According to the robotics community, it’s unlikely that any robot now on the market could hold your attention for more than 10 hours. (Actually, if you have a robot dog gathering dust on a closet shelf , you probably already know that.)
A new study, however, indicates that this threshold is poised to be broken—at least if the humans interacting with the machines are youngsters. Researchers found that a two-foot- (61-centimeter) tall metal man easily won over a classroom of tykes, aged 18 to 24 months, who intermittently spent time with it over a five-month period.
“Our results suggest that current robot technology is surprisingly close to achieving autonomous bonding and socialization with human toddlers for significant periods of time,” University of California, San Diego, researchers report in Proceedings of the National Academy of Sciences USA.
QRIO, a robot programmed with a slew of social functions, was placed in U.C. San Diego’s Early Childhood Education Center 45 times over the five-month observation period. For the first 27 sessions, the robot was allowed access to its full arsenal of programmed social behaviors. In addition, a controller could send commands to the humanoid, prompting it to wave, dance, sit, stand, etcetera (although there was a lag time between the prompt and when the robot made the movement).
photo credit: kaioshinThe tots began to increasingly interact with the robot and treat it more like a peer than an object during the first 11 sessions. The level of social activity increased dramatically when researchers added a new behavior to QRIO’s repertoire: If a child touched the humanoid on its head, it would make a giggling noise.
“The contingency coupled with the positive reaction of giggling made clear to the children that the robot was responsive to them and served often to initiate interaction episodes,” says study co-author Fumihide Tanaka, a researcher at U.C. San Diego’s Institute for Neural Computation and at Sony Intelligence Dynamics Laboratories, Inc.
For 15 sessions midway through the experiment, QRIO was programmed to repeatedly dance to the same song rather than interact with the kids. During these trials, the children became far less interested in the friendly automaton. For the final three sessions, however, QRIO could once again unleash its entire social arsenal.
Tanaka and his colleagues scored the quality of social interaction primarily based on where children touched the robot. A teddy bear and an inanimate toy robot named Robby accompanied QRIO during most of the observation period. The teddy bear was introduced first and prior to the introduction of the robots was very popular. But the stuffed animal was lost in the shuffle when QRIO and Robby came on the scene. Though the toddlers often manhandled Robby, they eventually began touching QRIO in a pattern similar to the way they touched one another—mostly on its arms and hands.
The only time they deviated from this behavior was when QRIO was programmed to giggle, at which point they frequently petted its face and head. Another indication that the little humans viewed robo-kid as a compeer was the way they reacted when QRIO ran out of juice and lay down as if to take a nap: Some of the children would try to wake and help it up, whereas others would cover it with a blanket.
“Our work suggests that touch integrated on the time-scale of a few minutes is a surprisingly effective index of social connectedness,” Tanaka says. “Something akin to this index may be used by the human brain to evaluate its own sense of social well-being.” He adds that social robots like QRIO could greatly enrich classrooms and assist teachers in early learning programs.
By Nikhil Swaminathan
The thing observed in this article points out that we can use robots as information gathers we could gather world knowledge this way through the interactions with toddlers surely this will pay off over time? As the toddlers grow so will the program through its various manifestations and soon the component of world knowledge will increase. Now this is vital for the success of any program all humans come into this world with a brain their gene program decides that this is going to be different from a monkey and that it’s going to be along human lines in brain design space (which could be imagined as a huge room stretching off to infinity which is the sum representation of all the brains that can be made in the observable 3 dimensions, we are so similar that we occupy just a point in this infinite space, a frikkin’ point). Next, something wonderful happens, we are born and we interact, we cry, we laugh, we learn. This world knowledge form the basis of the instinctive “common sense” seen in humans we know water is wet because as a baby we touched it and the brain recorded the sensation and maybe our parents told us that you are getting wet, so that word was associated with that feeling.
We are basically bottom-up systems we learn and we interact and then learn some more. This positive feedback loop is something, I feel, we need to mimic in order to create machines that are somewhat intelligent. Let’s just throw the complex symbolic reasoning all the complex mathematics that are used to simulate creatures out of the window and let’s just create a child, a one day old child and let it learn. We could provide the foundation of this by putting in facilities like reasoning through bayesian systems, how the memory storage should be done, basically reverse engineer the brain, like kurzweil says, and put whatever is in it and let it loose. Then we wait and observe and let it interact to see what happens over time. Does the system grow? Is any knowledge gathered through sensory feedback used? So on and so forth, in this scenario we only maintain the basic mathematics, pattern recognition algorithms so and so forth but nothing else, will it work?