Leveraging time-dependent lexical features for offensive language detection

Barbara McGillivray; Malithi Alahapperuma; Jonathan Cook; Chiara Di Bonaventura; Albert Merono Penuela; Gareth Tyson; Steven Wilson

Leveraging time-dependent lexical features for offensive language detection

Barbara McGillivray, Malithi Alahapperuma, Jonathan Cook, Chiara Di Bonaventura, Albert Merono Penuela, Gareth Tyson, Steven Wilson

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

204 Downloads (Pure)

Abstract

We present a study on the integration of time-sensitive information in lexicon-based offensive language detection systems. Our focus is on Offenseval sub-task A, aimed at detecting offensive tweets. We apply a semantic change detection algorithm over a short time span of two years to detect words whose semantics has changed and we focus particularly on those words that acquired or lost an offensive meaning between 2019 and 2020. Using the output of this semantic change detection approach, we train an Support Vector Machine (SVM) classifier on the Offenseval 2019 training set. We build on the already competitive SINAI system submitted to Offenseval 2019 by adding new lexical features, including those that capture the change in usage of words and their association with emerging offensive usages. We discuss the challenges, opportunities and limitations of integrating semantic change detection in offensive language detection models. Our work draws attention to an often neglected aspect of offensive language, namely that the meanings of words are constantly evolving and that NLP systems that account for this change can achieve good performance even when not trained on the most recent training data.

Original language	English
Title of host publication	Proceedings of the 1st Workshop of Ever Evolving NLP, EMNLP 2022
Publication status	Published - 7 Dec 2022

Access to Document

Accepted_versionAccepted author manuscript, 250 KBLicence: CC BY
Offensive_language_semantic_changeFinal published version, 249 KB

https://aclanthology.org/2022.evonlp-1.7Licence: CC BY

Cite this

@inbook{116be7531da84598bc0686119e7b4020,

title = "Leveraging time-dependent lexical features for offensive language detection",

abstract = "We present a study on the integration of time-sensitive information in lexicon-based offensive language detection systems. Our focus is on Offenseval sub-task A, aimed at detecting offensive tweets. We apply a semantic change detection algorithm over a short time span of two years to detect words whose semantics has changed and we focus particularly on those words that acquired or lost an offensive meaning between 2019 and 2020. Using the output of this semantic change detection approach, we train an Support Vector Machine (SVM) classifier on the Offenseval 2019 training set. We build on the already competitive SINAI system submitted to Offenseval 2019 by adding new lexical features, including those that capture the change in usage of words and their association with emerging offensive usages. We discuss the challenges, opportunities and limitations of integrating semantic change detection in offensive language detection models. Our work draws attention to an often neglected aspect of offensive language, namely that the meanings of words are constantly evolving and that NLP systems that account for this change can achieve good performance even when not trained on the most recent training data. ",

author = "Barbara McGillivray and Malithi Alahapperuma and Jonathan Cook and {Di Bonaventura}, Chiara and {Merono Penuela}, Albert and Gareth Tyson and Steven Wilson",

year = "2022",

month = dec,

day = "7",

language = "English",

booktitle = "Proceedings of the 1st Workshop of Ever Evolving NLP, EMNLP 2022",

}

TY - CHAP

T1 - Leveraging time-dependent lexical features for offensive language detection

AU - McGillivray, Barbara

AU - Alahapperuma, Malithi

AU - Cook, Jonathan

AU - Di Bonaventura, Chiara

AU - Merono Penuela, Albert

AU - Tyson, Gareth

AU - Wilson, Steven

PY - 2022/12/7

Y1 - 2022/12/7

N2 - We present a study on the integration of time-sensitive information in lexicon-based offensive language detection systems. Our focus is on Offenseval sub-task A, aimed at detecting offensive tweets. We apply a semantic change detection algorithm over a short time span of two years to detect words whose semantics has changed and we focus particularly on those words that acquired or lost an offensive meaning between 2019 and 2020. Using the output of this semantic change detection approach, we train an Support Vector Machine (SVM) classifier on the Offenseval 2019 training set. We build on the already competitive SINAI system submitted to Offenseval 2019 by adding new lexical features, including those that capture the change in usage of words and their association with emerging offensive usages. We discuss the challenges, opportunities and limitations of integrating semantic change detection in offensive language detection models. Our work draws attention to an often neglected aspect of offensive language, namely that the meanings of words are constantly evolving and that NLP systems that account for this change can achieve good performance even when not trained on the most recent training data.

AB - We present a study on the integration of time-sensitive information in lexicon-based offensive language detection systems. Our focus is on Offenseval sub-task A, aimed at detecting offensive tweets. We apply a semantic change detection algorithm over a short time span of two years to detect words whose semantics has changed and we focus particularly on those words that acquired or lost an offensive meaning between 2019 and 2020. Using the output of this semantic change detection approach, we train an Support Vector Machine (SVM) classifier on the Offenseval 2019 training set. We build on the already competitive SINAI system submitted to Offenseval 2019 by adding new lexical features, including those that capture the change in usage of words and their association with emerging offensive usages. We discuss the challenges, opportunities and limitations of integrating semantic change detection in offensive language detection models. Our work draws attention to an often neglected aspect of offensive language, namely that the meanings of words are constantly evolving and that NLP systems that account for this change can achieve good performance even when not trained on the most recent training data.

M3 - Conference paper

BT - Proceedings of the 1st Workshop of Ever Evolving NLP, EMNLP 2022

ER -

Leveraging time-dependent lexical features for offensive language detection

Abstract

Access to Document

Fingerprint

Cite this