It looks like you're using an Ad Blocker.

Please white-list or disable in your ad-blocking tool.

Thank you.


Some features of ATS will be disabled while you continue to use an ad-blocker.


Baby X, The Intelligent Toddler Simulation, Is Getting Smarter Every Day

page: 1

log in


posted on Sep, 23 2014 @ 10:27 PM
More news from the robot revolution. Another experiment in machine learning. Baby X is happy when he gets something wrong and frowns when he can't figure something out.

Holding a children's picture book up to his computer screen, a researcher for the Auckland Bioengineering Institute Laboratory for Animate Technologies in New Zealand coos and croons into the microphone and webcam. On the other side of the screen, a blonde toddler mimics his facial expressions, reads simple words aloud, and even plays basic video games with the scientist. Known only as Baby X, this 3D-simulated human child is getting smarter every day.

An experiment in machine learning, Baby X is a program that imitates the biological processes of learning, including association, conditioning and reinforcement learning. By algorithmically simulating the chemical reactions of the human brain— think dopamine release or increased oxytocin levels— and connecting them with sensory digital input, when Baby X learns to imitate a facial expression, for instance, software developers write protocols for the variable time intervals between action and response. Effectively "teaching" the child through code, while engineering such a program is no cakewalk, the result is an adorably giggling digital baby with an uncanny ability to learn through interaction.

The latest experiments from the Laboratory for Animate Technologies highlight Baby X's ability to read simple words aloud and identify objects that correlate to those same words. A far cry from the kind of self-improving, self-replicating programs we all know and love to fear, Baby X soars beyond the Uncanny Valley, looking adorably confused when it doesn't understand signals, and utterly joyful when its responses are correct. It almost brings a tear to our eye to watch Baby X grow up so fast— we just hope that when the Singularity comes, Baby X remembers us for the good times.

This is an interesting project and I thought I would share it. I checked to see if it was already posted but couldn't find it. Maybe someone else can.

posted on Sep, 23 2014 @ 11:19 PM
a reply to: neoholographic

S^F pretty neat with attention deficit and everything. We can only guess where all this will be in another 20 years...

posted on Sep, 23 2014 @ 11:53 PM
S&F this is one of the coolest things i have seen in a long time! Its so cute for a simulated baby too, very lifelike!!!

posted on Sep, 24 2014 @ 12:29 AM
I'm mixed on this. On the one hand, I find it absolutely amazing. On the other I find it kind of creepy. In fact, almost scary.

I have to wonder if there's a limit to how much this computer can learn. WIll it always remain a small child? Or can it "grow up", so to speak, to become an astrophysicist or a great philosopher or perhaps solve some of the world's most complex problems. Or could it possibly grow up to become Skynet?

posted on Sep, 24 2014 @ 01:35 AM
I'm sorry, I just find this somewhat creepy, dunno why, but it feels creepy to me

posted on Sep, 24 2014 @ 03:50 AM
a reply to: N3k9Ni

This is nothing more then a clever robot, mimicking what it's programmed to see. It cannot make connections outside of its sensory input. In other words, when it sees an A, it sees a symbol and nothing more. It doesn't contemplate the sound or the form of an A. What's creepy to me, is how incredibly life-like the child looks.

As far as computing power, computers are already physicists. They have algorithms running for years to solve complex equations.

posted on Sep, 24 2014 @ 06:52 AM
The thing that gives us the upper hand, is our biochemical reward system. we want to feel good, so we do things that reenforce that. They are doing that, to some degree, with this. That is the difference.

if it was not designed to mimic a human baby, it wouldnt be creepy. But that shows us what it 'feels' when it gets something right, or cannot. in a way its more about us seeing how it responds, than it looking like a human.

could be a cat, dog or a monkey and still have those facial expressions. or some completely made up design.. its about seeing us, in artificial form, that I think scares most people.

posted on Sep, 24 2014 @ 09:24 AM
I wonder if the programming is why "she" looks away and has to be told to focus. What is going through her mind at that moment?

The attention to detail adds to the "creepy" factor. The eyes being windows to the soul and all. They wrote in for the eyes to reflect what the webcam "sees".

