DE-ABUSE@TamilNLP-ACL 2022: Transliteration as Data Augmentation for Abuse Detection in Tamil

Abstract

With the rise of social media and internet, there is a necessity to provide an inclusive space and prevent the abusive topics against any gender, race or community. This paper describes the system submitted to the ACL-2022 shared task on fine-grained abuse detection in Tamil. In our approach we transliterated code-mixed dataset as an augmentation technique to increase the size of the data. Using this method we were able to rank 3rd on the task with a 0.290 macro average F1 score and a 0.590 weighted F1 score.