backtop


Print 18 comment(s) - last by oTAL.. on Mar 1 at 9:28 AM


Imprinting robots bond with study coordinator Dr. Lola Canamero.
Robots have feelings, too -- or at least they will -- pending the completion of a pan-European research project being led by a group of British scientists

The Feelix Growing project aims to design and build a series of robots that can interact with humans on an emotional level, and actually adapt their behavior in response to emotional cues from their human counterparts.

The official goal of the project is to conduct "interdisciplinary investigation of socially situated development ... as a key paradigm towards achieving robots that interact with humans in their everyday environments in a rich, flexible, autonomous, and user-centered way." To achieve this, the 2.3 million-Euro effort has assembled more than two dozen roboticists, developmental psychologists and neuroscientists from six nations.

The robots used in the project are simple designs, including some "off-the-shelf" models. The complexity lies in the software, which will construct artificial neural networks to pick up on human emotions exhibited via facial expressions, voice intonation, gestures and other behaviors.

The leader of the European Commission-funded project, Dr. Lola Canamero is a senior lecturer in the School of Computer Science at the University of Hertfordshire, England. She also serves as the principal investigator and coordinator of the university's Adaptive Systems Research Group, which focuses on researching socially intelligent agents, emotion modeling, developmental robotics, human-robot and human-computer interaction, embodied artificial intelligence and robotics, sensor evolution and artificial life.

In a recent interview with BBC, Canamero said the robots will be designed to detect basic human emotional states such as  anger, happiness, and loneliness. Once the states are detected, the robots will be programmed to perform a support role, by seeking to soothe an angry human, or cheer up a lonely or depressed one.

One of the first fruits of the Feelix Growing effort has been the production of a robot capable of "imprinting" behavior. The behavior is similar to the way many baby animals develop an instinctual attachment to the first adult they see at the time of birth, usually the mother. The imprinting robot prototype learns to recognize a particular human and follow that individual around, gradually adapting to the human's actions and emotional state, then interacting accordingly, Canamero said.

At the conclusion of the project, the scientists intend to build two robots that will combine aspects of the research being conducted at each of the eight partner sites scattered around Europe.



Comments     Threshold


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

Little scenario
By Axbattler on 2/27/2007 12:12:27 PM , Rating: 3
The following scenario just ccame into my mind: In a decade or two, we'll have our first (legal) case regarding abused e-bot (emotion-bot). It might not be a case where the robot itself press charges, but an eyewitness of 'robot bullying' pressing charge on what it appears to the person as inrobotic (i.e. inhuman). They'll get a bunch of computer scientists, psychologist etc. to analyse the robots 'emotional damage'. At first, most people will laugh it off, but gradually, a group of robot sympathisers will make themselves known. And by the time we are old, we will be telling our grandchildren how 'Back in the days, the concept of robot right was completely absurd - robots are only created to serve us nothing else' xD;

... but seriously, I think that bots that are capable of detecting human emotions can have their for some people in the same way that some people like to have pets around (companionship). But I wouldn't like parents to start using them to use those as substitute for spending times with their kids - something I can see happening.

I also don't think it is necessary for robots to have emotions themselves. It'll save the scenario I described earlier (written humorously, but I don't find implausible).




RE: Little scenario
By supaflydaddyc on 2/27/2007 12:41:54 PM , Rating: 2
Axbattler,

I don't know when robots will reach the level of functionality that you speak of, but when they do, you're dead-on that there will be activists groups clamoring for robots' rights.

It gives me an uneasy feeling.

I'm only in my early 20's but who knows what the world will be like when I'm in my 60's. Florida could begin to be submerged under the ocean due to global warming (maybe), robots could roam the Earth, maybe the United States and other Western nations successfully switch over to some form of fuel cell for which we can produce hydrogen cleanly (crosses fingers), but the rest of the developing world will still be reliant on oil and fossil fuels, and the countries of OPEC will be at war with one another.

Let's face it, 40 years is a long time, anything can happen. Heck, five years is a long time. Can you say that you are exactly where you thought you'd at this point in your life five years ago?


