Wired Explaining Dialogue Corpus
收藏资源简介:
Wired Explaining Dialogue Corpus是由帕德博恩大学计算机科学系创建的数据集,包含65个英语对话,来自Wired杂志的5 Levels视频系列。数据集涵盖13个科学相关主题,每个主题由一位专家向五位不同专业水平的解释对象解释。数据集中的所有1550个对话轮次均由五位独立专业人士手动标注,包括讨论的主题、对话行为和解释动作。该数据集旨在支持可解释人工智能(XAI)的研究,帮助AI学习如何在对话中与人类交互解释,以解决AI决策和行为的理解问题。
The Wired Explaining Dialogue Corpus is a dataset developed by the Department of Computer Science at the University of Paderborn. It consists of 65 English dialogues sourced from the 5 Levels video series of Wired magazine. The corpus covers 13 science-related topics, where each topic is explained by a single expert to five audience members with distinct levels of professional expertise. All 1,550 dialogue turns in this dataset were manually annotated by five independent professionals, covering the discussed topics, dialogue acts, and explanatory actions. This corpus is intended to support research in Explainable Artificial Intelligence (XAI), enabling AI systems to learn how to interactively explain concepts with humans during conversational interactions, thereby addressing the problem of human comprehension of AI decision-making and behavioral outputs.




