backtop


Print 31 comment(s) - last by Alexvrb.. on Oct 17 at 11:28 AM


Morphy the robot  (Source: University of Washington)
UW Researchers Study How Babies Perceive Robots

University of Washington researchers are studying what makes babies decide the difference between sentient and inanimate objects, and also how they interact and learn from these objects, such as robots. 

Babies are curious and interested in almost everything their parents do. They learn and socialize by mimicking an adult's actions. For instance, if an adult touches their nose to teach the baby where their nose is, the baby will learn to touch their nose as well when asked where it is. 

Andrew Meltzoff, lead author of the study and co-director of of the University of Washington's Institute for Learning and Brain Sciences; Rechele Brooks, co-author of the study and a UW research assistant professor; and Rajesh Rao, co-author and UW associate professor of computer science and engineering, have studied babies' interactions with adults and inanimate objects in an effort to understand how they perceive and learn from different "teachers." 

"Babies learn best through social interactions, but what makes something 'social' for a baby?" said Meltzoff. "It is not just what something looks like, but how it moves and interacts with others that gives it special meaning to the baby."

To further investigate this question, 64 babies were used in the study to observe certain interactions. All were around the age of 18-months-old, and were allowed to play with toys for awhile in order to become comfortable with the experimental setting. Once the babies were adjusted to this setting, Brooks brought out a "metallic humanoid robot" with a torso, legs, arms, head and eyes that were actually camera lenses. The robots name was Morphy, and was controlled by a researcher who was hidden from the babies' view. 

Brooks then followed a script where she interacted with Morphy, asking the robot questions like "Where is your tummy?" and "Where is your head?" Morphy would then act accordingly, pointing to its head and tummy. 

After 90 minutes of interaction, Brooks left the room while Morphy continued to "move on its own," and researchers measured if the babies thought Morphy was real or not. Morphy continued to make subtle movements and sounds to hold the babies' attention, and results showed that 13 out of 16 babies would follow Morphy's gaze when it would look at a nearby toy or make a subtle movement. For the control group of babies who did not witness Brooks playing with the robot first, only three out of 16 babies followed the Morphy's movements. 

"We are using modern technology to explore an age-old question about the essence of being human," said Meltzoff. "The babies are telling us that communication with other people is a fundamental feature of being human."

This study was published in Neural Networks



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

16 + 16 = 64?
By Shining Arcanine on 10/14/2010 8:23:31 PM , Rating: 2
I think that the numbers are a bit off. The experimental group had 16 and the control group had 16, yet 64 participated? What happened?




RE: 16 + 16 = 64?
By iceonfire1 on 10/14/2010 8:32:13 PM , Rating: 2
Clearly, this is the axiom of choice at work. The researchers wanted to improve their sample size, so they simply rearranged the babies to form 64 of them from the original 32.


RE: 16 + 16 = 64?
By Camikazi on 10/14/2010 9:03:03 PM , Rating: 3
Sounds illegal and very messy to me...


RE: 16 + 16 = 64?
By Visual on 10/15/2010 7:17:55 AM , Rating: 5
It might just be a result of a robot in development trying to touch the babies' noses. It takes trial and error before success.

Now excuse me while I go read up on some dead babies jokes.


RE: 16 + 16 = 64?
By nstott on 10/15/2010 10:21:12 AM , Rating: 4
You're all missing the important point here: the stimulus money used to fund this study created or saved 64 jobs! Thus, 16 + 16 does indeed equal 64.


RE: 16 + 16 = 64?
By Alexvrb on 10/17/2010 11:28:14 AM , Rating: 2
Those jobs also then went on to indirectly create or save another 512 supporting jobs in the green sector, too! After all, the robot was windsolar powered and was constructed entirely from recycled cell phones.


RE: 16 + 16 = 64?
By SpaceOddity85 on 10/14/10, Rating: 0
RE: 16 + 16 = 64?
By Tiffany Kaiser on 10/15/2010 12:03:56 AM , Rating: 3
The abstract notes that there are four groups being tested. There are a total of 64 babies, so that would mean 16 babies per group. Here's the study's layout, hope this helps:

http://www.sciencedirect.com/science?_ob=ArticleUR...


RE: 16 + 16 = 64?
By sviola on 10/15/2010 11:02:23 AM , Rating: 2
quote:
The abstract notes that there are four groups being tested. There are a total of 64 babies, so that would mean 16 babies per group.


You should have added more information to the article (as there were different type of interactions with the 4 groups)


RE: 16 + 16 = 64?
By Siki on 10/17/10, Rating: 0
RE: 16 + 16 = 64?
By amanojaku on 10/15/2010 12:07:38 AM , Rating: 3
2.4.1. Phase 1: scripted robot experience
2.4.1.1. Group 1 (social interaction)
2.4.1.2. Group 2 (robot movement, passive adult)
2.4.1.3. Group 3 (robot–adult mismatch)
2.4.1.4. Group 4 (passive robot baseline)

What's unclear is whether there were 64 unique babies or 16 babies reused 4 times.


RE: 16 + 16 = 64?
By MozeeToby on 10/15/2010 10:33:57 AM , Rating: 2
It would have to be 64 unique babies, unless the scientists really didn't care about the research in the slightest. Otherwise the subjects (in this case, the babies) would have preconceived notions about the robot before 2-4th trials began, notions that were created in the earlier part of the experiment. They could easily remember that the adult already interacted with the robot earlier and thus throw off the results.


RE: 16 + 16 = 64?
By Jjoshua2 on 10/15/2010 3:43:53 AM , Rating: 2
That is awesome! I took a class from Professor Raj last quarter, and saw him today at the Steve Ballmer presentation. I could ask him, but there's no use wasting his time on such a trivial question.


