COVID-19 Mythbusters in World Languages

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 3048–3055 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 3048 COVID-19 Mythbusters in World Languages Mana Ashida1, Jin-Dong Kim2, Seunghun J. Lee3,4 1 Yahoo Japan Corporation, 2 Database Center for Life Science, ROIS, 3 International Christian University, 4 IITG maashida@yahoo-corp.jp, jdkim@dbcls.rois.ac.jp, seunghun@icu.ac.jp Abstract This paper introduces a multi-lingual database containing translated texts of COVID-19 mythbusters. The database has translations into 115 languages as well as the original English texts, of which the original texts are published by World Health Organization (WHO). This paper then presents preliminary analyses on latin-alphabet-based texts to see the potential of the database as a resource for multilingual linguistic analyses. The analyses on latin-alphabet-based texts gave interesting insights into the resource. While the amount of translated texts in each language was small, character bi-grams with normalization (lowercasing and removal of diacritics) turned out to be an effective proxy for measuring the similarity of the languages, and the affinity ranking of language pairs could be obtained. Additionally, the hierarchical clustering analysis is performed using the character bigram overlap ratio of every possible pair of languages. The result shows the cluster of Germanic languages, Romance languages, and Southern Bantu languages. In sum, the multilingual database not only offers fixed set of materials in numerous languages, but also serves as a preliminary tool to identify the language family using text-based similarity measure of bigram overlap ratio. Keywords: COVID-19 Mythbusters, cross-linguistic database, clustering analysis 1. Introduction The rapid progress of natural language processing (NLP) applications are often limited to languages that already have multitude of resources such as English, French or Japanese. As such, a significant number of languages do not have any resources for NLP ap- plications (Joshi et al., 2020). Such a disparity be- tween languages turned out to be problematic under the COVID-19 pandemic situation where information sharing became important all over the world. Addition- ally, preventing the spreading of misinformation be- came equally important. The lack of NLP applications that target languages with low resources prevented nec- essary health measures from circulating promptly to re- gions where these languages are spoken. At the onset of the COVID-19 pandemic, the World Health Organization (WHO) recognized problems that arise due to misinformation that circulated via Social Networking Sites or Applications. Soon, a webpage that aims to bust the myths concerning COVID-19 was created to raise awareness of these false beliefs that were spreading online (Lee and Won, 2021). This paper has two aims. First, it reports the creation of a freely downloadable web-based database in about 116 versions, which was created based on the initiative “COVID-19 Mythbusters in World Languages” created by the third author. The 115 languages include 40 lan- guages that do not have a single page on wikipedia which is written in these languages. As of January 14, 2022, Wikipedia is available in 341 languages1. Sec- ond, results of preliminary analyses of the latin-based texts in the …

Content truncated to 500 words