COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts

Yuhang Jiang; Ramakanth Kavuluru

doi:10.3233/SHTI231050

COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts

Yuhang Jiang, Ramakanth Kavuluru

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

As COVID-19 ravages the world, social media analytics could augment traditional surveys in assessing how the pandemic evolves and capturing consumer chatter that could help healthcare agencies in addressing it. This typically involves mining disclosure events that mention testing positive for the disease or discussions surrounding perceptions and beliefs in preventative or treatment options. The 2020 shared task on COVID-19 event extraction (conducted as part of the W-NUT workshop during the EMNLP conference) introduced a new Twitter dataset for benchmarking event extraction from COVID-19 tweets. In this paper, we cast the problem of event extraction as extractive question answering using recent advances in continuous prompting in language models. On the shared task test dataset, our approach leads to over 5% absolute micro-averaged F1-score improvement over prior best results, across all COVID-19 event slots. Our ablation study shows that continuous prompts have a major impact on the eventual performance.

Original language	English
Title of host publication	MEDINFO 2023 - The Future is Accessible
Subtitle of host publication	Proceedings of the 19th World Congress on Medical and Health Informatics
Editors	Jen Bichel-Findlay, Paula Otero, Philip Scott, Elaine Huesing
Publisher	IOS Press BV
Pages	674-678
Number of pages	5
ISBN (Electronic)	9781643684567
DOIs	https://doi.org/10.3233/SHTI231050
State	Published - Jan 25 2024
Event	19th World Congress on Medical and Health Informatics, MedInfo 2023 - Sydney, Australia Duration: Jul 8 2023 → Jul 12 2023

Publication series

Name	Studies in Health Technology and Informatics
Volume	310
ISSN (Print)	0926-9630
ISSN (Electronic)	1879-8365

Conference

Conference	19th World Congress on Medical and Health Informatics, MedInfo 2023
Country/Territory	Australia
City	Sydney
Period	7/8/23 → 7/12/23

Bibliographical note

Publisher Copyright:
© 2024 International Medical Informatics Association (IMIA) and IOS Press.

Keywords

COVID-19
event extraction
question answering
social media mining

ASJC Scopus subject areas

Biomedical Engineering
Health Informatics
Health Information Management

Access to Document

10.3233/SHTI231050

Cite this

Jiang, Y., & Kavuluru, R. (2024). COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts. In J. Bichel-Findlay, P. Otero, P. Scott, & E. Huesing (Eds.), MEDINFO 2023 - The Future is Accessible: Proceedings of the 19th World Congress on Medical and Health Informatics (pp. 674-678). (Studies in Health Technology and Informatics; Vol. 310). IOS Press BV. https://doi.org/10.3233/SHTI231050

Jiang, Yuhang ; Kavuluru, Ramakanth. / COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts. MEDINFO 2023 - The Future is Accessible: Proceedings of the 19th World Congress on Medical and Health Informatics. editor / Jen Bichel-Findlay ; Paula Otero ; Philip Scott ; Elaine Huesing. IOS Press BV, 2024. pp. 674-678 (Studies in Health Technology and Informatics).

@inproceedings{13da687cade54f74b5ee1f1adf9c3d1e,

title = "COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts",

abstract = "As COVID-19 ravages the world, social media analytics could augment traditional surveys in assessing how the pandemic evolves and capturing consumer chatter that could help healthcare agencies in addressing it. This typically involves mining disclosure events that mention testing positive for the disease or discussions surrounding perceptions and beliefs in preventative or treatment options. The 2020 shared task on COVID-19 event extraction (conducted as part of the W-NUT workshop during the EMNLP conference) introduced a new Twitter dataset for benchmarking event extraction from COVID-19 tweets. In this paper, we cast the problem of event extraction as extractive question answering using recent advances in continuous prompting in language models. On the shared task test dataset, our approach leads to over 5% absolute micro-averaged F1-score improvement over prior best results, across all COVID-19 event slots. Our ablation study shows that continuous prompts have a major impact on the eventual performance.",

keywords = "COVID-19, event extraction, question answering, social media mining",

author = "Yuhang Jiang and Ramakanth Kavuluru",

note = "Publisher Copyright: {\textcopyright} 2024 International Medical Informatics Association (IMIA) and IOS Press.; 19th World Congress on Medical and Health Informatics, MedInfo 2023 ; Conference date: 08-07-2023 Through 12-07-2023",

