Slashdot Mirror
Robot Dogs Evolve Their Own Language
bab00n writes According to this article at The Engineer Online, researchers led by the Institute of Cognitive Science and Technology in Italy are developing robots that evolve their own language, bypassing the limits of imposing human rule-based communication. The technology, dubbed Embedded and Communicating Agents, has allowed researchers at Sony's Computer Science Laboratory in France to add a new level of intelligence to the AIBO dog. The robot dog has learnt to see a ball and tell another one where the ball is, if it's moving and what colour it is, and the other is capable of recognising it.
It's an interesting step in the direction of AI. I'd like to see how far they can get this to go on current computing power.
I, for one, welcome our new robot dog overlords.
The guy there has video of an Aibo following a ball and differentiating colors from a few years back.
3 billion human legs were humped on August 29th, 2007. The survivors of the frottage called the war Judgment Day. They lived only to face a new nightmare: the war against the cute little machines.
"Hey! Hey! He's got the ball! He's got the ball!"
"Oh boy, gimme the ball! I want the ball!"
"Ooh, a squirrel! Hey! Squirrel! Gotta get the squirrel!"
"Oh, gimme a treat! Please.....Gimme a treat!"
"Oh boy! Someone new! I wonder what his crotch smells like?"
Gifts for Geeks - Stuff that really matters!
but can they run linux?
This seems a little hard to believe. I could believe that they programmed it to be able to speak and hear statements that are directly connected to thoughts, but I just can't see an AIBO learning, much less inventing, the syntax to be able to say something like "The red ball is behind you, rolling to the right." It just seems a little far-fetched.
What the article doesn't explain is at what level the language system is attached to the brain. Does it talk about raw thoughts, or specific ideas (like the ball)? Do AIBO's have "raw thoughts", or can they only think about what they were programmed to know about?
ttuttle is a rankmaniac
How does this work? Is it a neural network, where sounds are associated with objects? That would make sense for the first part, but how does a neural network represent more complex ideas like "the red ball is behind the blue ball"? Or do the AIBO's not have thoughts that complex?
ttuttle is a rankmaniac
Hey, your one of those robot dogs aren't you? -Todd
-Todd
Put down the sig, and step away from the computer.
""What has been achieved at Sony shows that the technology gives the robot the ability to develop its own language with which to describe its environment and interact with other AIBOs. It sees a ball and it can tell another one where the ball is, if it's moving and what colour it is, and the other is capable of recognising it," Nolfi said."
These quadrapedal Terminators can now coordinate their efforts to get our balls. The rise of the machines has clearly begun. We shouldn't give robots the ability to scheme in their own langauge - how embarrasing would it be if the human race were wiped out by cute robot dogs?
If we can put a man on the moon, why can't we shoot people for Apollo-related non-sequiturs?
Very interesting.
.. paranoid crackpot leftover from the days of Amiga.
And of course, the dogs will be used to track down those who try to circumvent Sony DRM.
for "Sieze Control"? hrm. might be a tall order for a robot dog. No opposable thumbs.
"If god did not exist, it would be necessary to invent him" --Voltaire
This is all very fine and dandy, but I don't believe that mimicking what is presently known about human language capabilities will help us understand it better.
The technology was, if I understood the article correctly, built on the foundation laid by cognitive science. It mimics chldren's curiosity, it begins from the general semantics (i.e. selecting an entity), goes on to phonology (i.e. the shape of the symbol for the entity), and deals with finer points (morphology, syntax) in the end...
I'd be very interested to see how it goes on, but I really don't think we'll be seeing a huge breakthrough in cognitive science.
NLP, maybe... almost definitely, if we can get machines to learn human languages.
But I really doubt the humans and animals part.
Ignore this signature. By order.
As far as I can remember from my student days, brains in living creatures grow special node cells and link them together in order to create memories and associations, including all successes and mistakes.
I'm curious to know just how the robot learning is stored in this case. I have always thought the biggest hurdle on robot learning (including walking, knowing not to grip an egg to hard, etc) would be the space available for all the information. Would a 512MB memory stick be enough..? Surely more like 60GBs...
THE HONOUR OF THE KNIGHTS - CC Licensed Sci-Fi Novel
"I'll be back..."
"We are all geniuses when we dream"
- E.M. Cioran
...wonder what the word for "your butt smells like Alpo" is?
Crunch!
"Play? PLAY!"
Ignore this signature. By order.
Robot Dogs Evolve Their Own Language? That honestly sounds more like a Fark headline.
FTA:
Researchers led by the Institute of Cognitive Science and Technology in Italy are developing robots that evolve their own language, bypassing the limits of imposing human rule-based communication (my bold).
