Multi-Level Attention Based BLSTM Neural Network for Biomedical Event Extraction

Xinyu HE; Lishuang LI; Xingchen SONG; Degen HUANG; Fuji REN

doi:10.1587/transinf.2018EDP7268

Multi-Level Attention Based BLSTM Neural Network for Biomedical Event Extraction

Xinyu HE, Lishuang LI, Xingchen SONG, Degen HUANG, Fuji REN

Full Text Views

0

Cite this

Summary :

Biomedical event extraction is an important and challenging task in Information Extraction, which plays a key role for medicine research and disease prevention. Most of the existing event detection methods are based on shallow machine learning methods which mainly rely on domain knowledge and elaborately designed features. Another challenge is that some crucial information as well as the interactions among words or arguments may be ignored since most works treat words and sentences equally. Therefore, we employ a Bidirectional Long Short Term Memory (BLSTM) neural network for event extraction, which can skip handcrafted complex feature extraction. Furthermore, we propose a multi-level attention mechanism, including word level attention which determines the importance of words in a sentence, and the sentence level attention which determines the importance of relevant arguments. Finally, we train dependency word embeddings and add sentence vectors to enrich semantic information. The experimental results show that our model achieves an F-score of 59.61% on the commonly used dataset (MLEE) of biomedical event extraction, which outperforms other state-of-the-art methods.

Publication: IEICE TRANSACTIONS on Information Vol.E102-D No.9 pp.1842-1850

Publication Date: 2019/09/01

Publicized: 2019/04/26

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2018EDP7268

Type of Manuscript: PAPER

Category: Natural Language Processing

Authors

Xinyu HE
  Dalian University of Technology
Lishuang LI
  Dalian University of Technology
Xingchen SONG
  Dalian University of Technology
Degen HUANG
  Dalian University of Technology
Fuji REN
  University of Tokushima

Keyword

event extraction, trigger detection, argument detection, BLSTM neural network, multi-level attention

Cite this

Copy

Xinyu HE, Lishuang LI, Xingchen SONG, Degen HUANG, Fuji REN, "Multi-Level Attention Based BLSTM Neural Network for Biomedical Event Extraction" in IEICE TRANSACTIONS on Information, vol. E102-D, no. 9, pp. 1842-1850, September 2019, doi: 10.1587/transinf.2018EDP7268.
Abstract: Biomedical event extraction is an important and challenging task in Information Extraction, which plays a key role for medicine research and disease prevention. Most of the existing event detection methods are based on shallow machine learning methods which mainly rely on domain knowledge and elaborately designed features. Another challenge is that some crucial information as well as the interactions among words or arguments may be ignored since most works treat words and sentences equally. Therefore, we employ a Bidirectional Long Short Term Memory (BLSTM) neural network for event extraction, which can skip handcrafted complex feature extraction. Furthermore, we propose a multi-level attention mechanism, including word level attention which determines the importance of words in a sentence, and the sentence level attention which determines the importance of relevant arguments. Finally, we train dependency word embeddings and add sentence vectors to enrich semantic information. The experimental results show that our model achieves an F-score of 59.61% on the commonly used dataset (MLEE) of biomedical event extraction, which outperforms other state-of-the-art methods.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7268/_p

Copy

@ARTICLE{e102-d_9_1842,
author={Xinyu HE, Lishuang LI, Xingchen SONG, Degen HUANG, Fuji REN, },
journal={IEICE TRANSACTIONS on Information},
title={Multi-Level Attention Based BLSTM Neural Network for Biomedical Event Extraction},
year={2019},
volume={E102-D},
number={9},
pages={1842-1850},
abstract={Biomedical event extraction is an important and challenging task in Information Extraction, which plays a key role for medicine research and disease prevention. Most of the existing event detection methods are based on shallow machine learning methods which mainly rely on domain knowledge and elaborately designed features. Another challenge is that some crucial information as well as the interactions among words or arguments may be ignored since most works treat words and sentences equally. Therefore, we employ a Bidirectional Long Short Term Memory (BLSTM) neural network for event extraction, which can skip handcrafted complex feature extraction. Furthermore, we propose a multi-level attention mechanism, including word level attention which determines the importance of words in a sentence, and the sentence level attention which determines the importance of relevant arguments. Finally, we train dependency word embeddings and add sentence vectors to enrich semantic information. The experimental results show that our model achieves an F-score of 59.61% on the commonly used dataset (MLEE) of biomedical event extraction, which outperforms other state-of-the-art methods.},
keywords={},
doi={10.1587/transinf.2018EDP7268},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Multi-Level Attention Based BLSTM Neural Network for Biomedical Event Extraction
T2 - IEICE TRANSACTIONS on Information
SP - 1842
EP - 1850
AU - Xinyu HE
AU - Lishuang LI
AU - Xingchen SONG
AU - Degen HUANG
AU - Fuji REN
PY - 2019
DO - 10.1587/transinf.2018EDP7268
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2019
AB - Biomedical event extraction is an important and challenging task in Information Extraction, which plays a key role for medicine research and disease prevention. Most of the existing event detection methods are based on shallow machine learning methods which mainly rely on domain knowledge and elaborately designed features. Another challenge is that some crucial information as well as the interactions among words or arguments may be ignored since most works treat words and sentences equally. Therefore, we employ a Bidirectional Long Short Term Memory (BLSTM) neural network for event extraction, which can skip handcrafted complex feature extraction. Furthermore, we propose a multi-level attention mechanism, including word level attention which determines the importance of words in a sentence, and the sentence level attention which determines the importance of relevant arguments. Finally, we train dependency word embeddings and add sentence vectors to enrich semantic information. The experimental results show that our model achieves an F-score of 59.61% on the commonly used dataset (MLEE) of biomedical event extraction, which outperforms other state-of-the-art methods.
ER -