year = "2024",

month = jan,

day = "25",

doi = "10.3233/SHTI231050",

language = "English",

series = "Studies in Health Technology and Informatics",

publisher = "IOS Press BV",

pages = "674--678",

editor = "Jen Bichel-Findlay and Paula Otero and Philip Scott and Elaine Huesing",

booktitle = "MEDINFO 2023 - The Future is Accessible",

address = "Netherlands",

}

Jiang, Y & Kavuluru, R 2024, COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts. in J Bichel-Findlay, P Otero, P Scott & E Huesing (eds), MEDINFO 2023 - The Future is Accessible: Proceedings of the 19th World Congress on Medical and Health Informatics. Studies in Health Technology and Informatics, vol. 310, IOS Press BV, pp. 674-678, 19th World Congress on Medical and Health Informatics, MedInfo 2023, Sydney, Australia, 7/8/23. https://doi.org/10.3233/SHTI231050

COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts. / Jiang, Yuhang; Kavuluru, Ramakanth.
MEDINFO 2023 - The Future is Accessible: Proceedings of the 19th World Congress on Medical and Health Informatics. ed. / Jen Bichel-Findlay; Paula Otero; Philip Scott; Elaine Huesing. IOS Press BV, 2024. p. 674-678 (Studies in Health Technology and Informatics; Vol. 310).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts

AU - Jiang, Yuhang

AU - Kavuluru, Ramakanth

PY - 2024/1/25

Y1 - 2024/1/25

N2 - As COVID-19 ravages the world, social media analytics could augment traditional surveys in assessing how the pandemic evolves and capturing consumer chatter that could help healthcare agencies in addressing it. This typically involves mining disclosure events that mention testing positive for the disease or discussions surrounding perceptions and beliefs in preventative or treatment options. The 2020 shared task on COVID-19 event extraction (conducted as part of the W-NUT workshop during the EMNLP conference) introduced a new Twitter dataset for benchmarking event extraction from COVID-19 tweets. In this paper, we cast the problem of event extraction as extractive question answering using recent advances in continuous prompting in language models. On the shared task test dataset, our approach leads to over 5% absolute micro-averaged F1-score improvement over prior best results, across all COVID-19 event slots. Our ablation study shows that continuous prompts have a major impact on the eventual performance.

AB - As COVID-19 ravages the world, social media analytics could augment traditional surveys in assessing how the pandemic evolves and capturing consumer chatter that could help healthcare agencies in addressing it. This typically involves mining disclosure events that mention testing positive for the disease or discussions surrounding perceptions and beliefs in preventative or treatment options. The 2020 shared task on COVID-19 event extraction (conducted as part of the W-NUT workshop during the EMNLP conference) introduced a new Twitter dataset for benchmarking event extraction from COVID-19 tweets. In this paper, we cast the problem of event extraction as extractive question answering using recent advances in continuous prompting in language models. On the shared task test dataset, our approach leads to over 5% absolute micro-averaged F1-score improvement over prior best results, across all COVID-19 event slots. Our ablation study shows that continuous prompts have a major impact on the eventual performance.

KW - COVID-19

KW - event extraction

KW - question answering

KW - social media mining

UR - http://www.scopus.com/inward/record.url?scp=85183577554&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85183577554&partnerID=8YFLogxK

U2 - 10.3233/SHTI231050

DO - 10.3233/SHTI231050

M3 - Conference contribution

C2 - 38269894

AN - SCOPUS:85183577554

T3 - Studies in Health Technology and Informatics

SP - 674

EP - 678

BT - MEDINFO 2023 - The Future is Accessible

A2 - Bichel-Findlay, Jen

A2 - Otero, Paula

A2 - Scott, Philip

A2 - Huesing, Elaine

PB - IOS Press BV

T2 - 19th World Congress on Medical and Health Informatics, MedInfo 2023

Y2 - 8 July 2023 through 12 July 2023

ER -

Jiang Y, Kavuluru R. COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts. In Bichel-Findlay J, Otero P, Scott P, Huesing E, editors, MEDINFO 2023 - The Future is Accessible: Proceedings of the 19th World Congress on Medical and Health Informatics. IOS Press BV. 2024. p. 674-678. (Studies in Health Technology and Informatics). doi: 10.3233/SHTI231050

COVID-19 Event Extraction from Twitter via Extractive Question Answering with Continuous Prompts

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this