Ferrari races are high-octane and blazing fast showdowns. The entire complexion of a race can completely change in the blink of an eye. To the unseasoned onlooker these automobile competitions can be hard to keep up with while watching on a TV screen.
The California-based technology company Intel unveiled a new chip with deep learning capabilities at the NIPS 2017 conference back in December. They’re calling it the “Nervana Neural Network Processor,” and the company is putting it to use by partnering with Ferrari North America to bring real-time, dynamic statistics to the asphalt.
“We’re using A.I. to annotate broadcast drone footage in real-time,” Naveen Rao, the general manager of the Artificial Intelligence Products Group at Intel, tells Inverse. “This will improve fan experience by displaying the team the car is on, the time gap for the driver behind him, and other details that would have had to be edited in manually before.”
Not only will this improve how fans view the race, but the teams will also gain access to data that, until now, could only be analyzed after the race.
Information about cars’ engine performance is crucial for pit crews and fans that want to know how their favorite driver’s metallic steed is fairing in the race. The processor feeds these figures, known as telemetric data, to stat-hungry fans and teams.
Intel’s chip will also be able to analyze live video taken by drones to compare the exit angle a driver takes in each of their turns from lap to lap. The tech can then identify which turn was the fastest, providing newfound statistics that had never been presented during a race before.
“This goes to show that this technology can be used in way more industries rather than the healthcare, agriculture, and finance markets that every A.I. companies are going after.” said Rao. “Of course we’re going after all of those things, as well, but things like this are more fun.”
A lot machine learning technology is ran on graphical processing units , which Rao says serve their purpose but aren’t ideal for running these types of systems. The NNP uses human brain-like features to make these next-generation Ferrari statistics possible.
Rao explains that the chip diffuses information similarly to the human brain, by storing similar data in multiple locations to make it easier and quicker to access.
He also said the human brain is highly effective because it doesn’t deal extremely precise data. Intel’s new processor does something similar. It downgrades the level of its precision, which is so minute you wouldn’t be able to tell, in favor of quicker and more effective processing. Think of it like rounding: You don’t really need to go to the tenth decimal place to get the answer you want.
These two factors make for the type of lightning fast number crunching that motorsports needs.
So the next time you’re watching a fleet of Italian engineering bombing down the race track, remember that all of those stats are thanks to a tiny replica of the human brain.