AmQA

AmQA数据集是首个公开的阿姆哈拉语问答数据集，由亚的斯亚贝巴大学和汉堡大学合作创建。该数据集包含2628个问题-答案对，来源于378篇阿姆哈拉语维基百科文章。数据集通过众包方式创建，确保了问题和答案的质量。AmQA数据集适用于评估和提升阿姆哈拉语问答系统的性能，特别是阅读理解和检索式问答模型。此外，该数据集的建立有助于推动低资源语言的自然语言处理研究。

The AmQA dataset is the first publicly available Amharic question answering dataset, co-created by Addis Ababa University and the University of Hamburg. It contains 2,628 question-answer pairs sourced from 378 Amharic Wikipedia articles. The dataset was developed through crowdsourcing to ensure the quality of both questions and answers. It is applicable for evaluating and enhancing the performance of Amharic question answering systems, especially reading comprehension and retrieval-based question answering models. Furthermore, the establishment of this dataset helps promote natural language processing research on low-resource languages.