clips/beir-nl-trec-covid

BEIR-NL是BEIR基准的荷兰语翻译版本，涵盖了多个领域的任务，如事实核查、问答、生物医学信息检索等。数据集结构包括corpus、queries和qrels文件，分别包含文档、查询和查询文档相关性判断。数据集创建的目的是为了促进荷兰语信息检索模型的发展，通过自动翻译BEIR数据集实现。此外，数据集存在一些局限性，如非原生荷兰语资源、数据污染和基准有效性随时间变化的问题。

BEIR-NL is a Dutch-translated version of the BEIR benchmark, covering various tasks such as fact-checking, question-answering, and biomedical information retrieval. The dataset structure includes corpus, queries, and qrels files, containing documents, queries, and query-document relevance judgments, respectively. The dataset was created to promote the development of Dutch information retrieval models by automatically translating the BEIR dataset. Additionally, the dataset has some limitations, such as non-native Dutch resources, data contamination, and the diminishing validity of the benchmark over time.