DWIE Dataset
收藏paperswithcode.com2025-03-23 收录
下载链接:
https://paperswithcode.com/dataset/dwie
下载链接
链接失效反馈官方服务:
资源简介:
The 'Deutsche Welle corpus for Information Extraction' (DWIE) is a multi-task dataset that combines four main Information Extraction (IE) annotation sub-tasks: (i) Named Entity Recognition (NER), (ii) Coreference Resolution, (iii) Relation Extraction (RE), and (iv) Entity Linking. DWIE is conceived as an entity-centric dataset that describes interactions and properties of conceptual entities on the level of the complete document.
《德意志之声信息提取语料库》(DWIE)是一项多任务数据集,该数据集融合了四个主要的信息提取(IE)标注子任务:(i) 命名实体识别(NER),(ii) 指代消解,(iii) 关系抽取(RE),以及(iv)实体链接。DWIE被视为一个以实体为中心的数据集,它描述了在完整文档层面上概念实体的交互和属性。
提供机构:
paperswithcode.com



