ViWOZ

ViWOZ是由越南国立大学创建的第一个越南语多轮、多领域任务导向对话数据集，旨在解决低资源语言在任务导向对话系统中的性能问题。该数据集包含5000个对话，总计60,946个完全标注的语句，涵盖多个领域，如旅游、餐饮等。创建过程中，研究团队通过众包方式进行翻译和校正，确保对话的自然性和准确性。ViWOZ数据集的应用领域主要集中在开发和评估多语言任务导向对话系统，特别是在低资源语言环境下。

ViWOZ is the first Vietnamese multi-turn, multi-domain task-oriented dialogue dataset developed by Vietnam National University, aiming to address the performance limitations of low-resource languages in task-oriented dialogue systems. This dataset contains 5,000 dialogues, with a total of 60,946 fully annotated utterances, covering multiple domains including tourism, catering, and others. During the dataset construction, the research team conducted translation and correction via crowdsourcing to guarantee the naturalness and accuracy of the dialogues. The main application scenarios of the ViWOZ dataset focus on the development and evaluation of multilingual task-oriented dialogue systems, particularly in low-resource language environments.