NewsQA

NewsQA是由Maluuba Research在加拿大蒙特利尔创建的一个大规模机器理解数据集，包含超过119,633个人类生成的问题-答案对，基于12,744篇CNN新闻文章。该数据集通过四阶段过程收集，旨在鼓励探索性问题，需要推理能力。数据集中的答案是从相应文章中提取的文本片段。NewsQA的应用领域主要集中在机器阅读理解，旨在通过复杂的问题解决机器理解文本的挑战。

NewsQA is a large-scale machine comprehension dataset developed by Maluuba Research in Montreal, Canada. It comprises more than 119,633 human-generated question-answer pairs derived from 12,744 CNN news articles. This dataset is collected via a four-stage workflow, designed to elicit exploratory questions that necessitate reasoning abilities. The answers within the dataset are text snippets extracted from their corresponding source articles. The core application domain of NewsQA is machine reading comprehension, with the objective of addressing the challenge of machine understanding of natural language text through complex queries.