RE: Little scenario
By UserDoesNotExist on 2/27/2007 5:38:31 PM , Rating: 2
I must point out one thing, namely that I believe your comparision of e-bots and pets is mistaken. The difference is that I don't doubt whether a dog or cat has a similiar emotional capacity as I do, and I don't think I'm alone of this matter. In case you're wondering why I phrased that so awkwardly, it's because I know that there's some philosophy major lurking out there just waiting to prove how smart he is to a bunch of strangers online by posting some smart ass comment about the nature of existance.

As an aside, the more time I spend training my puppy, the more convinced I am of his sentience as well as mine. No matter what the nature shows say about how the dog is "honing his hunting instinct as a pup" or "developing a submissive bond with his pack master", I know he's just playing with me and likes to have me around. To those who say that he's just a biological machine that has been programmed to the point where he appears to be self-aware, well, I'll be the first to admit that I have no really good, surefire rebuttal. That doesn't automatically make me wrong, especially when you consider that one who propagates such an arguement has no good, surefire rebuttal against the "ghost in the machine" arguement.

With a robot, I just don't have that same personal assurance as to the emotional capacity, and once again I don't think I'm alone on in this matter. This doesn't require that I be truly self-aware, just that I think myself to be self-aware and that I hold views of the sentience of other objects in the universe.

You know what? Nevermind the last three paragraphs. Dogs sentient. Me sentient. Robot not sentient. Me like dogs. Me not like robot. HULK SMASH!


RE: Little scenario
By oTAL on 3/1/2007 9:28:12 AM , Rating: 2
I disagree. From someone who has worked with Aibo's I can tell you that, as idiotic as it may sound, one can easily find himself attached to a stupid machine like that.
The major impediment to further attachment is that it is just too stupid. It's like the normal attachment people have to fish (I know some people get very attached to their fish, but you know what I mean... the normal "hey... I care about them...").
People with knowledge on biologically inspired AI have built models around emotions. Emotions and feelings evolved in nature as a tool to give positive and negative feedback on the learning process. This same tool is useful when used in AI models allowing the emergence of a large variety of behaviors. Feelings are a large amount of variables that affect the subject's current "emotional state". That can be copied and adapted to artificial models... Right now, the largest difference between a machine's software and an animals brain is that the machine's software is highly serialized and the animal's brain is highly parallelized. This higher parallelism allows for a much larger amount of possible states and behaviors, and a greater unpredictability. Still, it can be simulated and is, step by step, being adapted.


What the hell is that headline about?
By nrb on 2/27/2007 4:48:58 AM , Rating: 2
This story is not about "robots getting emotional" it is about "robots becoming able to recognise human emotional states" which is completely different . Merely because I am able to tell the difference between a dog and a cat doesn't mean that I am a dog or a cat. Similarly the fact that a machine may be able to recognise anger or happiness doesn't mean that it is able to be angry or happy.

Get it right, please!




Research is needed
By bassjunkie on 2/27/2007 6:02:32 AM , Rating: 2
Why do people insult stuff they don’t understand, it might be a waist of time now for you but lets say 100 years time or 200 years time, when we will need to be living in space using robots to help us do this.. if we don’t research our ideas now then we won’t get anywhere, fine the moon landings may be fake this i am a skeptic about but from what i remember it was theorized many many many years before 1 attempt was made.. these theories are needed to start the ball rolling, so rather than say this is a waist of time, in fact it isn’t your wasting your time pointlessly complaining about stuff you don’t even pay for, and if it comes outta your tax money rather than than illegals in your country.




Great!
By Mitch101 on 2/27/2007 11:37:50 AM , Rating: 2
Robots with Road Rage thats all we need. Thats probably how skynet really started someone cut off and gave the finger to an emotional robot.




Dangerous
By Azsen on 2/28/2007 4:12:52 AM , Rating: 2
I don't see the need to make artificially intelligent, emotional robots. Hasn't anyone watched enough movies to know that it could all turn to crap and have them turn on humans?

The more research they do, the more likely the movies could turn into reality. Sure, create robots that can do generic tasks like cleaning or work in a dangerous factory or whatever. There's absolutely no need to try and make them like humans and give them emotions because that's where it'll turn bad.