FACT: Human communication is definitively NOT rule-based, neither syntax, semantics nor phonology. Unless you accept that most rules have more exceptions than cases where it applies. And even the exceptions have (typically) several layers of exceptions. Google for 'machine learning of natural languege' for several millions hits on several decades of scientific research on the subject.
maybe they meant 'human imposed rule-based communication' ?
I guess they are intelligently designed?
Would you kindly mod me +1 insightful?
But, how long before it can fetch my slippers?
Ugh, never mind.
This is interesting. If these robot toy dogs can do this, that puts engineers one step closer for any U*V fleets the military will use. If they can "talk" to each other, and one is fired upon, it can tell the other crafts where to target and return fire. But, now it can also do more complex things in the future. If one craft can tell the other crafts where a target it is, we can have stealth attacks done autonamously.
That which does not kill me only postpones the inevitable.
Nice doggy. . . (I hope).
What?
The initial prototype (named 'Maxthon') is the first in this new line of robotic dogs (which, oddly enough, resembles the "Shinese" breed). 'Max' is supposedly using his new language to circumvent Chinese censhorship.
What is this Zebra program you speak of? --ThinkingInBinary
Before long, not only will we have real dogs keeping us awake at night.. but robotic ones too
This is really exciting but the prospect of swarms of any kind of robot is a bit scary - hopefully designers will build in a simple, easily exploitable flaw so that an out-of-control swarm could be easily deactivated.
To me, it would seem the most seminal part of creating AI is to somehow instill "wants" and "needs" into machines. Without those, there's really no intelligence. When it comes down to it, the only reason we (humans) do anything is to be happy and to survive; how the hell do we make a machine want/need to be happy/survive? Interesting stuff, to be sure, but we've really got quite a long way to go.
Reprogramming AIBO dogs is a DMCA violation is it not ; )
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
it seems some occurred on that link (missing semicolon huh?) =) http://yro.slashdot.org/article.pl?sid=01/10/28/00 5233
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
to me, it sounds like they just had the dogs make up word for things, not exactly creating a language itself. I think the concepts of a language to describe things was already programmed into these dogs, so they only had to make up sounds to fill already given attributes like "thing.size" "thing.direction". Doesn't sound like a step in actual AI, but i coulda be usefull for translators ...
or at least tech him to stop humping the roomba. everytime the roomba comes off its charger Rex jumps it like a 16 year old on prom night. I would throw water on him but that seems like a bad idea.
WTF?
researchers led by the Institute of Cognitive Science and Technology in Italy are developing robots that evolve their own language, bypassing the limits of imposing human rule-based communication.
... are developing robot that evolve their own language, bypassing the limits of the imposing human rule-based communication.
... developing ... evolve ... bypass ... human rule-based
Key things to note:
researchers
and one more:
researchers
Well, if the engineers develop them so that (I am guessing) they can agree on their own sound based communication (if indeed they are limiting themselves to audible communication, for purposes of being quaint or stupid) then they are still the 'limited imposing rules' that they will 'learn' by.
Like the story about "robotic 'baby' rats that learn their own behaviours" - they put 'baby' in the title as a loaded statement (babies learn, and are born, and alive). Just because the guy can't program and all 'robots' (motors connected to wheels....) end up in the same corner, he says this is a breakthrough to get more grant money, yet it filters into the consciousness.
It is also a bit like people looking over the fact that 'more accurate textbooks' that contain 'evolutionary theory' are deliberately including known untruths rather than erase and go back on assumptions that supported most of the theory.
Utter fucking stupidity.
please type the word in this image: realist
verification text - if you are visually impaired, please email us at pater@slashdot.org
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
playing..
Spot: [the AIBO's are running the gauntlet toward the Red Ball] The newspapers- they've stopped!
Rex: [realizes why] Stabilize your tails... Watch for enemy cats!.
Rover: They're coming in! Three marks and 2-10!
[Spot] is slain by Darth Puddles and his wingmen; Rover starts to panic]
Rover: It's no good down here, I can't maneuver!
Rex: Stay on target.
Rover: *We're too close!*
Rex: Stay on target!
Rover: [shouts] Loosen up!
[he too is picked off by Puddles and Company; Rex tries to escape but is fatally winged]
Rex: Rex to Skippy, I Lost three, Lost three. They came from... behind!
[crashes]
"There is only a one in six billion chance that you actually exist"
This sounds like intelligent design to me. The AI may change over time but thats what its designed to do, obviously by some intelligent folks in robotics :)
..."Do Robot Dogs Dream of Electric Cats."
It will eventually be made into a movie starring Harrison Ford as "Shaggy" - an aging inventor who is being tortured by his robotic great dane. The great dane constantly comes up to him and goes, "ruh roh!"
