MIT's Furry Robot Is Capable Of Human Learning

For some reason I can't look at MIT's Leonardo robot without an involuntary shudder - I think it's the lifelike fur and the evil, calculating eyes.  Even scarier is to see it in motion (and there are videos after the cut), when complex facial mapping techniques have taken human expressions and reworked them for the robot's face (which has 32 degrees of freedom).

Of course the clever people at MIT aren't just making a particularly impressive toy; their intent is to research new ways of human/machine interaction, where robots are used to give a computer a naturalistic interface that users are comfortable working with. 

"Once a task is learned, the robot should then be competent in its ability to provide assistance; understanding how to perform the task as well as how to perform it in partnership with a human"

Capable of being instructed by tutelage, where a user guides Leonardo to complete a task, by imitation, where Leonardo copies a demonstration, and social referencing, where Leonardo responds to a users reaction to a task, their goal is to end up with a machine that can learn from humans in their own terms, rather than them needing to be experts in programming.

[googlevideo]http://video.google.co.uk/videoplay?docid=7599178959058990307[/googlevideo]

[googlevideo]http://video.google.co.uk/videoplay?docid=2677188667049278109[/googlevideo]

I still think it looks like an evil squirrel monster.

MIT Leonardo [via Hacked Gadgets]