Part 4/8:

NVIDIA's Role in Advancing Physical AI

To facilitate the development of these advanced robotic systems, NVIDIA has emerged as a key player. The company has introduced three types of computing systems tailored to support developers in creating robust physical AI applications.

DGX systems are initially used for training complex models.
The AI models are then refined and tested using a technique known as reinforcement learning-physics feedback in an innovative environment called Omniverse.
Finally, the trained AI is implemented on NVIDIA Jetson AGX robotics computers, allowing for practical applications in robotics.