Lune

ACL2024

EmpathicStories++: A Multimodal Dataset for Empathy Towards Personal Experiences

Jocelyn Shen, Yubin Kim, Mohit Hulse, Wazeer Zulfikar, Sharifa Alghowinem, Cynthia Breazeal, Hae Won Park

DOI出版方

摘要

Figure 1 : The EMPATHICSTORIES++ dataset is collected from a month-long in-the-wild deployment of 41 participants (across 269 sessions with 53 hours of data) telling personal stories and reading other people's personal stories with an AI agent. We publicly release all video, audio, and text data in addition to psychometric surveys in order to advance computational empathy research, and more broadly, social-emotional reasoning in AI.