BanglaQuAD
收藏资源简介:
BanglaQuAD是由BanglaAI和德累斯顿工业大学合作创建的一个高质量的孟加拉语开放领域问答数据集。该数据集包含30,808个由母语者标注的问题-答案对,基于孟加拉语维基百科的658篇文章构建。数据集的创建过程包括文章筛选、文本清洗和人工标注,确保了数据的高质量。BanglaQuAD涵盖了广泛的词汇和问题类型,旨在用于机器阅读理解和信息检索任务,以解决孟加拉语自然语言处理中的低资源问题。
BanglaQuAD is a high-quality Bengali open-domain question answering dataset jointly created by BanglaAI and Technische Universität Dresden. It contains 30,808 question-answer pairs annotated by native speakers, and is constructed based on 658 articles from the Bengali Wikipedia. The dataset creation process includes article screening, text cleaning and manual annotation, which ensures the high quality of the data. BanglaQuAD covers a wide range of vocabulary and question types, and is intended for machine reading comprehension and information retrieval tasks to address the low-resource issue in Bengali natural language processing.
BanglaQuAD-LREC-COLING-24
数据集概述
- 名称: BanglaQuAD-LREC-COLING-24
- 相关论文: LREC-COLING 2024 paper Incorporating Query Recommendation for Improving In-Car Conversational Search

- 1BanglaQuAD: A Bengali Open-domain Question Answering Dataset德累斯顿工业大学 · 2024年



