Multilingual offensive language identification for low-resource languages
Abstract
Offensive content is pervasive in social media and a reason for concern to companies and government organizations. Several studies have been recently published investigating methods to detect the various forms of such content (e.g., hate speech, cyberbullying, and cyberaggression). The clear majority of these studies deal with English partially because most annotated datasets available contain English data. In this article, we take advantage of available English datasets by applying cross-lingual contextual word embeddings and transfer learning to make predictions in low-resource languages. We project predictions on comparable data in Arabic, Bengali, Danish, Greek, Hindi, Spanish, and Turkish. We report results of 0.8415 F1 macro for Bengali in TRAC-2 shared task [23], 0.8532 F1 macro for Danish and 0.8701 F1 macro for Greek in OffensEval 2020 [58], 0.8568 F1 macro for Hindi in HASOC 2019 shared task [27], and 0.7513 F1 macro for Spanish in in SemEval-2019 Task 5 (HatEval) [7], showing that our approach compares favorably to the best systems submitted to recent shared tasks on these three languages. Additionally, we report competitive performance on Arabic and Turkish using the training and development sets of OffensEval 2020 shared task. The results for all languages confirm the robustness of cross-lingual contextual embeddings and transfer learning for this task.Citation
Ranasinghe, T. and Zampieri, M. (2021) Multilingual offensive language identification for low-resource languages. ACM Transactions on Asian and Low-Resource Language Information Processing, 21(1), 4. https://doi.org/10.1145/3457610.Publisher
Association for Computing MachineryJournal
ACM Transactions on Asian and Low-Resource Language Information ProcessingDOI
10.1145/3457610Additional Links
https://dl.acm.org/doi/10.1145/3457610Type
Journal articleLanguage
enDescription
This is an accepted manuscript of a paper published by ACM on 10/11/2021, available online: https://doi.org/10.1145/3457610. The accepted manuscript of the publication may differ from the final published version.ISSN
2375-4699EISSN
2375-4702ae974a485f413a2113503eed53cd6c53
10.1145/3457610
Scopus Count
Except where otherwise noted, this item's license is described as https://creativecommons.org/licenses/by-nc-nd/4.0/