Uber Sees Gold in Autonomous Vehicle Field
Dec 7, 2016 9:49 AM PT
Uber on Monday announced the acquisition of startup Geometric Intelligence, a move designed to increase its presence in the autonomous vehicle technology field.
Geometric CEO Gary Marcus will become the chief of Uber's new artificial intelligence business. He will lead the newly formed Uber AI Labs, which will be dedicated to conducting research into artificial intelligence and machine learning for Uber, according to Jeff Holden, the company's chief product officer.
The new AI lab initially will be staffed with the 15 employees from Geometric Intelligence, and additional hires are expected in the future.
Marcus, a professor of psychology and neural science at New York University in Manhattan, in 2014 cofounded Geometric Intelligence with Zoubin Ghahramani, a Cambridge professor of machine learning; Kenneth Stanley, a professor of computer science at the University of Central Florida; and Douglas Bemis, a PhD graduate in neurolinguistics at NYU.
The firm grew out of a business incubator at NYU through a partnership between its Tandon School of Engineering and the New York City Data Future Labs.
Additional AI and machine learning will be used to help the company with a variety of scenarios, according to Uber's Holden, including everything from finding an optimal route for the arrival of your car to placing an order using UberEATS, or matching up riders for UberPOOL.
It could be applied to the development of new types of robotics, or helping a self-driving machine navigate the world using autonomous vehicles, or even helping guide an aircraft through busy airspace.
Uber did not disclose terms of the Geometric Intelligence acquisition.
The agreement will further Uber's efforts to expand into autonomous vehicle technology. The company already is testing autonomous vehicles in Pittsburgh as part of a pilot program to examine how well self-driving cars perform while transporting passengers under real-world conditions in the city.
Autonomous Vehicle Demand
"Uber's ultimate goal is to create a massive fleet of self-driving vehicles for every major global market," said Karl Brauer, executive publisher of Kelley Blue Book.
"Processing the amount of data these vehicles will collect is impossible without advanced AI," he told the E-Commerce Times.
Machine learning is a necessary technology to make self-driving vehicles able to navigate real-world conditions, he pointed out, but Uber wants to have more than just safe self-driving vehicles. It wants to implement the most efficient methodology for deploying those cars, and machine learning will be a key component of making that happen.
Artificial intelligence "seems to be the critical element" for the safe use of autonomous cars, observed Steven Polzin, director of mobility policy research at the University of South Florida's Center for Urban Transportation Research.
Transportation fatalities are measured in terms of incidents per 100 million miles of driving, he said. Overall, there are 1.15 fatalities per 100 million miles, but the autonomous vehicle industry has yet to reach the level of development and testing necessary to match that level of road experience.
"I think Uber is realizing there is a finite ability to scale up services dependent on drivers," Polzin told the E-Commerce Times. "If the economy takes off and the job market gets even stronger, it could be increasingly difficult to find drivers."
Other factors can play into the increasing demand as well. For example, with the legalization of recreational marijuana in Colorado, the city of Denver recently was unable to find enough qualified bus drivers who could pass a drug test, he said.
Uber previously had announced plans to develop a joint autonomous vehicle testing center with Carnegie Mellon University. However those plans failed to materialize, and a number of the researchers involved went to work for Uber directly to assist in the firm's creation of a testing and development center.