By Laura Shin
Posting in Environment
One of the main reasons humans are different from robots is that we can recognize ourselves in a mirror. One robot begins to break down that barrier.
The fact that we can recognize ourselves in a mirror.
But one robot is starting to break down that wall with some preliminary achievements that foretell a day when robots may recognize themselves.
Basically, the robot, who is named Nico, is pictured above and hails from Yale, can look at things in a mirror and recognize that what it sees gives it information about its own environment.
The paper describing its feat, "Mirror Perspective-Taking With a Humanoid Robot," by Yale PhD candidate Justin Hart and his advisor Brian Scasselati, was presented last month at the Conference on Artificial Intelligence in Toronto, Canada.
Nico's great achievement is to do something humans do all the time in the presence of a mirror: To look at it and infer information about the room we are in. So, for instance, if there is a lamp to the right of you, you will see it in the mirror, and you could reach out and touch it without taking your eyes off the mirror.
Nico's achievement was a bit similar. It is able to look in the mirror and see its hand, which has a visual token attached. Learning from its view of its hand, it then can estimate where other objects in the room are, without needing to look at the actual room.
Many animals can't do this, because they can't understand what mirrors are. That's why animals often look in mirrors and think other animals are approaching them. As Hart explained in an email to NBC News:
What makes this exciting, in terms of self-awareness, is that the robot is able to use this knowledge that it has learned about itself in order to reason about a thing in its environment, the mirror, in a way that robots really haven't been able to do before.
The holy grail of robot self-awareness would be for a robot to recognize itself in a mirror -- a test passed only by a few animals, such as dolphins, elephants and chimps.
If a robot could do this, it could be made much more useful in everyday life. Its awareness of its own body in space could enable it to "live" in people's homes without causing harm to the furniture of other housemates.
Hart plans to put Nico through this more complex mirror test in the next few months.
Related on SmartPlanet:
- Video: Rats works to free caged mates, showing empathy
- Coming soon: Robots in the sky that recognize and track you
- The future of manual labor: no people, just robots?
- Small, speedy robots zip, roll and swarm through the air
- A robot that uses whiskers to get around
- This robot can jump on water
- Video: Scientists figure out mechanics of wet-dog shake
photo: Nico. (Justin Hart/Yale University)
Aug 30, 2012
When a robot sees itself in the mirror, would it think (A) what am I seeing? (B)is that another robot? Or (C), is that me? First of all it needs to be aware of itself, without a mirror. Personally I'm not happy about robots thinking for themselves. It is a lot safer if they are told what to do and do only that. Autonomous drones come to mind, roaming the skies, looking for cars to shoot at.I'm not just thinking Afghanistan or Pakistan. These drones will shortly be filling the skies in America, looking for dissidents (i.e. anybody who doesn't agree with the government)
Really, I'd rather just adopt an elephant - and leave it where it lives. We don't need robots that are that are sooner or later going to succumb to existential crises. Sartre and Blade Runner. Please. It will just be a mess.
how do we no it wasn't programed to recognise a printed/programed in shape? I am quite happy with the AI personalities and human brain backup, as I'm working on AI law:) there is a paper on AI Criminal Liability :)