Bangla Key2Text: Text Generation from Keywords for a Low Resource Language

2026-04-21

AI Summary

This work addresses the challenge of keyword-driven text generation for Bengali, a low-resource language, where large-scale supervised data is scarce. To overcome this limitation, the authors construct the first large-scale dataset comprising 2.6 million keyword–text pairs, automatically generated from news corpora using a BERT-based keyword extraction pipeline. They then perform task-specific fine-tuning on mT5 and BanglaT5 models. Experimental results demonstrate that the proposed approach significantly outperforms zero-shot large language models across both automatic metrics and human evaluations, achieving high-quality and controllable text generation. The study further advances research in controllable generation for low-resource languages by publicly releasing the dataset, models, and code.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

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.

Problem

• low-resource language
• keyword-to-text generation
• Bangla NLP
• text generation
• dataset scarcity

Innovation

• keyword-to-text generation
• low-resource language
• Bangla NLP
• BERT-based keyword extraction
• sequence-to-sequence fine-tuning