TempQA-WD

TempQA-WD数据集由IBM研究院创建，专注于时序问题回答，基于Wikidata这一广泛维护的开放知识库。数据集包含839个问题，涉及单跳、多跳及时序推理问题，通过SPARQL查询进行评估，支持跨多个知识库如Freebase和Wikidata的通用性。创建过程中，数据集利用Wikidata的结构化和实时更新特性，通过人工标注确保SPARQL查询的准确性和答案的时效性。该数据集主要用于评估和推动时序知识库问题回答系统的研究，特别是在复杂时序推理方面的应用。

The TempQA-WD dataset, developed by IBM Research, focuses on temporal question answering and is built upon Wikidata, a widely maintained open knowledge base. It contains 839 questions covering single-hop, multi-hop, and temporal reasoning tasks, which are evaluated via SPARQL queries and supports generalizability across multiple knowledge bases such as Freebase and Wikidata. During its curation, the dataset leverages the structured characteristics and real-time update capabilities of Wikidata, with manual annotations employed to guarantee the accuracy of SPARQL queries and the timeliness of corresponding answers. This dataset is primarily used to evaluate and advance research on temporal knowledge base question answering systems, particularly for applications involving complex temporal reasoning.