NTX

NTX数据集是由杜克大学等机构创建的多语言评估数据集，专注于时间和数值表达的提取与规范化，涵盖14种语言。数据集通过多年在商业应用中的实际使用积累而成，包含8种数值子类型和10种时间子类型，旨在提供精细化的实体覆盖和易于下游应用使用的数据。NTX数据集的应用领域广泛，包括信息检索、关系提取、对话语言理解等，旨在解决现有数据集在时间和数值表达处理上的不足。

The NTX dataset is a multilingual evaluation dataset created by institutions including Duke University, focusing on the extraction and normalization of temporal and numerical expressions, covering 14 languages. It is accumulated through years of practical deployment in commercial applications, containing 8 numerical subtypes and 10 temporal subtypes, with the objective of providing fine-grained entity coverage and easily usable data for downstream applications. The NTX dataset has a wide range of application fields, including information retrieval, relation extraction, conversational language understanding and more, and is designed to address the shortcomings of existing datasets in the processing of temporal and numerical expressions.