BNLI: A Linguistically-Refined Bengali Dataset for Natural Language Inference
This addresses the need for better resources for Bengali NLI, supporting research in low-resource languages, though it is incremental as it refines existing datasets rather than introducing a new method.
The authors tackled the problem of limited and inconsistent Bengali Natural Language Inference (NLI) datasets by introducing BNLI, a refined and linguistically curated dataset, which improved reliability and interpretability in benchmarking transformer-based models for Bengali text.
Despite the growing progress in Natural Language Inference (NLI) research, resources for the Bengali language remain extremely limited. Existing Bengali NLI datasets exhibit several inconsistencies, including annotation errors, ambiguous sentence pairs, and inadequate linguistic diversity, which hinder effective model training and evaluation. To address these limitations, we introduce BNLI, a refined and linguistically curated Bengali NLI dataset designed to support robust language understanding and inference modeling. The dataset was constructed through a rigorous annotation pipeline emphasizing semantic clarity and balance across entailment, contradiction, and neutrality classes. We benchmarked BNLI using a suite of state-of-the-art transformer-based architectures, including multilingual and Bengali-specific models, to assess their ability to capture complex semantic relations in Bengali text. The experimental findings highlight the improved reliability and interpretability achieved with BNLI, establishing it as a strong foundation for advancing research in Bengali and other low-resource language inference tasks.