The Kai's Semi-Updated Website Thingy
This article is all fluff. They don't say anything really interesting. Ok, they can communicate. If that's so, then engineers can record it and perform analysis on the lexicon and gramatical structure. I want to know something about that! I'm sure it won't match up well to human language, but that's okay, because human languages are themselves very diverse in the way things are represented. Would it kill them to give a few examples of 'words' (even if they're described in terms of musical notes or whatever), what they mean, and how they go together to form sentences?
The robot dog has learnt to see a ball and tell another one where the ball is, if it's moving and what colour it is, and the other is capable of recognising it.
Just one more step, and it would make a perfect domestic companion. That and a wet, velvety tongue.
We could call it the "Peanut Butter" paradigm.
-Styopa
Stefano Nolfi's team (responsible for the research in TFA) usually come up with very solid and ingenious research, and the project mentioned is a continuation of a large EU-funded project with interesting results. But it should be remembered that this research is quite far from any actual applications, in large part because people schooled in classical engineering instinctively distrust anything that has been "evolved". Understandable, as there is often no way to tell why the evolved AI is doing what it is doing, just that it is doing it, and no guarantees that it will continue doing it when you need it the most. Sometime in the far future, however, I believe artificial evolution might be the standard way of creating control and intelligence for various purposes.
I haven't gotten the chance to read this thoroughly, but I think this is the research paper the article was referring to:
http://ecagents.istc.cnr.it/imgs/blu_paper.pdf
Technology Quarterly
How to build a Babel fish
Jun 8th 2006
From The Economist print edition
Translation software: The science-fiction dream of a machine that understands any language is getting slowly closer
IMAGE
IT IS arguably the most useful gadget in the space-farer’s toolkit. In “The Hitchhiker’s Guide to the Galaxy”, Douglas Adams depicted it as a “small, yellow and leech-like” fish, called a Babel fish, that you stick in your ear. In “Star Trek”, meanwhile, it is known simply as the Universal Language Translator. But whatever you call it, there is no doubting the practical value of a device that is capable of translating any language into another.
Remarkably, however, such devices are now on the verge of becoming a reality, thanks to new “statistical machine translation” software. Unlike previous approaches to machine translation, which relied upon rules identified by linguists which then had to be tediously hand-coded into software, this new method requires absolutely no linguistic knowledge or expert understanding of a language in order to translate it. And last month researchers at Carnegie Mellon University (CMU) in Pittsburgh began work on a machine that they hope will be able to learn a new language simply by getting foreign speakers to talk into it and perhaps, eventually, by watching television.
Within the next few years there will be an explosion in translation technologies, says Alex Waibel, director of the International Centre for Advanced Communication Technology, which is based jointly at the University of Karlsruhe in Germany and at CMU. He predicts there will be real-time automatic dubbing, which will let people watch foreign films or television programmes in their native languages, and search engines that will enable users to trawl through multilingual archives of documents, videos and audio files. And, eventually, there may even be electronic devices that work like Babel fish, whispering translations in your ear as someone speaks to you in a foreign tongue.
This may sound fanciful, but already a system has been developed that can translate speeches or lectures from one language into another, in real time and regardless of the subject matter. The system required no programming of grammatical rules or syntax. Instead it was given a vast number of speeches, and their accurate translations (performed by humans) into a second language, for statistical analysis. One of the reasons it works so well is that these speeches came from the United Nations and the European Parliament, where a broad range of topics are discussed. “The linguistic knowledge is automatically extracted from these huge data resources,” says Dr Waibel.
Statistical translation encompasses a range of techniques, but what they all have in common is the use of statistical analysis, rather than rigid rules, to convert text from one language into another. Most systems start with a large bilingual corpus of text. By analysing the frequency with which clusters of words appear in close proximity in the two languages, it is possible to work out which words correspond to each other in the two languages. This approach offers much greater flexibility than rule-based systems, since it translates languages based on how they are actually used, rather than relying on rigid grammatical rules which may not always be observed, and often have exceptions.
Examples abound of the ridiculous results produced by rule-based systems, which are unable to cope in the face of similes, ambiguities or bad grammar. In one example, a sentence written in Arabic meaning “The White House confirmed the existence of a new bin Laden tape” was translated using a standard rule-based translator and became “Alpine white new presence tape registered for cof
The extreme centre is the paper's historical position. --Geoffrey Crowther
I wonder why these companies seem to prefer dogs over other critters. Why not a robot platypus, squirrel or iguana? what makes a robot dog better than other animal?
It'll go so well with my TARDIS.
