Introducing A Bangla Sentence - Gloss Pair Dataset for Bangla Sign Language Translation and Research
This addresses a low-resource NLP problem for Bangla Sign Language translation, providing a foundational dataset and baseline models, though it is incremental as it builds on existing methods for a specific domain.
The authors tackled the lack of large-scale datasets for Bangla Sign Language translation by introducing Bangla-SGP, a dataset of 1,000 human-annotated and 3,000 synthetically generated sentence-gloss pairs, and fine-tuned transformer models like mBart50 and GPT4.1-nano, achieving performance evaluated with BLEU scores compared to the RWTH-PHOENIX-2014T benchmark.
Bangla Sign Language (BdSL) translation represents a low-resource NLP task due to the lack of large-scale datasets that address sentence-level translation. Correspondingly, existing research in this field has been limited to word and alphabet level detection. In this work, we introduce Bangla-SGP, a novel parallel dataset consisting of 1,000 human-annotated sentence-gloss pairs which was augmented with around 3,000 synthetically generated pairs using syntactic and morphological rules through a rule-based Retrieval-Augmented Generation (RAG) pipeline. The gloss sequences of the spoken Bangla sentences are made up of individual glosses which are Bangla sign supported words and serve as an intermediate representation for a continuous sign. Our dataset consists of 1000 high quality Bangla sentences that are manually annotated into a gloss sequence by a professional signer. The augmentation process incorporates rule-based linguistic strategies and prompt engineering techniques that we have adopted by critically analyzing our human annotated sentence-gloss pairs and by working closely with our professional signer. Furthermore, we fine-tune several transformer-based models such as mBart50, Google mT5, GPT4.1-nano and evaluate their sentence-to-gloss translation performance using BLEU scores, based on these evaluation metrics we compare the model's gloss-translation consistency across our dataset and the RWTH-PHOENIX-2014T benchmark.