DialogUSR

该数据集是一个高质量的数据集，涵盖了23个领域。它通过多步骤的众包过程收集而成，旨在促进对多意图用户查询的理解。该数据集包含了复杂的用户查询，这些查询可能涉及话题切换和不完整的陈述，真实反映了用户的实际行为。该任务的目的是进行对话语句的分割与重组，以便于多意图检测。

This is a high-quality dataset covering 23 domains. It was collected via a multi-step crowdsourcing process, aiming to facilitate the understanding of multi-intent user queries. The dataset contains complex user queries that may involve topic switching and incomplete utterances, which authentically reflect users' actual behaviors. The objective of the corresponding task is to segment and reorganize conversational utterances to enable multi-intent detection.