Elayavilli, Ravikumar; Wang, Yanshan; Rastegar-Mojarad, Majid; Chaudhary, Vipin; Liu, Hongfang D.

In: Database : the journal of biological databases and curation, Vol. 2018, 01.01.2018.

Research output: Contribution to journalArticle

Liu, Sijia ; Shen, Feichen ; Komandur Elayavilli, Ravikumar ; Wang, Yanshan ; Rastegar-Mojarad, Majid ; Chaudhary, Vipin ; Liu, Hongfang D. / Extracting chemical-protein relations using attention-based neural networks. In: Database : the journal of biological databases and curation. 2018 ; Vol. 2018.
@article{2c81e43a2a8f4621842c329ad8d4c7fb,
title = "Extracting chemical-protein relations using attention-based neural networks",
abstract = "Relation extraction is an important task in the field of natural language processing. In this paper, we describe our approach for the BioCreative VI Task 5: text mining chemical-protein interactions. We investigate multiple deep neural network (DNN) models, including convolutional neural networks, recurrent neural networks (RNNs) and attention-based (ATT-) RNNs (ATT-RNNs) to extract chemical-protein relations. Our experimental results indicate that ATT-RNN models outperform the same models without using attention and the ATT-gated recurrent unit (ATT-GRU) achieves the best performing micro average F1 score of 0.527 on the test set among the tested DNNs. In addition, the result of word-level attention weights also shows that attention mechanism is effective on selecting the most important trigger words when trained with semantic relation labels without the need of semantic parsing and feature engineering. The source code of this work is available at https://github.com/ohnlp/att-chemprot.",
author = "Sijia Liu and Feichen Shen and {Komandur Elayavilli}, Ravikumar and Yanshan Wang and Majid Rastegar-Mojarad and Vipin Chaudhary and Liu, {Hongfang D}",
year = "2018",
month = "1",
day = "1",
doi = "10.1093/database/bay102",
language = "English (US)",
volume = "2018",
journal = "Database : the journal of biological databases and curation",
issn = "1758-0463",
publisher = "Oxford University Press",

}

TY - JOUR

T1 - Extracting chemical-protein relations using attention-based neural networks

AU - Liu, Sijia

AU - Shen, Feichen

AU - Komandur Elayavilli, Ravikumar

AU - Wang, Yanshan

AU - Rastegar-Mojarad, Majid

AU - Chaudhary, Vipin

AU - Liu, Hongfang D

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Relation extraction is an important task in the field of natural language processing. In this paper, we describe our approach for the BioCreative VI Task 5: text mining chemical-protein interactions. We investigate multiple deep neural network (DNN) models, including convolutional neural networks, recurrent neural networks (RNNs) and attention-based (ATT-) RNNs (ATT-RNNs) to extract chemical-protein relations. Our experimental results indicate that ATT-RNN models outperform the same models without using attention and the ATT-gated recurrent unit (ATT-GRU) achieves the best performing micro average F1 score of 0.527 on the test set among the tested DNNs. In addition, the result of word-level attention weights also shows that attention mechanism is effective on selecting the most important trigger words when trained with semantic relation labels without the need of semantic parsing and feature engineering. The source code of this work is available at https://github.com/ohnlp/att-chemprot.

AB - Relation extraction is an important task in the field of natural language processing. In this paper, we describe our approach for the BioCreative VI Task 5: text mining chemical-protein interactions. We investigate multiple deep neural network (DNN) models, including convolutional neural networks, recurrent neural networks (RNNs) and attention-based (ATT-) RNNs (ATT-RNNs) to extract chemical-protein relations. Our experimental results indicate that ATT-RNN models outperform the same models without using attention and the ATT-gated recurrent unit (ATT-GRU) achieves the best performing micro average F1 score of 0.527 on the test set among the tested DNNs. In addition, the result of word-level attention weights also shows that attention mechanism is effective on selecting the most important trigger words when trained with semantic relation labels without the need of semantic parsing and feature engineering. The source code of this work is available at https://github.com/ohnlp/att-chemprot.

UR - http://www.scopus.com/inward/record.url?scp=85054456100&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054456100&partnerID=8YFLogxK

U2 - 10.1093/database/bay102

DO - 10.1093/database/bay102

M3 - Article

VL - 2018

JO - Database : the journal of biological databases and curation

JF - Database : the journal of biological databases and curation

SN - 1758-0463

ER -