A partnership between Nvidia and Cadence Design Systems aims to use ultra-precise physics engines to train robots in simulations before deploying them in the social environment.
The two multinational technology companies announced a strategic partnership to accelerate the development of advanced robotics, aiming to combine Cadence’s “physics engines”—software that predicts with millimeter precision how real materials interact—with Nvidia’s AI models.
Nvidia CEO Jensen Huang, quoted by Reuters, explained that the idea is to create simulations so perfect that training a robot inside the computer is identical to that in the real world, but much faster. “We are working with you across the entire range of robotic systems,” Huang stated during a conference in Santa Clara, California.
One of the major bottlenecks in robotics today is the lack of data to train machines. It is difficult and expensive to film robots making mistakes in the physical world until they learn. Cadence’s solution is to generate highly accurate synthetic data. “The more accurate the training data generated, the better the [AI] model will be,” explained Cadence CEO Anirudh Devgan.
In addition to robotics, Cadence presented a tool that should shake up the semiconductor market: an AI agent capable of taking on tasks that were previously exclusive to human engineers.
Written by: Eduardo Quive
