以人类视频中的操作性作为机器人的通用表现形式（英文版）

英文标题：Affordances from Human Videos as a Versatile Representation for Robotics中文摘要：本论文探讨了如何通过利用互联网视频中的人类行为，训练一个可视化驱动的视觉能力模型，以此实现机器人在现实环境中的复杂任务执行。我们将该模型与四个机器人学习范式无缝连接，并在 4 个现实世界环境、超过 10 种不同任务和 2 个机器人平台中展示了其效力。英文摘要：Building a robot that can understand and learn to interact by watching humanshas inspired several vision problems. However, despite some successful resultson static datasets, it remains unclear how current models can be used on arobot directly. In this paper, we aim to bridge this gap by