Uber Buys New York’s Geometric Intelligence to Power New A.I. Lab
Uber is ratcheting up its efforts to advance artificial intelligence technologies as it pushes down the road toward self-driving vehicles.
The ride-hailing app company has launched Uber AI Labs, a new division based in San Francisco focused on A.I. and machine learning research. The initial core team is made up of the 15 employees of New York-based A.I. research startup Geometric Intelligence, which Uber has acquired, according to a Monday announcement on its website. Deal terms weren’t disclosed.
San Francisco-based Uber is best known for its mobile app that enables people to summon a private driver to shuttle them to local destinations, but the company has also moved into other areas, such as delivery of food and other items. Uber envisions its A.I. lab developing computing technologies that enhance its capabilities in determining optimal routes and calculating arrival times, for example.
The bigger goal, of course, is delivering on the promise of autonomous vehicles. Uber began experimenting with self-driving cars in Pittsburgh in September, and it acquired self-driving truck startup Otto in August.
The acquisition of Geometric Intelligence continues the A.I. land grab by all the big tech companies, including Google, Facebook, Amazon, Apple, Microsoft, and Samsung. The deal also adds a feather in the cap of the small but burgeoning A.I. startup cluster in New York. Other players in the Big Apple’s A.I. scene include Cortica and Clarifai.
Geometric Intelligence is interesting because it has taken a different approach to solving A.I., one that involves applying multiple research techniques, the New York Times reported. Its CEO and co-founder, Gary Marcus—a New York University psychology and neural science professor—doesn’t subscribe to some of the beliefs about deep learning that are popular among tech companies these days. Rather than feeding computer systems humongous amounts of data in order to teach them, Marcus believes they might be capable of learning more like humans, picking things up quickly from just one example, or a few.
Similar concepts have been explored in recent years at places like MIT, NYU, and the University of Toronto, the New York Times reported. Now, with the resources of Uber at their disposal, Marcus and his team will get a chance to further test their ideas, and potentially shape the future of transportation—and much more.