VIP-MINGLE
收藏资源简介:
VIP-MINGLE是由纽约大学创建的多模态对话数据集,旨在研究面对面与视频会议两种环境下群体语言互动的行为差异。该数据集包含约59小时录音,涵盖32个小组的105名参与者,共计7,077个标注片段,数据来源为受控实验任务中采集的原始音视频、心理测量数据及多模态特征。数据集通过严格的被试内设计,使同一组参与者在两种环境下完成相同协作任务,从而支持跨环境的直接行为对比。该资源主要应用于多模态对话分析、人机交互建模及远程通信研究,致力于解决跨环境对话行为差异的量化问题,并为开发适应性强的群体交互模型提供关键数据支持。
VIP-MINGLE is a multimodal dialogue dataset developed by New York University, which aims to investigate the behavioral differences in group linguistic interaction between face-to-face and video conferencing environments. The dataset contains approximately 59 hours of recordings, covering 105 participants from 32 groups, with a total of 7,077 annotated segments. Its data sources include raw audio and video, psychometric data, and multimodal features collected during controlled experimental tasks. The dataset employs a strict within-subjects design, where the same group of participants complete the same collaborative tasks in both environments, thus enabling direct cross-environment behavioral comparisons. This resource is mainly applied to multimodal dialogue analysis, human-computer interaction modeling and remote communication research. It is committed to solving the problem of quantifying the differences in dialogue behaviors across environments, and provides key data support for developing robust group interaction models.
数据集概述
- 名称: VIP-MINGLE: A Corpus for Videoconference and In-Person Multimodal Interaction in Group Language Engagement
- 版本: v1(2026年6月12日发布)
- DOI: 10.5281/zenodo.20670131(代表所有版本)
- 资源类型: 数据集
- 发布机构: Zenodo
- 作者/创作者:
- Andrew Chang(数据管理者),所属机构:纽约大学
- David Poeppel(导师),所属机构:纽约大学
- 许可协议: Creative Commons Attribution Non Commercial No Derivatives 4.0 International
- 状态: 未完全公开(完整数据集将在论文发表后公开)
- 说明: 该语料库专注于在群体语言交流活动中,视频会议与面对面两种模式下的多模态交互研究。
文件信息
- 当前仅包含一个说明文件:README.md(4.9 kB,MD5: 0f6cef93daf8676751a5018a1cc2264e)
统计数据(截至检索时,仅该版本)
- 总浏览量: 10 次
- 总下载量: 3 次
- 总数据量: 14.7 kB
相关标识
- 版本DOI: 10.5281/zenodo.20670132(v1 版本特有)
- 索引: 收录于 OpenAIRE
- 引用格式示例(APA): Chang, A. & Poeppel, D. (2026). VIP-MINGLE: A Corpus for Videoconference and In-Person Multimodal Interaction in Group Language Engagement [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.20670132




