A robotic hand can juggle a cube - with lots of training
How long does it take a robotic hand to learn to juggle a cube? About 100 years, give or take. That’s how much virtual computing time it took researchers at OpenAI, the non-profit artificial intelligence lab funded by Elon Musk and others, to train its disembodied hand. The team paid Google $3,500 to run its software on thousands of computers simultaneously, crunching the actual time to 48 hours. After training the robot in a virtual environment, the team put it to a test in the real world.
The hand, called Dactyl, learned to move itself, the team of two dozen researchers disclosed this week. Its job is simply to adjust the cube so that one of its letters - “O,” ”P,“”E,“”N,“”A“or ”I” - faces upward to match a random selection.
Ken Goldberg, a University of This undated photo provided by OpenAI shows a robotic hand holding a cube at the company’s research lab in San Francisco.
California, Berkeley robotics professor called Dex-Net, though its hand who isn’t affiliated with the is simpler and the objects it manipulates project, said OpenAI’s achievement are more complex. is a big deal because it demonstrates “The key is the idea that you how robots trained in a can make so much progress in virtual environment can operate simulation,” he said. in the real world. His lab is trying “This is a plausible path forward, something similar with a robot when doing physical experiments is very hard.”
Dactyl’s real-world fingers are tracked by infrared dots and cameras. In training, every simulated movement that brought the cube closer to the goal gave Dactyl a small reward. Dropping the cube caused it to feel a penalty 20 times as big.
The process is called reinforcement learning. The robot software repeats the attempts millions of times in a simulated environment, trying over and over to get the highest reward. OpenAI used roughly the same algorithm it used to beat human players in a video game, “Dota 2.”
In real life, a team of researchers worked about a year to get the mechanical hand to this point.
Why? For one, the hand in a simulated environment doesn’t understand friction. So even though its real fingers are rubbery, Dactyl lacks human understanding about the best grips.