Vārta: A Large-Scale Headline-Generation Dataset for Indic Languages
This provides a valuable resource for researchers in Indic NLP and multilinguality, though it is incremental as it focuses on dataset creation rather than novel methods.
The authors tackled the lack of large-scale datasets for headline generation in Indic languages by creating Vārta, a dataset of 41.8 million news articles in 14 Indic languages and English, and showed that it challenges state-of-the-art models and enables pretraining of language models that outperform baselines in NLU and NLG benchmarks.
We present Vārta, a large-scale multilingual dataset for headline generation in Indic languages. This dataset includes 41.8 million news articles in 14 different Indic languages (and English), which come from a variety of high-quality sources. To the best of our knowledge, this is the largest collection of curated articles for Indic languages currently available. We use the data collected in a series of experiments to answer important questions related to Indic NLP and multilinguality research in general. We show that the dataset is challenging even for state-of-the-art abstractive models and that they perform only slightly better than extractive baselines. Owing to its size, we also show that the dataset can be used to pretrain strong language models that outperform competitive baselines in both NLU and NLG benchmarks.