Hate Speech detection in English and Malayalam Code-Mixed Text using BERT embedding

Hate speech detection is a very popular research area for past few years. Hate speech is given various definition by various researchers. In this paper we try to analyse the use of BERT embedding in hate speech detection in low resource language like Malayalam. The crucial challenge faced by researchers in this area are that most non-English languages are represented in code-mixed form in Social media. Here we work with transformer-based models to classify tweets as hate or non-hate content. Hence this is a novel approach that uses BERT in non-English text.