Deep Learning Gives Nvidia Realism Advantage
According to a story in The Verge, Nvidia has developed a video game system that generates AI-generated images that are far superior to any others currently on the market.
Already, AI can be used in film editing and to manipulate images, but the new technology creates video that’s closer to photo realism than other systems.
“It’s a new way to render video content using deep learning,” Nvidia’s vice president of applied deep learning, Bryan Catanzaro, told The Verge. “Obviously Nvidia cares a lot about generating graphics [and] we’re thinking about how AI is going to revolutionize the field.”
A video game demo of a car driving through streets shows how the pairing of deep learning and video game systems render a more realistic experience than in the standard graphics available now.
The graphics were created using Nvidia’s most powerful supercomputing GPU, the Titan V, which retails for $3,000. This means a mass market product is still a long way from being available. However, MarketWatch predicts the company’s new supercomputing chips will move even farther ahead in its market.
read more at theverge.com