Leonardo, the creation of MIT scientists, looks like a cuddly toy with movable appendages. But this "toy" is, in at least one way, more intelligent than a human pre-schooler.
Our Guest Blogger, Lisa Zyga, is a science writer who is interested in many areas of science and technology, including the impact of science on society. She is excited about exploring new inventions in science, medicine, and space, and inspiring and fascinating readers interested in the creative frontier of science with the readers of InventorSpot.com. Here's her article: * * * * *
Two robots, Leonardo and Nico, are the first of their kind to demonstrate some upper level skills...which begs the question, how exactly do you measure a robot's abilities?
Leonardo, the creation of MIT scientists Cynthia Breazeal, Matt Berlin, and Jesse Gray, looks like a cuddly toy with movable appendages. But this "toy" is, in at least one way, more intelligent than a human pre-schooler.
Psychologists know that before the age of four or five, humans can't put themselves in another person's shoes, or see the world from someone else's perspective.
To test this, psychologists have made children watch as a boy puts a candy bar in a drawer and leaves the room. Then, an adult enters the room and moves the candy to a cupboard. The children are then asked where they think the boy will look for the candy when he comes back in the room. Younger children predict the boy will look where the adult moved the candy, while four- and five-year-olds understand the boy still thinks the candy is in the drawer.
What does Leonardo think? According to the scientists, he can understand when someone else has false knowledge, and knows that someone outside the room can't know what's going on inside the room. He predicts the drawer. He passes this so-called "false-belief test."
Leonardo's "brain" is an array of computers and software programs that allow him to recognize voices, images, and faces. In his brain, he stores a list of objects and people he sees, constantly updating his list.
In addition, for every new face he sees, Leonardo can create a separate "brain" for them. In this brain, he stores only information that that person knows. The robot only adds to these brains when he observes those people doing or seeing something.
In a way, Leonardo's system sounds straightforward (aside from all the technical details that undoubtedly go into actually making it a reality). But the premise sounds simple. What's strange about this is that children's brains lack the ability to do this task that we adults do many times a day, unconsciously. And yet eventually, they naturally learn it.
The point I'm making is that human brains are wired to evolve, without any kind of special exterior influence (besides daily interactions with the world). But for Leonardo-and any other robot-the brain is not wired to evolve. It can add to its list of knowledge, and perhaps gather even more information than we can, but its basic configuration is still the same.
So referring to my question in the intro, "how do you measure a robot's abilities?", I wonder if, just because abilities appear the same on the outside, if you can really say that the robot has mastered a skill. After all, it was directly programmed to do this. I would look at it more as a sophisticated piece of software, and I'd applaud the engineers for their computer programming.
Can you measure a robot's abilities in the same way that you'd measure a human's abilities-as this example measured a type of cognitive ability? Just because my dog can shake when you hold out your hand to him, doesn't mean he understands American customs or the importance of polite etiquette. We've just wired his brain in a way that only works if you're a dog: cookies.
Likewise, the way the scientists wired Leonardo to know others' minds (with electronics) would be a means to the end, but only for a robot, right?
There seems to be a gap between the result and the meaning of the result, and I can offer no solution on how to bridge the gap.
Then again, I wonder if when the day comes when they make a robot that can fully function on its own-and even evolve-if my concerns will be irrelevant.
We'll find out more about another robot, Nico, in my next article. (via NewScientistTech )