Georgia Gkioxari proposes shifting computer vision research toward object affordances, physical-space reasoning, and efficient robotics models

From: Georgia Gkioxari proposes shifting computer vision research toward object affordances, physical-space reasoning, and efficient robotics models

Reposted by Georgia Gkioxari·7hView on

F. Güney@ftm_guney·18hView on

Georgia identified 3 directions for interesting research in vision: - we need to go beyond objects and semantics and start thinking about their utility, affordances. - we need to reason in the physical space, beyond LLMs/VLMs - we need efficient AI for applications like robotics.