Multitask Learning for Arabic Offensive Language and Hate-Speech Detection

Ibrahim Abu Farha; Walid Magdy

Multitask Learning for Arabic Offensive Language and Hate-Speech Detection

Ibrahim Abu Farha, Walid Magdy

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

Abstract / Description of output

Offensive language and hate-speech are phenomena that spread with the rising popularity of social media. Detecting such content is crucial for understanding and predicting conflicts, understanding polarisation among communities and providing means and tools to filter or block inappropriate content. This paper describes the SMASH team submission to OSACT4’s shared task on hate-speech and offensive language detection, where we explore different approaches to perform these tasks. The experiments cover a variety of approaches that include deep learning, transfer learning and multitask learning. We also explore the utilisation of sentiment information to perform the previous task. Our best model is a multitask learning architecture, based on CNN-BiLSTM, that was trained to detect hate-speech and offensive language and predict sentiment.

Original language	English
Title of host publication	Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools
Publisher	European Language Resources Association (ELRA)
Chapter	90
Pages	86
Number of pages	5
ISBN (Electronic)	979-10-95546-51-1
Publication status	Published - 12 May 2020
Event	The 4th Workshop on Open-Source Arabic Corpora and Processing Tools - Marseille, France Duration: 12 May 2020 → 12 May 2020 http://edinburghnlp.inf.ed.ac.uk/workshops/OSACT4/

Workshop

Workshop	The 4th Workshop on Open-Source Arabic Corpora and Processing Tools
Abbreviated title	OSACT4
Country/Territory	France
City	Marseille
Period	12/05/20 → 12/05/20
Internet address	http://edinburghnlp.inf.ed.ac.uk/workshops/OSACT4/

Keywords / Materials (for Non-textual outputs)

Arabic
hate-speech
offensive language

Access to Document

Multitask Learning_ABU-FARHA_DOA13032020_AFVAccepted author manuscript, 172 KB
Multitask Learning_ABU-FARHA_DOA13032020_VOR_CC-BY-NCFinal published version, 198 KBLicence: Creative Commons: Attribution Non-Commercial (CC-BY-NC)

https://www.aclweb.org/anthology/2020.osact-1.14Licence: Creative Commons: Attribution Non-Commercial (CC-BY-NC)

Cite this

@inproceedings{fb2f963314064b81aa0d0adc41d32fee,

title = "Multitask Learning for Arabic Offensive Language and Hate-Speech Detection",

abstract = "Offensive language and hate-speech are phenomena that spread with the rising popularity of social media. Detecting such content is crucial for understanding and predicting conflicts, understanding polarisation among communities and providing means and tools to filter or block inappropriate content. This paper describes the SMASH team submission to OSACT4{\textquoteright}s shared task on hate-speech and offensive language detection, where we explore different approaches to perform these tasks. The experiments cover a variety of approaches that include deep learning, transfer learning and multitask learning. We also explore the utilisation of sentiment information to perform the previous task. Our best model is a multitask learning architecture, based on CNN-BiLSTM, that was trained to detect hate-speech and offensive language and predict sentiment.",

keywords = "Arabic, hate-speech, offensive language",

author = "{Abu Farha}, Ibrahim and Walid Magdy",

year = "2020",

month = may,

day = "12",

language = "English",

pages = "86",

booktitle = "Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools",

publisher = "European Language Resources Association (ELRA)",

note = "The 4th Workshop on Open-Source Arabic Corpora and Processing Tools, OSACT4 ; Conference date: 12-05-2020 Through 12-05-2020",

url = "http://edinburghnlp.inf.ed.ac.uk/workshops/OSACT4/",

}

Abu Farha, I & Magdy, W 2020, Multitask Learning for Arabic Offensive Language and Hate-Speech Detection. in Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools. European Language Resources Association (ELRA), pp. 86, The 4th Workshop on Open-Source Arabic Corpora and Processing Tools, Marseille, France, 12/05/20. <https://www.aclweb.org/anthology/2020.osact-1.14>

TY - GEN

T1 - Multitask Learning for Arabic Offensive Language and Hate-Speech Detection

AU - Abu Farha, Ibrahim

AU - Magdy, Walid

PY - 2020/5/12

Y1 - 2020/5/12

N2 - Offensive language and hate-speech are phenomena that spread with the rising popularity of social media. Detecting such content is crucial for understanding and predicting conflicts, understanding polarisation among communities and providing means and tools to filter or block inappropriate content. This paper describes the SMASH team submission to OSACT4’s shared task on hate-speech and offensive language detection, where we explore different approaches to perform these tasks. The experiments cover a variety of approaches that include deep learning, transfer learning and multitask learning. We also explore the utilisation of sentiment information to perform the previous task. Our best model is a multitask learning architecture, based on CNN-BiLSTM, that was trained to detect hate-speech and offensive language and predict sentiment.

AB - Offensive language and hate-speech are phenomena that spread with the rising popularity of social media. Detecting such content is crucial for understanding and predicting conflicts, understanding polarisation among communities and providing means and tools to filter or block inappropriate content. This paper describes the SMASH team submission to OSACT4’s shared task on hate-speech and offensive language detection, where we explore different approaches to perform these tasks. The experiments cover a variety of approaches that include deep learning, transfer learning and multitask learning. We also explore the utilisation of sentiment information to perform the previous task. Our best model is a multitask learning architecture, based on CNN-BiLSTM, that was trained to detect hate-speech and offensive language and predict sentiment.

KW - Arabic

KW - hate-speech

KW - offensive language

M3 - Conference contribution

SP - 86

BT - Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools

PB - European Language Resources Association (ELRA)

T2 - The 4th Workshop on Open-Source Arabic Corpora and Processing Tools

Y2 - 12 May 2020 through 12 May 2020

ER -

Multitask Learning for Arabic Offensive Language and Hate-Speech Detection

Abstract / Description of output

Workshop

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Cite this