Bridging the Data Gap: Creating a Hindi Text Summarization Dataset from the English XSUM

TL;DR

This paper addresses the data scarcity for Hindi text summarization by proposing a cost-effective, automated pipeline to create a Hindi version of the English XSUM dataset. It translates and linguistically adapts XSUM using forward translation, backtranslation, error correction, paraphrasing, and selective LLM translation, with validation guided by COMET and other automated metrics. The approach yields a diverse, high-quality Hindi summarization corpus mirroring XSUM’s complexity, and demonstrates a practical, scalable framework for expanding NLP resources to other low-resource languages. The Hindi XSUM dataset is publicly released, offering a direct resource for Hindi NLP research and a blueprint for democratizing NLP through automated, scalable data creation.

Abstract

Current advancements in Natural Language Processing (NLP) have largely favored resource-rich languages, leaving a significant gap in high-quality datasets for low-resource languages like Hindi. This scarcity is particularly evident in text summarization, where the development of robust models is hindered by a lack of diverse, specialized corpora. To address this disparity, this study introduces a cost-effective, automated framework for creating a comprehensive Hindi text summarization dataset. By leveraging the English Extreme Summarization (XSUM) dataset as a source, we employ advanced translation and linguistic adaptation techniques. To ensure high fidelity and contextual relevance, we utilize the Crosslingual Optimized Metric for Evaluation of Translation (COMET) for validation, supplemented by the selective use of Large Language Models (LLMs) for curation. The resulting dataset provides a diverse, multi-thematic resource that mirrors the complexity of the original XSUM corpus. This initiative not only provides a direct tool for Hindi NLP research but also offers a scalable methodology for democratizing NLP in other underserved languages. By reducing the costs associated with dataset creation, this work fosters the development of more nuanced, culturally relevant models in computational linguistics.

Figures (16)

• Figure 1: Process Architecture
• Figure 2: sample Annotation UI used for Human Annotation
• Figure 3: Document BertScore
• Figure 4: Document BLEU Score
• Figure 5: Document CHRF score