Task-Oriented Dialogue (TOD) systems allow users to accomplish tasks by giving directions to the system using natural language utterances. With the widespread adoption of conversational agents and chat platforms, TOD has become mainstream in NLP research today. However, developing TOD systems require massive amounts of data, and there has been limited work done for TOD in low-resource languages like Tamil. Towards this objective, we introduce TamilATIS - a TOD dataset for Tamil which contains 4874 utterances. We present a detailed account of the entire data collection and data annotation process. We train state-of-the-art NLU models and report their performances. The joint BERT model with XLM-Roberta as utterance encoder achieved the highest score with an intent accuracy of 96.26% and slot F1 of 94.01%.

Sentiment classification is a fundamental NLP task of detecting the sentiment polarity of a given text. In this paper we show how solving sentiment span extraction as an auxiliary task can help improve final sentiment classification performance in a low-resource code-mixed setup. To be precise, we don’t solve a simple multi-task learning objective, but rather design a unified transformer framework that exploits the bidirectional connection between the two tasks simultaneously. To facilitate research in this direction we release gold-standard human-annotated sentiment span extraction dataset for Tamil-english code-switched texts. Extensive experiments and strong baselines show that our proposed approach outperforms sentiment and span prediction by 1.27% and 2.78% respectively when compared to the best performing MTL baseline. We also establish the generalizability of our approach on the Twitter Sentiment Extraction dataset. We make our code and data publicly available on GitHub