INDICXNLI: A Dataset for Studying NLI in Indic Languages

Abstract

While Indic NLP has made rapid advances recently in terms of availability of corpora and pre-trained models, benchmark dataset on standard NLU tasks are limited. To this end, we introduce INDICXNLI, an NLI dataset for 11 Indic languages. It has been created by high-quality machine translation of the original English XNLI dataset and out analysis attests to the quality of INDICXNLI. By finetuning different pre-trained LMs on this INDICXNLI, we analyze various cross-lingual transfer techniques with respect to the impact of choice of language models, languages, multi-linguality, mix-language input, etc. These experiments provide us with useful insights into the behaviour of pre-trained models for a diverse set of languages. INDICXNLI will be publicly available for research.

Tasks

Reproductions