Meta today introduced V-JEPA 2, a 1.2-billion-parameter world model trained primarily on video to support robotic systems.

Format: The course is divided in two parts. In the first part there will be a series of formal lectures on the principles of robot planning, dynamics and control followed by a midterm project ...

One popular technique is to use pre-trained LLMs and VLMs as components in modular systems for task planning ... action models (VLAs) from the ground up to directly generate robot control actions.

Meta has released V-JEPA 2, a powerful new AI world model designed to help robots and self-driving cars better understand and ...

Meta challenges rivals with V-JEPA 2, its new open-source AI world model. By learning from video, it aims to give robots ...

Model helps machines reason like humans using raw video data—no labels required Meta has launched a new artificial ...