Babies learn by watching and imitating what adults are doing, and robots can "learn" in much the same way, according to computer scientists and developmental psychologists at the University of Washington (UW).
The researchers have demonstrated robots can learn by amassing data through exploration, watching humans do something and determining how to perform that task on their own.
"You can look at this as a first step in building robots that can learn from humans in the same way that infants learn from humans," says UW professor Rajesh Rao.
The team used research on how infants follow an adult's gaze to develop machine-learning algorithms that enable a robot to explore how its own actions result in different outcomes. The researchers' robot used this learned probabilistic model to infer what a human wants it to do and complete the task, and even "ask" for help if it is not certain that it can.
The researchers tested the robotic model in a computer simulation experiment in which a robot learned to follow a human's gaze. In another experiment, an actual robot imitated a human moving toy food objects around a tabletop.
From UW Today
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found