Towards Offensive Language Identification for Tamil Code-Mixed YouTube Comments and Posts

Offensive Language detection in social media platforms has been an active field of research over the past years. In non-native English-speaking countries, social media users mostly use a code-mixed form of text in their posts/comments. This poses several challenges for offensive content identification tasks and considering the low resources available for the Tamil language, the task becomes much more challenging. The current study presents extensive experiments using multiple deep learning and transfers learning models to detect offensive content on YouTube. We propose a novel and flexible approach of selective translation and transliteration techniques to reap better results from fine-tuning and ensembling multilingual transformer networks like BERT, DistilBERT, and XLM-RoBERTa. The experimental results showed that ULMFiT is the best model for this task. The best performing models were ULMFiT and mBERT-BiLSTM for this Tamil code-mix dataset instead of more popular transfer learning models such as DistilBERT and XLM-RoBERTa and hybrid deep learning models. The proposed model ULMFiT and mBERT-BiLSTM yielded good results and are promising for effective offensive speech identification in low-resourced languages.