Biomedical Event Extraction as Sequence Labeling

Alan Ramponi; Rob van der Goot; Rosario Lombardo; Barbara Plank

Biomedical Event Extraction as Sequence Labeling

Alan Ramponi, Rob van der Goot, Rosario Lombardo, Barbara Plank

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

Abstract

We introduce Biomedical Event Extraction as Sequence Labeling (BeeSL), a joint end-to-end neural information extraction model. BeeSL recasts the task as sequence labeling, taking advantage of a multi-label aware encoding strategy and jointly modeling the intermediate tasks via multi-task learning. BeeSL is fast, accurate, end-to-end, and unlike current methods does not require any external knowledge base or preprocessing tools. BeeSL outperforms the current best system on the Genia 2011 benchmark by 1.57 % absolute F1 score reaching 60.22 % F1, establishing a new state of the art for the task.
Importantly, we also provide first results on biomedical event extraction without gold entity information. Empirical results show that BeeSL's speed and accuracy makes it a viable approach for large-scale real-world scenarios.

Original language	English
Title of host publication	Conference on Empirical Methods in Natural Language Processing
Publisher	Association for Computational Linguistics
Publication date	Nov 2020
Publication status	Published - Nov 2020

Keywords

Biomedical Event Extraction
Sequence Labeling
Multi-task Learning
End-to-End Neural Model
Genia 2011 Benchmark

Access to Document

EMNLP2020__BEESLAccepted author manuscript, 457 KB

https://www.aclweb.org/anthology/2020.emnlp-main.431/

Multi-Task Sequence Labeling Under Adverse Conditions
Plank, B. (PI) & van der Goot, R. (CoI)
Amazon.com, Inc.
01/04/2019 → 31/08/2020
Project: Other

Cite this

@inproceedings{b8a97a71d16d4aa5ad539c2dbc292346,

title = "Biomedical Event Extraction as Sequence Labeling",

abstract = "We introduce Biomedical Event Extraction as Sequence Labeling (BeeSL), a joint end-to-end neural information extraction model. BeeSL recasts the task as sequence labeling, taking advantage of a multi-label aware encoding strategy and jointly modeling the intermediate tasks via multi-task learning. BeeSL is fast, accurate, end-to-end, and unlike current methods does not require any external knowledge base or preprocessing tools. BeeSL outperforms the current best system on the Genia 2011 benchmark by 1.57 % absolute F1 score reaching 60.22 % F1, establishing a new state of the art for the task. Importantly, we also provide first results on biomedical event extraction without gold entity information. Empirical results show that BeeSL's speed and accuracy makes it a viable approach for large-scale real-world scenarios.",

keywords = "Biomedical Event Extraction, Sequence Labeling, Multi-task Learning, End-to-End Neural Model, Genia 2011 Benchmark, Biomedical Event Extraction, Sequence Labeling, Multi-task Learning, End-to-End Neural Model, Genia 2011 Benchmark",

author = "Alan Ramponi and {van der Goot}, Rob and Rosario Lombardo and Barbara Plank",

year = "2020",

month = nov,

language = "English",

booktitle = "Conference on Empirical Methods in Natural Language Processing",

publisher = "Association for Computational Linguistics",

address = "United States",

}

TY - GEN

T1 - Biomedical Event Extraction as Sequence Labeling

AU - Ramponi, Alan

AU - van der Goot, Rob

AU - Lombardo, Rosario

AU - Plank, Barbara

PY - 2020/11

Y1 - 2020/11

N2 - We introduce Biomedical Event Extraction as Sequence Labeling (BeeSL), a joint end-to-end neural information extraction model. BeeSL recasts the task as sequence labeling, taking advantage of a multi-label aware encoding strategy and jointly modeling the intermediate tasks via multi-task learning. BeeSL is fast, accurate, end-to-end, and unlike current methods does not require any external knowledge base or preprocessing tools. BeeSL outperforms the current best system on the Genia 2011 benchmark by 1.57 % absolute F1 score reaching 60.22 % F1, establishing a new state of the art for the task. Importantly, we also provide first results on biomedical event extraction without gold entity information. Empirical results show that BeeSL's speed and accuracy makes it a viable approach for large-scale real-world scenarios.

AB - We introduce Biomedical Event Extraction as Sequence Labeling (BeeSL), a joint end-to-end neural information extraction model. BeeSL recasts the task as sequence labeling, taking advantage of a multi-label aware encoding strategy and jointly modeling the intermediate tasks via multi-task learning. BeeSL is fast, accurate, end-to-end, and unlike current methods does not require any external knowledge base or preprocessing tools. BeeSL outperforms the current best system on the Genia 2011 benchmark by 1.57 % absolute F1 score reaching 60.22 % F1, establishing a new state of the art for the task. Importantly, we also provide first results on biomedical event extraction without gold entity information. Empirical results show that BeeSL's speed and accuracy makes it a viable approach for large-scale real-world scenarios.

KW - Biomedical Event Extraction

KW - Sequence Labeling

KW - Multi-task Learning

KW - End-to-End Neural Model

KW - Genia 2011 Benchmark

KW - Biomedical Event Extraction

KW - Sequence Labeling

KW - Multi-task Learning

KW - End-to-End Neural Model

KW - Genia 2011 Benchmark

M3 - Article in proceedings

BT - Conference on Empirical Methods in Natural Language Processing

PB - Association for Computational Linguistics

ER -

Biomedical Event Extraction as Sequence Labeling

Abstract

Keywords

Access to Document

Fingerprint

Projects

Multi-Task Sequence Labeling Under Adverse Conditions

Cite this