Some of Silicon Valley's heaviest hitters, including Elon Musk, Peter Thiel, and Reid Hoffman are all backing OpenAI, a non-profit AI research company with a focus on on deep learning. The first challenge on the list? Robot maids that can teach themselves how to clean.
Fetch Robotics currently makes warehouse bots, the type that will follow employees around as they load products for carrying. Fetch also makes the software running their robots, and the plan OpenAI has is to use deep learning to have them train themselves through a process of trial and error. The company has brought on Pieter Abbeel—a professor at the University of California, Berkeley who focuses robot learning—to teach these robots how to learn tasks that are performed by hand, in this case common household tasks.
It's a deeply ambitious Jetsons-type future project that Musk and fellow investors are known for, although that's no guarantee of results. Abbeel has a good track record with helping robots learn—last year, he and his team from Berkeley got a robot to put a clothes hanger on a rack, assemble a toy plane, and screw a cap on a water bottle without any preprogrammed instructions.
The key, he said in a university press release, was that "when a robot is faced with something new, we won't have to reprogram it. The exact same software, which encodes how the robot can learn, was used to allow the robot to learn all the different tasks we gave it." It's no Rosie yet, but BRETT—the Berkeley Robot for the Elimination of Tedious Tasks—is getting there.
If Abbeel's methods really take off, the bulk of robotic programming could soon be a thing of the past. Especially considering that OpenAI is far from the only company on the case. Microsoft has recently begun training AI using an experimental version of its game Minecraft. The trick is just making sure that these robots don't learn anything that will end the human race.
Source: MIT Technology Review
David Grossman is a staff writer for PopularMechanics.com. He's previously written for The Verge, Rolling Stone, The New Republic and several other publications. He's based out of Brooklyn.
