Fine-tuned BERT Model for Multi-Label Tweets Classification

Hamada M. Zahera; Ibrahim Elgendy; Rricha Jalota; Mohamed Ahmed Sherif

Fine-tuned BERT Model for Multi-Label Tweets Classification

Hamada M. Zahera^*, Ibrahim Elgendy, Rricha Jalota, Mohamed Ahmed Sherif

^*Corresponding author for this work

Research output: Contribution to conference › Paper › peer-review

29 Scopus citations

Abstract

In this paper, we describe our approach to classify disaster-related tweets into multi-label information types (i.e, labels). We aim to filter first relevant tweets during disasters. Then, we assign tweets relevant information types. Information types can be SearchAndRescue, MovePeople and Volunteer. We employ a fine-tuned BERT model with 10 BERT layers. Further, we submitted our approach to the TREC-IS 2019 challenge, the evaluation results showed that our approach outperforms the F1-score of median score in identifying actionable information.

Original language	English
State	Published - 2019
Externally published	Yes
Event	28th Text REtrieval Conference, TREC 2019 - Gaithersburg, United States Duration: 13 Nov 2019 → 15 Nov 2019

Conference

Conference	28th Text REtrieval Conference, TREC 2019
Country/Territory	United States
City	Gaithersburg
Period	13/11/19 → 15/11/19

Bibliographical note

Publisher Copyright:
© 2019 28th Text REtrieval Conference, TREC 2019 - Proceedings. All Rights Reserved.

ASJC Scopus subject areas

Language and Linguistics
Computer Science Applications
Linguistics and Language

Cite this

@conference{fe3590807c8047329a5a88cba6d7cb6d,

title = "Fine-tuned BERT Model for Multi-Label Tweets Classification",

abstract = "In this paper, we describe our approach to classify disaster-related tweets into multi-label information types (i.e, labels). We aim to filter first relevant tweets during disasters. Then, we assign tweets relevant information types. Information types can be SearchAndRescue, MovePeople and Volunteer. We employ a fine-tuned BERT model with 10 BERT layers. Further, we submitted our approach to the TREC-IS 2019 challenge, the evaluation results showed that our approach outperforms the F1-score of median score in identifying actionable information.",

author = "Zahera, \{Hamada M.\} and Ibrahim Elgendy and Rricha Jalota and Sherif, \{Mohamed Ahmed\}",

note = "Publisher Copyright: {\textcopyright} 2019 28th Text REtrieval Conference, TREC 2019 - Proceedings. All Rights Reserved.; 28th Text REtrieval Conference, TREC 2019 ; Conference date: 13-11-2019 Through 15-11-2019",

year = "2019",

language = "English",

}

TY - CONF

T1 - Fine-tuned BERT Model for Multi-Label Tweets Classification

AU - Zahera, Hamada M.

AU - Elgendy, Ibrahim

AU - Jalota, Rricha

AU - Sherif, Mohamed Ahmed

PY - 2019

Y1 - 2019

N2 - In this paper, we describe our approach to classify disaster-related tweets into multi-label information types (i.e, labels). We aim to filter first relevant tweets during disasters. Then, we assign tweets relevant information types. Information types can be SearchAndRescue, MovePeople and Volunteer. We employ a fine-tuned BERT model with 10 BERT layers. Further, we submitted our approach to the TREC-IS 2019 challenge, the evaluation results showed that our approach outperforms the F1-score of median score in identifying actionable information.

AB - In this paper, we describe our approach to classify disaster-related tweets into multi-label information types (i.e, labels). We aim to filter first relevant tweets during disasters. Then, we assign tweets relevant information types. Information types can be SearchAndRescue, MovePeople and Volunteer. We employ a fine-tuned BERT model with 10 BERT layers. Further, we submitted our approach to the TREC-IS 2019 challenge, the evaluation results showed that our approach outperforms the F1-score of median score in identifying actionable information.

UR - https://www.scopus.com/pages/publications/85180126131

M3 - Paper

AN - SCOPUS:85180126131

T2 - 28th Text REtrieval Conference, TREC 2019

Y2 - 13 November 2019 through 15 November 2019

ER -

Fine-tuned BERT Model for Multi-Label Tweets Classification

Abstract

Conference

Bibliographical note

ASJC Scopus subject areas

Fingerprint

Cite this