MCWQ-R
收藏arXiv2023-06-20 更新2024-06-21 收录
下载链接：
https://github.com/ziwang-klvk/CFQ-RBMT
下载链接
链接失效反馈官方服务：
资源简介：
MCWQ-R是由哥本哈根大学的研究团队创建的多语言组合泛化评估数据集，包含从英语翻译到中文和日语的问题共124,187个。数据集通过规则基础机器翻译方法创建，确保了翻译的忠实性和系统性。MCWQ-R主要用于研究多语言环境下的组合泛化问题，特别是在语义解析领域，旨在解决由于语言差异导致的泛化能力评估难题。
MCWQ-R is a multilingual compositional generalization evaluation dataset developed by a research team from the University of Copenhagen. It contains 124,187 question pairs translated from English into Chinese and Japanese. The dataset was constructed via rule-based machine translation approaches, which ensures the faithfulness and systematicity of the translations. MCWQ-R is primarily designed for researching compositional generalization in multilingual scenarios, especially within the field of semantic parsing, with the goal of resolving the challenges in evaluating generalization capabilities arising from linguistic disparities.
提供机构：
哥本哈根大学
创建时间：
2023-06-20
