Language Identification of Hindi-English tweets using code-mixed BERT

Language identification of social media text has been an interesting problem of study in recent years. Social media messages are predominantly in code mixed in non-English speaking states. Prior knowledge by pre-training contextual embeddings have shown state of the art results for a range of downstream tasks. Recently, models such as Bidirectional Encoder Representations from Transformers (BERT) have shown that using a large amount of unlabeled data, the pre-trained language models are even more beneficial for learning common language representations. Extensive experiments exploiting transfer learning and fine-tuning BERT models to identify language on Twitter are presented in this paper. The work utilizes a data collection of Hindi-English-Urdu code-mixed text for language pre-training and Hindi-English code-mixed for subsequent word-level language classification. The results show that the representations pre-trained over code-mixed data produce better results by their monolingual counterpart.