Topical-Chat

该数据集是一个以知识为基础的人类间对话集合，覆盖了8个广泛的主题，其中的对话伙伴并没有明确的角色设定。该数据集旨在推动开放领域对话人工智能的研究，并使用自动化和人工评估指标来评估模型性能。这项工作的任务是开放域对话人工智能。

This dataset is a knowledge-based corpus of human-to-human conversations covering eight broad topics, with no explicit role assignments for the dialogue participants. It aims to advance research in open-domain conversational artificial intelligence, and uses both automated and human evaluation metrics to assess model performance. The core task of this work is open-domain conversational AI.