中文自然话题对话(CNTD)数据集
收藏资源简介:
中文自然话题对话(CNTD)数据集是由苏州大学计算机科学与技术学院创建,旨在填补中文自然对话话题数据集的空白。该数据集包含1308条高质量对话,内容涵盖多个领域,如健康、教育、技术等,每条对话均基于一则新闻,增加了对话的真实性和多样性。数据集的创建过程经过精细的手工标注,确保了数据的质量。CNTD数据集主要用于对话话题转移检测的研究,帮助对话系统更有效地理解和预测话题的变化,提升对话的自然度和连贯性。
The Chinese Natural Topic Dialogue (CNTD) dataset was developed by the School of Computer Science and Technology, Soochow University, aiming to fill the research gap of Chinese natural dialogue topic datasets. It contains 1308 high-quality dialogues covering multiple domains such as healthcare, education and technology. Each dialogue is based on a news article, which enhances the authenticity and diversity of the dataset. The construction of the CNTD dataset underwent rigorous manual annotation to ensure data quality. The CNTD dataset is primarily used for research on dialogue topic shift detection, helping dialogue systems more effectively understand and predict topic changes, and improving the naturalness and coherence of conversations.

- 1Topic Shift Detection in Chinese Dialogues: Corpus and Benchmark苏州大学计算机科学与技术学院 · 2023年



