Teaching Robots to Learn

Intelligent robots that can learn from their environment, and figure out how to execute novel complex tasks are looking like a very real possibility. Researchers from Iowa State University are working on software that will help robots learn at the level of a two year old child.

Scientific American explains:

In one set of experiments, the robot was presented with 36 different objects, including hockey pucks and Tupperware. It could perform five different actions with each one—grasping, pushing, tapping, shaking and dropping—and had to identify and classify them based only on the sounds they made. After just one action the robot had a 72 percent success rate, but its accuracy soared with each successive action, reaching 99.2 percent after all five. The robot had learned to use a perceptual model to recognize and classify objects—and it could rely on this model to estimate how similar two objects were with only the sounds they made to guide it.

Intel’s HERB, (the “Home Exploring Robotic Butler”)part of the company’s Personal Robotics Project, is also making impressive progress in learning new skills.

In the meantime, learning robots will sometimes surprise their own creators. Srinivasa tells how an early version of HERB puzzled researchers when it was grabbing coffee cups to place them in a dishwasher rack. It used a strange hand position, with one of its “thumbs” pointing down. Then they realized this was a “far more efficient motion” used by professional bartenders, Srinivasa says:  “They lift from underneath and pour in a single smooth motion, like in the movie Cocktail.”  He calls these surprises “one of joys of doing manipulation research.”

Scientific American: “Can Robots Be Programmed to Learn from Their Own Experiences?”