No, they wont
By Sharky974 on 2/27/07, Rating: -1
RE: No, they wont
By Sharky974 on 2/27/07, Rating: -1
RE: No, they wont
By simonjwfg on 2/27/2007 4:01:44 AM , Rating: 2
It's really stupid so far, but who knows the future?
Anyway, the human behavior can gradually be analized. Then someday, the comprehensive high tech in every field of Mechanical,MEMS, Signal & systems, advanced control algorithm and things like come out, It, the robot, may sniffers at you. "how stupid this man"....



RE: No, they wont
By ButterFlyEffect78 on 2/27/07, Rating: 0
RE: No, they wont
By Visual on 2/27/2007 7:53:01 AM , Rating: 2
the matrix is a joke, and pretty poor example. how come out of all the good movies with robots taking over, you picked this one? the terminator trilogy is the standard example for scenarios where AI gets just too smart, next time use it instead :p

indeed we'll be more an more dependent on machines, and a lot of sci-fi movies about AI takeover get closer and closer to becoming real possibilities. with brain-computer interface advances of a couple years ago even the matrix seems possible now, but i'm pretty sure our actual future will be different. there are far more revolutions just waiting to happen in the next couple of decades besides AI breakthroughs.

we'll get better and better understanding of the inner workings of our brains, which may actually help us in creating artificial intelligence. or it may be the other way around, AI developed in some other way can help us understand how our own brains work...

developments in brain-computer interfaces could mean a whole different way of interfacing with computers - including matrix-like virtual reality, controlling machinery as if it is part of our bodies, controlling wholly artificial bodies either as a replacement of our biological ones or in addition to them. it may even turn out to be possible to control multiple bodies at once if your mind can cope with it.
it opens a whole different way of communicating between ourselves too - forget mobile phones, we can technically have technological equivalents of controlled telepathy. generations that grow up with such a technology might even develop new languages, not suitable for speaking but perfect for electronic communication, much faster and more accurate, perhaps similar to how computers communicate with each-other. and since the development of our way of thinking is quite dependent on communication, i imagine the new communication means will result in vastly different intellect for future humans. probably for the better, but again some sci-fi scenarios suggest a bad side is possible, i.e. losing emotional sensitivity and becoming too much like machines. i think even developing a collective conscience could be possible.
but even before any such changes have time to occur, BCI coupled with improved understanding of the brain can lead to funky stuff like mind reading, mind control etc.

another field with great potential is quantum mechanics. it, or a more general understanding of the universe's physics, has exciting possibilities we might one day unlock, some of which could be quite soon. large-scale quantum teleportation or duplication of matter, breaking the speed of light, manipulating space-time for instant travel or time travel are all very likely to turn out impossible or unusable for any practical application, but who knows, the future might surprise us. even simpler things like developing ways for observation or manipulation of events at the quantum level has great potential for nanomaterials and even nanomachinery production. another example is data storage at the atomic or quantum levels, etc. for computing.

quantum computing can potentially bring a revolution by itself, and can be one of the keys of emulating a whole world on the sub-atomic scale, matrix-style or just for scientific simulations to help all the other areas. it can turn out to be the key to AI as well.

nanotechnology is already starting to offer us advantages in the form of new materials, but this can get much further in the future. nanobots for production, medical and military purposes might be a reality soon.

we may be quite close to a revolution in power production too, specifically achieving cold fusion or cheaper antimatter synthesis and exploitation.

also, we're getting more and more efficient in designing various processes with minimal entropy and energy loss, for example in conversion between various types of energy or in transmitting energy, even wirelessly.

biology and biological engineering advances, especially dna reading and engineering improves more and more, and can allow us to understand and change nature and even ourselves both for good and for bad. cloning and organ transplantation, including eventual brain transplantation or whole-body transplantation depending on the viewpoint, can increase our lifetimes possibly indefinitely, and is a good alternative to the 'brain in a jar' scenario we can get from the BCI revolution.

the space race has been at a standstill for a while now, but using the advancements of all the other fields it might soon be feasible to reach and colonise the moon, mars, then farther around the solar system like jupiter or saturn's moons. even settling a different solar system or indeed the whole galaxy doesn't seem impossible now.

