COMMON-TOM
收藏arXiv2024-03-05 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2403.02451v1
下载链接
链接失效反馈官方服务：
资源简介：
COMMON-TOM是由石溪大学和吉兰大学合作创建的第一个基于自然发生的口语对话的理论心智（ToM）数据集。该数据集包含1710个对话轮次，来源于CALLHOME语料库，旨在评估语言模型在理解对话参与者心理状态方面的能力。数据集通过标记对话中的共同基础（CG）来评估ToM，强调了在对话中理解和预测他人信念的重要性。COMMON-TOM的应用领域主要集中在人工智能和自然语言处理中，特别是在开发能够理解和模拟人类心理状态的智能系统方面。
COMMON-TOM is the first theory of mind (ToM) dataset built on naturally occurring spoken dialogues, jointly created by Stony Brook University and Gilan University. It consists of 1,710 dialogue turns sourced from the CALLHOME corpus, and is designed to evaluate the ability of language models to understand the mental states of dialogue participants. The dataset assesses ToM by annotating common ground (CG) in conversations, highlighting the importance of understanding and predicting others' beliefs during dialogue interactions. Its application fields mainly focus on artificial intelligence and natural language processing, especially in the development of intelligent systems capable of understanding and simulating human mental states.
提供机构：
计算机科学系 ♦ 语言学系 ♠ 高级计算科学研究所 † 石溪大学 ‡ 吉兰大学
创建时间：
2024-03-05
