BNLI: A Linguistically-Refined Bengali Dataset for Natural Language Inference

TL;DR

BNLI is a carefully curated Bengali NLI dataset that improves linguistic diversity and annotation quality, enabling more effective model training and evaluation for Bengali language understanding.

Contribution

We introduce BNLI, a linguistically refined Bengali NLI dataset with a rigorous annotation process, addressing previous resource limitations and inconsistencies.

Findings

State-of-the-art models perform better on BNLI, indicating improved semantic understanding.

BNLI enhances the reliability and interpretability of Bengali NLI tasks.

Benchmark results establish BNLI as a strong foundation for future Bengali NLP research.

Abstract

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.