YAGO-WIKI50K

该数据集从Wikidata中提取了前50,000个实体，并将它们与等效的YAGO实体进行关联。此外，该数据集源自tkbc中的时间信息，提供了Wikidata QIDs到YAGO实例的映射。数据集分为两个子集：YAGO-WIKI50K-5K包含5,000个对齐种子，而YAGO-WIKI50K-1K包含1,000个对齐种子。其任务是对实体进行对齐。

This dataset extracts the top 50,000 entities from Wikidata and aligns them with their equivalent YAGO entities. Additionally, built upon temporal information obtained from tkbc, this dataset offers mappings between Wikidata QIDs and YAGO instances. The dataset is partitioned into two subsets: YAGO-WIKI50K-5K consists of 5,000 alignment seeds, whereas YAGO-WIKI50K-1K includes 1,000 alignment seeds. The fundamental task associated with this dataset is entity alignment.

搜集汇总

背景与挑战

背景概述

YAGO-WIKI50K数据集是YAGO3.1知识图谱的一部分，包含多语言的事实、分类和标签数据，数据量庞大且覆盖多种语言领域。