MUTANT: A Multi-sentential Code-mixed Hinglish Dataset

Findings of the Association for Computational Linguistics: EACL 2023, pages 744–753 May 2-6, 2023 ©2023 Association for Computational Linguistics MUTANT: A Multi-sentential Code-mixed Hinglish Dataset Rahul Gupta IIT Gandhinagar Gandhinagar, Gujarat, India gupta.rahul@iitgn.ac.in Vivek Srivastava TCS Research Pune, Maharashtra, India srivastava.vivek2@tcs.com Mayank Singh IIT Gandhinagar Gandhinagar, Gujarat, India singh.mayank@iitgn.ac.in Abstract The multi-sentential long sequence textual data unfolds several interesting research directions pertaining to natural language processing and generation. Though we observe several high- quality long-sequence datasets for English and other monolingual languages, there is no sig- nificant effort in building such resources for code-mixed languages such as Hinglish (code- mixing of Hindi-English). In this paper, we pro- pose a novel task of identifying multi-sentential code-mixed text (MCT) from multilingual arti- cles. As a use case, we leverage multilingual ar- ticles from two different data sources and build a first-of-its-kind multi-sentential code-mixed Hinglish dataset i.e., MUTANT. We propose a token-level language-aware pipeline and ex- tend the existing metrics measuring the degree of code-mixing to a multi-sentential framework and automatically identify MCT in the multilin- gual articles. The MUTANT dataset comprises 67k articles with 85k identified Hinglish MCTs. To facilitate future research directions, we will make the dataset and the code publicly avail- able upon publication. 1 Introduction Over the years, we have seen enormous down- stream applications of multi-sentential datasets in the areas such as question-answering (Joshi et al., 2017; Tapaswi et al., 2016), summarization (Sharma et al., 2019; Cachola et al., 2020), machine translation (Bao et al., 2021), etc. The existing state-of-the-art methods prove challenging to scale effectively and efficiently on multi-sentential long sequence text (Ainslie et al., 2020), which unplugs several exciting research avenues. Unfortunately, to a large extent, the majority of the research on multi-sentential data is dominated by a few popular monolingual languages such as English, Chinese, and Spanish. Due to this, code-mixed languages (among other low-resource and under-explored lan- guages) suffer from non-existent works in the afore- mentioned areas of interest. Figure 1: Example MCT and the corresponding article’s title form two multilingual data sources: (A) Dainik Jagran news article and (B) Man-ki-baat speech tran- script. We color code the tokens as: English, Hindi, and language independent. We posit that due to several inherent challenges, the NLP community hold back on building multi- sentential datasets for the low-resource and code- mixed languages. One of the most significant bot- tlenecks in building such resources is the unavail- ability of MCT on traditional and widely popular data sources such as social media platforms where the short-length and noisy code-mixed text is avail- able in abundance. It presents several challenges such as the difficulty in curating a large-scale multi- sentential dataset at ease. Another major challenge is the lack of metrics to measure the degree of code- mixing in the multi-sentential framework. The ex- isting metrics such as code-mixing index (Das and Gambäck, 2014) and multilingual-index (Barnett et al., 2000) already suffers from major limitations (Srivastava and Singh, 2021a) in the short-length text format. In such a scenario, it …

Content truncated to 500 words