WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments

Skye  Morgan; Tharindu Ranasinghe; Marcos Zampieri

doi:10.48550/arXiv.2108.00057

WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments

Skye Morgan, Tharindu Ranasinghe, Marcos Zampieri

Research output: Chapter in Book/Published conference output › Conference publication

Abstract

This paper addresses the identification of toxic, engaging, and fact-claiming comments on social media. We used the dataset made available by the organizers of the GermEval-2021 shared task containing over 3,000 manually annotated Facebook comments in German. Considering the relatedness of the three tasks, we approached the problem using large pre-trained transformer models and multitask learning. Our results indicate that multitask learning achieves performance superior to the more common single task learning approach in all three tasks. We submit our best systems to GermEval-2021 under the team name WLV-RIT.

Original language	English
Title of host publication	Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments
Pages	32-38
DOIs	https://doi.org/10.48550/arXiv.2108.00057
Publication status	Published - Sept 2021

Bibliographical note

This is an accepted manuscript of a paper published in Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. This accepted manuscript is distributed under the terms of the Creative Commons Attribution License CC BY [https://creativecommons.org/licenses/by/4.0/], which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Access to Document

10.48550/arXiv.2108.00057Licence: CC BY 4.0

Morgan et al_arXiv_AAM_2108.00057
This is an accepted manuscript of a paper published in Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. This accepted manuscript is distributed under the terms of the Creative Commons Attribution License CC BY [https://creativecommons.org/licenses/by/4.0/], which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Accepted author manuscript, 264 KBLicence: CC BY 4.0

Cite this

@inproceedings{e5309813fad545489a7ced26f2fee413,

title = "WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments",

abstract = "This paper addresses the identification of toxic, engaging, and fact-claiming comments on social media. We used the dataset made available by the organizers of the GermEval-2021 shared task containing over 3,000 manually annotated Facebook comments in German. Considering the relatedness of the three tasks, we approached the problem using large pre-trained transformer models and multitask learning. Our results indicate that multitask learning achieves performance superior to the more common single task learning approach in all three tasks. We submit our best systems to GermEval-2021 under the team name WLV-RIT.",

author = "Skye Morgan and Tharindu Ranasinghe and Marcos Zampieri",

note = "This is an accepted manuscript of a paper published in Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. This accepted manuscript is distributed under the terms of the Creative Commons Attribution License CC BY [https://creativecommons.org/licenses/by/4.0/], which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.",

year = "2021",

month = sep,

doi = "10.48550/arXiv.2108.00057",

language = "English",

pages = "32--38",

booktitle = "Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments",

}

WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments. / Morgan, Skye ; Ranasinghe, Tharindu; Zampieri, Marcos.
Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. 2021. p. 32-38.

Research output: Chapter in Book/Published conference output › Conference publication

TY - GEN

T1 - WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments

AU - Morgan, Skye

AU - Ranasinghe, Tharindu

AU - Zampieri, Marcos

N1 - This is an accepted manuscript of a paper published in Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments. This accepted manuscript is distributed under the terms of the Creative Commons Attribution License CC BY [https://creativecommons.org/licenses/by/4.0/], which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PY - 2021/9

Y1 - 2021/9

N2 - This paper addresses the identification of toxic, engaging, and fact-claiming comments on social media. We used the dataset made available by the organizers of the GermEval-2021 shared task containing over 3,000 manually annotated Facebook comments in German. Considering the relatedness of the three tasks, we approached the problem using large pre-trained transformer models and multitask learning. Our results indicate that multitask learning achieves performance superior to the more common single task learning approach in all three tasks. We submit our best systems to GermEval-2021 under the team name WLV-RIT.

AB - This paper addresses the identification of toxic, engaging, and fact-claiming comments on social media. We used the dataset made available by the organizers of the GermEval-2021 shared task containing over 3,000 manually annotated Facebook comments in German. Considering the relatedness of the three tasks, we approached the problem using large pre-trained transformer models and multitask learning. Our results indicate that multitask learning achieves performance superior to the more common single task learning approach in all three tasks. We submit our best systems to GermEval-2021 under the team name WLV-RIT.

UR - https://aclanthology.org/2021.germeval-1.5

U2 - 10.48550/arXiv.2108.00057

DO - 10.48550/arXiv.2108.00057

M3 - Conference publication

SP - 32

EP - 38

BT - Proceedings of the GermEval 2021 Shared Task on the Identification of Toxic, Engaging, and Fact-Claiming Comments

ER -

WLV-RIT at GermEval 2021: Multitask Learning with Transformers to Detect Toxic, Engaging, and Fact-Claiming Comments

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this