RE: 16 + 16 = 64?
By tastyratz on 10/15/2010 11:01:56 AM , Rating: 2
I blame it on the unions... Damn United Baby Workers


Why some people post.
By Thelookingglass on 10/15/2010 12:36:39 AM , Rating: 2
Its easy to read through a decent article and look past the errors, fill in the blanks, or, through looking at context, understand what is being said.

What life do people have on here always pointing out grammer mistakes, slips of the tongue(fingers I suppose), or strictly reading articles to find errors?

Get a life. Get a job. Stop posting. No one cares. The articles here are well written and span a variety of topics. Anandtech.com is awesome.




RE: Why some people post.
By CrazyBernie on 10/15/2010 12:44:59 AM , Rating: 4
Dude, you might as well tell the sky to stop being blue.


RE: Why some people post.
By geddarkstorm on 10/15/2010 2:09:19 PM , Rating: 2
Or tell vampires to stop sparkling.


RE: Why some people post.
By AstroGuardian on 10/15/10, Rating: 0
RE: Why some people post.
By Murloc on 10/15/2010 7:27:09 AM , Rating: 3
this is not anandtech.
Anandtech articles only speak about computers and are specialized, of course, but they are 1000x better.


RE: Why some people post.
By GaryJohnson on 10/15/2010 9:57:28 AM , Rating: 5
This site used to be better. Kristopher Kubicki used to write awesome articles here.


RE: Why some people post.
By tmouse on 10/15/2010 8:16:01 AM , Rating: 4
Well if you like being a droid and mindlessly lap up everything you read, that’s fine. Scientific reading mandates being critical, real reporting mandates accurate interpretation and clear, concise communication. Asking how one goes from 64 members to 32 is a perfectly valid question. The writer should make these things clear. Most of the articles here have inexcusable errors in them. The problems we have today stem from the basic laziness of many “news writers”. They are nothing more than just parrots, regurgitating others reports. It certainly seem that most of the writers do not even bother to read the original source material of the story they are “reporting” on, they just do a little rearranging of someone else’s report which may also be second hand or worse and often pass along other reporters interpretations as “facts” or quotes from the original articles. There seems to be no proof reading and many come awfully close to plagiarism of the source material. Certainly some criticisms lean more to personal attacks but many are valid. In this day and age there are minimal reasons to allow some of these types of errors to get into print, it’s just sloppy.


RE: Why some people post.
By Denigrate on 10/15/2010 8:50:11 AM , Rating: 4
Exactly. There are far to many errors in most of the articles by this writer. Some are obviously intentional to attempt to drive readers to her point.

This article is better than most of her work, and I think she is improving. Sucks for DT readers that she is learing on the job, but at least there seems to be a glimmer of hope.


Really?
By TheSev on 10/15/2010 7:44:08 AM , Rating: 3
I'm sorry, but does anyone think that this is one of the (not the only) most pointless studies a group can do? I can't see any useful or beneficial result from this.




RE: Really?
By Paj on 10/15/2010 9:00:23 AM , Rating: 2
I disagree. As robots become more powerful and commonplace, learning how to improve the way they interact with humans is becoming increasingly important. Based on the outcomes of this study, designers will have useful information for designing future robots that perform better when interacting with humans.


RE: Really?
By Suntan on 10/15/2010 12:24:22 PM , Rating: 2
I would agree. At least, agree that it is pointless due to the fact that they are still just jumping to conclusions.

quote:
"We are using modern technology to explore an age-old question about the essence of being human," said Meltzoff. "The babies are telling us that communication with other people is a fundamental feature of being human."


How do they come to this conclusion? The conclusion I come to, given the facts in the article, is that children are more inclined to play with something if they see adults playing with it first.

This pretty much lines up with what every parent has ever experienced. Both my young kids have a big attraction to the remote control and the cell phones because they have seen Mommy and Daddy use them... does that make the remote or the phone “human-like.”

-Suntan


RE: Really?
By Anoxanmore on 10/15/2010 12:47:16 PM , Rating: 2
So, you are saying the robot they used was the substitute for an actual adult?

That made the study worth while all by itself.


RE: Really?
By Suntan on 10/15/2010 1:00:40 PM , Rating: 2
No. The article says that the tests showed that kids who witnessed a real adult interacting with the robot were more inclined to interact with and mimic the robot after the adult left the room. Which is pretty much what you would expect because kids that age take interest in anything that some other person shows interest in.

The failure of the experiement is that there is nothing related to the kids "thinking the robot is alive," and that they are jumping to an (unfounded) conclusion of such.

-Suntan


I don't need a study to tell me this.
By JonnyDough on 10/15/2010 4:44:20 AM , Rating: 2
If a robot was your only caretaker/friend as you grew you would see it as family. A good robot mimicks people. Just as pets try to mimick our behaviors, we too are animals that will mimick the behaviors of what we percieve as our elder beings. Its inherent. Ever read the Jungle Book or hear stories of men raised by wolves? It happens you know. Both monkeys and elephants have been known to take care of kittens and they aren't the only mammals to do this. Adoption doesn't just happen with mammals either I don't believe.




RE: I don't need a study to tell me this.
By nstott on 10/15/2010 10:17:32 AM , Rating: 2
quote:
I don't need a study to tell me this.

That's because you were raised by robots, which explains a lot about you.


By JonnyDough on 10/15/2010 12:27:31 PM , Rating: 2
Not just any robots, really smart and fashionable ones.


It all depends...
By mindless1 on 10/16/2010 4:48:23 PM , Rating: 2
... on how lifelike the robot is. The real question is can researchers tell robot babies from real ones, if not this study is tainted from the start.




"Spreading the rumors, it's very easy because the people who write about Apple want that story, and you can claim its credible because you spoke to someone at Apple." -- Investment guru Jim Cramer














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki