Without human intuition and our knowledge of goals and prerequisite actions acquired through years of experience, learning new tasks is difficult for robots, and considerable human effort goes into training them. But what if robots could learn from other robots?
Will Knight of MIT Technology Review interviews Assistant Professor Stefanie Tellex of Brown University's Department of Computer Science and Professor Ashutosh Saxena of Cornell University about their RoboBrain collaboration, which allowed one robot to teach another how to pick up mugs from a table and place them on overturned bowls. The story was also featured on Slashdot.
"This is really a baby step," says Tellex, toward the world in which a robot can translate information for itself by comparing its physical form to that of another robot.
The entire article is available here.