Bangla Key2Text: Text Generation from Keywords for a Low Resource Language
For researchers working on low-resource natural language generation, this provides a new benchmark and dataset for Bangla keyword-to-text tasks.
The authors created Bangla Key2Text, a dataset of 2.6 million keyword-text pairs for low-resource Bangla text generation, and showed that fine-tuning mT5 and BanglaT5 on this dataset substantially improves keyword-conditioned generation over zero-shot LLMs.
This paper introduces \textit{Bangla Key2Text}, a large-scale dataset of $2.6$ million Bangla keyword--text pairs designed for keyword-driven text generation in a low-resource language. The dataset is constructed using a BERT-based keyword extraction pipeline applied to millions of Bangla news texts, transforming raw articles into structured keyword--text pairs suitable for supervised learning. To establish baseline performance on this new benchmark, we fine-tune two sequence-to-sequence models, \texttt{mT5} and \texttt{BanglaT5}, and evaluate them using multiple automatic metrics and human judgments. Experimental results show that task-specific fine-tuning substantially improves keyword-conditioned text generation in Bangla compared to zero-shot large language models. The dataset, trained models, and code are publicly released to support future research in Bangla natural language generation and keyword-to-text generation tasks.