New dataset and benchmark advance Bangla text-to-gloss translation for BdSL

Researchers have developed the first dataset and benchmark for Bangla text-to-gloss translation, addressing a significant gap for the Bangla Sign Language (BdSL) community. The dataset includes manually annotated and synthetically generated sentence-gloss pairs, designed to aid low-resource translation efforts. Experiments showed that GPT-5.4 performed best overall, while a fine-tuned mBART model offered competitive results with a smaller size, and Qwen-3 excelled in human evaluations. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces a new dataset and benchmark for Bangla text-to-gloss translation, potentially improving accessibility for the deaf and hard-of-hearing community in Bangladesh.

RANK_REASON This is a research paper introducing a new dataset and benchmark for a specific NLP task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• Bangla Sign Language
• BdSL
• Bangla
• GPT-5.4
• mBART
• Qwen-3
• arXiv

• paper
• other