Calendar
July 2009
M T W T F S S
« Jun   Aug »
 12345
6789101112
13141516171819
20212223242526
2728293031  
Categories

Einstein Look-alike Robot Teaches Itself to Smile By Looking In Mirror - UCSD

A robot, designed by researchers at UCSD, can learn new facial expressions simply by monitoring itself in a mirror, a strange, and slightly eery, step towards computers and robots that learn on their own.

Courtesy of UCSD

The Einstein-like robot was first shown last February at the TED Conference last February. Now, the inventors have taken the next step by programming the robot to ‘learn’ to fine tune its own facial expressions via self-monitoring. The robot is designed by Hanson Robotics. Prior to this upgrade, the robot could demonstrate only the 31 preprogrammed expressions. Now the robot can tweak, modulate and improve upon those preprogrammed expressions creating a nearly limitless facial repertoire. Similar to the movie title ‘Flubber’, the material which the robot’s skin is made of is called ’Frubber’.

The idea was derived from the way in which babies learn vocalizations and expressions. Rather than relying on preprogramming to elicit it’s certain facial expressions, the UCSD robot uses trial and error while getting feedback of its own expressions from a mirror, and thereby slowly learns how certain micromovements lead to full emotional expressions.   The UCSD researchers presented a paper on the momentous feat last month at the 2009 IEEE Conference on Development and Learning.

The press release from UCSD states,

‘Once the robot learned the relationship between facial expressions and the muscle movements required to make them, the robot learned to make facial expressions it had never encountered.’

Creepy, but impressive, nonetheless. I’m not sure how comfortable I feel with the convergence of nanotechnology, emotionally expressive robots, artificial intelligence, cloning technology and the implantation of electronics into humans. It seems quite plausible, as many have argued, that a quick learning, human-like robot could be developed in the next 25 years. The problems seem to start when and if that robot becomes ’self-aware’ and gets a chance to self-replicate. Perhaps that’s needeless worrying, but it seems common sense to me to proceed with caution and mindfulness in these areas.

Have a wonderful week,

 

Dr. John Schinnerer

Guide To Self, Inc.

 

Leave a Reply

You must be logged in to post a comment.