MULTI2WOZ
收藏arXiv2022-05-21 更新2024-06-21 收录
下载链接：
https://github.com/umanlp/Multi2WOZ
下载链接
链接失效反馈官方服务：
资源简介：
MULTI2WOZ是由德国曼海姆大学数据与网络科学组开发的多语言多领域任务导向对话数据集。该数据集包含2000个对话，覆盖中文、德语、阿拉伯语和俄语，每个对话均经过精心翻译和校对，确保与英文数据集的直接可比性。MULTI2WOZ的创建旨在解决跨语言迁移在自然语言处理中的关键应用问题，通过提供高质量的跨语言对话数据，支持对跨语言迁移性能的系统研究和评估。
MULTI2WOZ is a multilingual, multi-domain task-oriented dialogue dataset developed by the Data and Network Science Group at the University of Mannheim, Germany. This dataset contains 2000 dialogues covering Chinese, German, Arabic and Russian, with each dialogue meticulously translated and proofread to ensure direct comparability with the English version of the dataset. MULTI2WOZ was developed to address key application issues of cross-lingual transfer in natural language processing. By providing high-quality cross-lingual dialogue data, it supports systematic research and evaluation of cross-lingual transfer performance.
提供机构：
数据与网络科学组，曼海姆大学，德国
创建时间：
2022-05-21
