研究人员推出了SG-Ego，一个扩展Ego4D的新数据集，包含时空场景图，以更好地理解第一人称视频中的人类活动。他们还开发了GLEN，一个基于图的模型，用于处理这些场景图序列以进行动作对齐和时间演化建模。提出的活动驱动图编辑预测（A-GEF）任务将场景动态构建为以动作为条件的结构化变换，从而能够对场景变化进行显式推理。 AI
影响 增强了具身AI和视频理解任务的结构化推理能力。
排序理由 该集群描述了一篇介绍新数据集、模型和视频理解任务的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2607.02425v1 Announce Type: new Abstract: Understanding human behavior while interacting with the surrounding world is crucial for many applications of embodied AI. First-person videos are particularly informative for this problem, as they well capture how activities reshap…
Understanding human behavior while interacting with the surrounding world is crucial for many applications of embodied AI. First-person videos are particularly informative for this problem, as they well capture how activities reshape the scene over time. However, existing approac…