Researchers from UC Berkeley have found a way to help robots obtain better motor tasks using a trial and error process, in a similar manner to how humans learn. The reinforcement learning technique is possible because of software algorithms that give robots a new ability to learn from previous mistakes.
The robots complete different tasks, such as screwing a cap on a water bottle, putting a clothes hanger on a rack, and other tasks without the need of pre-programmed details. Deep learning will continue to be a major research focus in artificial intelligence (AI) development, as the Willow Garage Personal Robot 2 used at UC Berkeley continues perfecting its motor tasks.
"It used to take hours on up to months of careful programming to give a robot the hand-eye coordination necessary to do a task," said Gary Bradski, founder of OpenCV, which provides machine vision software, in a statement published by the New York Times. "This new work enables robots to just learn the task by doing it."