SQuAD-es v1.1
收藏arXiv2019-12-13 更新2024-06-21 收录
下载链接：
https://github.com/ccasimiro88/TranslateAlignRetrieve
下载链接
链接失效反馈官方服务：
资源简介：
SQuAD-es v1.1是由TALP研究中心创建的第一个大规模西班牙语问答训练资源，基于自动翻译的斯坦福问答数据集(SQuAD) v1.1。该数据集包含近100,000个问题，这些问题的答案是从维基百科文章中提取的。创建过程中，研究团队开发了Translate-Align-Retrieve (TAR)方法，利用神经机器翻译和无监督对齐算法自动将英语QA数据集翻译成西班牙语。SQuAD-es v1.1主要用于训练西班牙语问答系统，旨在解决跨语言问答任务中的数据稀缺问题，提升西班牙语问答系统的性能。
SQuAD-es v1.1 is the first large-scale Spanish question answering (QA) training resource created by the TALP Research Center, based on the automatically translated Stanford Question Answering Dataset (SQuAD) v1.1. This dataset contains nearly 100,000 questions, whose answers are extracted from Wikipedia articles. During the development process, the research team proposed the Translate-Align-Retrieve (TAR) method, which leverages neural machine translation and unsupervised alignment algorithms to automatically translate the English QA dataset into Spanish. SQuAD-es v1.1 is primarily used for training Spanish QA systems, aiming to address the data scarcity issue in cross-lingual QA tasks and enhance the performance of Spanish QA systems.
提供机构：
TALP研究中心创建时间：
2019-12-11