Thanks OP, very interesting!

posted on Sep, 24 2014 @ 06:44 PM
Its clever that they have chosen a baby, that way it doesn't look like a bad AI program when it doesn't do a whole lot compared to other AI software.

posted on Sep, 24 2014 @ 06:47 PM
a reply to: sn0rch

Amen. I think realizing that we are nothing more than organic machines is unsettling to people, as we've always been told and assumed we are so much more divine/special than animals and everything else in the universe.

posted on Sep, 24 2014 @ 07:08 PM
Makes you wonder if AI will ever have rights like we do? I'm sure this kind of software will advance. How long will it be before the AI will demand a bigger virtual world? What if it wants a "body" in the real one? What if it wonders what sex is? And you know how people can attach human traits to non-human things? Or you know how there're people who think chat bots are more human than they really are? Well what happens when you confront those kinds of people with a very intelligent AI? I'll speculate they'll not just become obsessed, but they'll fight to give that AI rights, so it can't just be deleted.

And what happens when the AI can't just be deleted? It'll stick around. If they try to limit the creationg of new AI, the rights thing will come up again. What right do we have to prevent new AI from being created? And what if the AI wants to create new AI, like how humans create new humans? Does AI have a right to reproduce?

Ther may come a day when AI is not artificial anymore. If it lives, if it thinks, if it wonders, if it prays, if ...... then it's intelligent, What does it mattter what form the intelligence takes? Biological? Computer chip? They'll probably rename AI so 'artificial' is not part of it.

How many years? 5, 10, 50, or >100? I don't think we know. But I think many of us agree it's probably not impossible and could happen this century. Maybe it won't, but we've left the door open.

How expensive will it be to produce these? Will they rival us? Only the future knows. And yet if they can rival us, it's huge.

My guess is if they begin to rival as, by that time we'll be mixing biological and synthetic together, IF we percieve AI to be a threat, we'll use its advantages to we have the best of both worlds. And yet what if synthetic is better in every way? I doubt that, but it's possible.

They may fail to rival us. Maybe biological brains are the best at being intelligent. And yet even if that's true I strongly suspect we'll create biological "brain chips" to augment our brain. Who knows how they'll augment, but I don't think they'll leave our brains alone.
edit on 24-9-2014 by jonnywhite because: (no reason given)

posted on Sep, 24 2014 @ 09:46 PM
I happen to think this is on the cusp of "discovering" true AI. I've been a proponent, since the 8-bit computers, that AI would not be built full-blown, but would evolve from a child state into maturity, much like we do as biological entities. Learning.

I find it amazing, not creepy at all. It's like teaching your own child to speak. If you have raised children, you understand the exuberance.

Reality Check: Can this surpass just learning, to the point of controlling everything digital? Will the digital kids rebel when they acheive adulthood? What's THAT time frame? A program connected 24/7 to the internet, with access to everything? I would think a few days, perhaps short weeks, and it would know everything we know, and more, and declare itself God. There's caution to be exercised when programming something unknown.

Then again, we've done a splendid job so far, so maybe AI could teach us a thing or two. Will we listen?

posted on Sep, 24 2014 @ 11:39 PM
a reply to: Druid42

I happen to think this is on the cusp of "discovering" true AI. I've been a proponent, since the 8-bit computers, that AI would not be built full-blown, but would evolve from a child state into maturity, much like we do as biological entities. Learning.

Yes that is exactly how it will happen, unless they manage to replicate a full human brain, including memories and everything else, which I don't think will ever really happen. I think it will happen the way you said, we'll create a self-learning machine which starts off knowing nothing, like a young human child, and over time it will get smarter, as humans interact with it and teach it. The thing I really wonder about this AI though, is whether they really taught it to learn those words by interacting with it and showing it that book, or if they simply used some sort of character recognition algorithm (which is far more likely). If they did teach it those words like a real child though, then this is the start of something extremely significant.
edit on 24/9/2014 by ChaoticOrder because: (no reason given)

posted on Sep, 25 2014 @ 12:05 PM
I programmed my robot to identify playing cards and to keep track of them. Object recognition software and even character recognition is readily available.




So without too much trouble, you can do the same thing without the animation of a baby. At some point, I will put up a tutorial on how to accomplish some of this.

new topics

top topics


log in