yes, i took a bunch of sci-fi books and dumped the plot briefs here :p and yes, you've probably read or watched similar things, and can imagine each of the described "futures" with both good and bad sides of it... but i don't think any of those authors got it right. i doubt any man can even remotely guess what our future will be like, because, quite simply, it won't be just one of those revolutions. it will be many of them combined, and many more that we can't even see coming yet, and that combination is something i just can't imagine


RE: No, they wont
By Flunk on 2/27/2007 3:09:55 PM , Rating: 2
Nope, my job depends on computers. In fact the more computers there are the more jobs there are for me and everyone else in IT. In fact if things keep on like this there is going to be a serious shortage of qualified IT staff, once again great for me. So bring on the automation, the more complex the more IT people they need to fix it when it breaks.


RE: No, they wont
By Visual on 2/27/2007 5:46:47 AM , Rating: 3
are you kidding?

about walking, there are enough projects already dealing with it and having a decent success, humanoid robots being able to walk just fine... but more importantly, have you considered that bipedal locomotion is by far not the perfect variant, and is only necessary if you need to achieve humanoid shape. there may not be robots that can outrun you on two feet, but there are designs with 3-8 feet or more, with wheels, caterpillar tracks, whole-body rolling and even jumping or flying, and quite a few of those can beat you in a race.

but of course even a perfect locomotion system is useless if the robot is blind or just too dumb to not run into walls and stuff. that's why they have to work on artificial intelligence before they can allow the robot to run, or "do anything useful" for that matter. and they are working on it.

robots may be stupider than monkeys for a lot of things, but for a few they are smarter than humans. and they'll improve.

granted, this article isn't directly about making the robots intelligent. but it is about an interesting (even if minor) part of the artificial intelligence field, and you still scoff at it.

and i totally don't understand your point about robots being dead. they are, and so what?


RE: No, they wont
By sviola on 2/27/2007 7:51:46 AM , Rating: 2
Have you ever heard of iRobot's Roomba robot? It's a house cleaning robot. Also, there has been intense use of robots in the car maufacturing industry for years. Robots are being used in a lot of other industries as well.


RE: No, they wont
By rushfan2006 on 2/27/2007 12:00:30 PM , Rating: 3
quote:
Also, maybe they work on getting robots to fucking walk first.


Wow, this and your other post -- geez talk about a closed mind. Have you no ability to imagine the future, have you no comprehension or capacity to be open-minded? And yet you are saying that the robots are "stupider" than monkeys. ;>

In any regard, remember a few things with scienctific research - it all has to start somewhere. You think all the inventions with take for granted today just happened over night? You think someone guy said "I'm bored tomorrow I'm going to invent a device I'll call a radio" and *poof* the next day the world had radio. Or perhaps one day somoene said "you know what it sucks to ride this damn bike to work...." and *poof* the next day he drove a car to work.

Lastly, you are comparing a machine that is trying to mimic the most complex "machine" the world has ever known --- the human.

In the least seeing how difficult it is for a machine to think like we do or walk as effortlessly as we can - should bring profound appreciation for how complex and sophisticated our bodies really are. Its pretty amazing when you really think about it.


RE: No, they wont
By oTAL on 2/28/2007 9:01:20 AM , Rating: 2
quote:
Its pretty amazing when you really think about it.


If everybody who posts here was to "think about it" before posting, we wouldn't have half the posts we do now!... and we would hardly have any red dotting the comments section.

On the other hand your post was pretty good. Science is an evolving thing. The guy who invented the telephone (PLEASE let's not get into an argument there - Meucci, Bell, whatever) probably based his concept on the two cups connected by a cord thing.
And when he did it, the invention sucked. Bad sound quality, interference, needed power and it had to be stuck to a place. What a piece of shit. Why didn't the idiot invent a cell phone in the first place? Maybe he was afraid of the radiation or something... idiot!


"We basically took a look at this situation and said, this is bullshit." -- Newegg Chief Legal Officer Lee Cheng's take on patent troll Soverain











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