Mobileye to train self-driving cars with data submitted by automakers

Ever since MobilEye, maker of the EyeQ3 chip used by Tesla’s Autopilot front-facing camera, severed ties with the Silicon Valley-based automaker, the Israeli company has been busy forging new alliances with other car companies. It’s ultimate goal is the same as Elon Musk’s — developing self driving technology as quickly as possible. The two former allies turned adversaries agree on the target but disagree on how to get there.

MobilEye says it is now concentrating on ways to train cars to drive themselves without using real-world testing through human drivers. The company intends to feed large amounts of data into computers that will create highly accurate simulations of the real world. It is asking the world’s automakers to provide it with the data from their own autonomous driving research in order to speed up development efforts.

What an autonomous car sees

“If you want to leverage many, many cars, you need to leverage as many carmakers as possible,” says Amnon Shashua, co-founder and CTO of Mobileye. According to the MIT Technology Review, he says his company has figured out a way to let carmakers share their information without losing control over their data. Shashua says Mobileye will announce several deals before the end of the year, starting with Volkswagen.

The Israeli company relies heavily on deep learning networks which are trained to recognize visual information accurately. Until now, human editing of the learning process has been required but Mobileye says it is pioneering the use of “reinforcement learning.” That’s a process that rewards the computers involved for positive results, such as learning to recognize unfamiliar road signs without external assistance. Using reinforcement learning, a computer can learn to do things that would be exceedingly difficult to program.

Mobileye’s biggest obstacle is not software or hardware. Rather, it is convincing automakers to cooperate and share their data. “It makes a ton of sense for car companies to share data, particularly for a problem like this where a vast amount of diverse data is required.” says Karl Iagnaema, CEO of a startup called nuTonomy which is testing automated taxis in Singapore. “Typically, however, leaders are unwilling to pool resources, for fear of diluting their advantage. It makes sharing of resources difficult.”

Shashua says he hopes that Mobileye’s platform will serve as the gold standard for testing and verifying self-driving algorithms.

To Top