clips/beir-nl-fiqa

BEIR-NL是BEIR基准测试的荷兰语翻译版本，涵盖了多个领域的信息检索任务，包括事实核查、问答、生物医学信息检索、论点检索、重复问题检索、引用预测和实体检索等。数据集结构遵循原始BEIR基准，包含语料库、查询和相关性判断文件。数据集通过自动翻译BEIR数据集生成，旨在促进荷兰语信息检索模型的发展。

BEIR-NL is a Dutch-translated version of the BEIR benchmark, a diverse and heterogeneous collection of datasets covering various domains from biomedical and financial texts to general web content. BEIR-NL contains the following tasks: fact-checking (e.g., FEVER, Climate-FEVER, SciFact), question-answering (e.g., NQ, HotpotQA, FiQA-2018), bio-medical IR (e.g., TREC-COVID, NFCorpus), argument retrieval (e.g., Touche-2020, ArguAna), duplicate question retrieval (e.g., Quora, CqaDupstack), citation-prediction (e.g., SCIDOCS), and entity retrieval (e.g., DBPedia). The dataset structure includes corpus, queries, and qrels files, formatted in jsonlines and tsv. The dataset creation is based on automatically translating publicly accessible BEIR datasets into Dutch, with some translations done using GPT-4o-mini and Google Translate. Considerations for using the dataset include limitations such as non-native Dutch resources, data contamination issues, and the potential obsolescence of the benchmark over time.