"The Sage treasures Unity and measures all things by it" - Lao Tzu
...the last time we let two machines develop their own language that we couldn't understand,things didn't turn out so well.
Just sayin', is all.
~Philly
"The technology could lead to robots able to carry out rescue operations fire attacks by swarming over inaccessible areas to find any left living humans people,"
-Dickens
Despite the generated jokes about dogs and the French, and the "oohing and aahing of the crowd at the AIBO robotics soccer games broadcast on U.S. national television, this is not merely "cute". This may be the most important research that you have ever read about.
Researchers Luc Steels and colleagues at Sony's Paris Computer Science Laboratory in France have performed a series of remarkable experiments demonstrating the development, from naught, of spoken language among robots. Words, grammar and semantics evolve spontaneously among cooperating robotic agents initially programmed with a small base set of ground perceptions and behaviors. And from the development of language arises cooperative group (intelligent) behavior.
Enhanced AIBOs are initially programmed to recognise simple stimuli from their surprisingly limited hardware sensors. Over the course of several hours or days, the AIBOs learn to distinguish objects and how to interact with them. A built-in curiosity system ('metabrain') continually directs the AIBOs to look for new and more challenging tasks and to cease activities that are not fruitful. In time they develop more complex tasks, just as do human children.
Like children, the enhanced Sony AIBOs initially babble ("argue?") until two or more settle on a sound to describe an object or aspect of their environment. Over time the group gradually builds a lexicon and grammatical rules through which to communicate. Agreement on word usage spreads through the population as terms for similar meanings compete for acceptance. For example, the robots develop the language structures to express that a red ball is rolling to the left. Just as human twins sometimes develop a unique language in which only they can communicate, the enhanced AIBOs (which are clone-like and similar to twins) develop their own language.
Language analysis and generation are part of Good Old Fashioned AI (GOFAI) and have been studied extensively for decades by AI researchers. In the past several decades GOFAI was challenged by Nouvelle AI (Situated AI) championed by Hans Moravec and Rodney Brooks. This alternative approach holds that true AI will not arise from formal mathematical systems but instead from robotic behaviors which have a subsumption architecture as an overall organising principle for the individual robot. This architecture consists of layers of behavioural modules, each capable of carrying out a complete but simple task. Steels' enhanced AIBOs are embodiments of just such a subsumption architecture and provide strong support for Moravec's and Brooks' hypotheses
Prior to Luc Steels' experiments, no one had experimentally demonstrated how language develops among intelligent agents. Steels' experiments are no less than stunning: in a controlled environment AIBO robots develop their own words and grammars for objects in their environment. All aspects of human language development are mirrored in these experiments: words compete for acceptance in the population, new words are created, and grammatical structures arise spontaneously. Steels' work also addresses the idea of a "robot culture", since it is in the context of a population of cooperating agents that language becomes most useful.
Contrast this with the W3C's Semantic Web effort, which has received much more interest and money in recent years due to the growth of the Internet yet has proven far less fertile. In the Semantic Web there are multiple competing "ontologies" (roughly, data dictionaries wherein all terms are strictly defined by specialists from their
Everyone knows that dogs are colorblind.
Everything you know is wrong, Just forget the words and sing along.
I think it is going to be awfully hard for a human being to learn Aibo. Would it not be more useful to make the dog learn a human language?
Oh well, what the hell...
And so in the experiments new words are created; old less useful words decline in use. At any time, there may be multiple words for the same thing in the population, but eventually one of those words mostly "wins over" the other words (although the older word may continue to be used by a small part of the population).
BTW what's an IPC? [See, here we're attempting to resolve terms in our separate "ontologies" (dictionaries): doing what the W3C's Semantic Web cannot do and what the enhanced AIBOs _can_ do].
Now the fun questions...
The enhanced AIBOs used in the experiments have defined inputs (perceptions) and defined actions (outputs). The best way to think of this is to pretend there's a "little man" (homunculus) inside your brain: he only see the inputs from your eyes and controls the output signals to your limbs. Then replace the little man with a computer program. Yes, the computer program can only do what it is programmed to do, but as "Good Old Fashioned AI" has shown, the program can also can alter itself, so there is no lack of variety in resultant behaviors. The key is to choose an appropriate set of initial behaviors for the particular environment.
You should skim some of the papers of Luc Steels, who is the primary researcher behind this work. Don't dig into the details; look for the summaries and the insights. Steels is a good writer and has gone out of his way to encourage research. Here are two sources:
A favorite, How to do Experiments in Artificial Language Evolution and Why. For an example that addresses your question, scroll down to Figure 3 which shows the AIBO and a plot of what it "sees" with it's rather primitive visual system.
Unfortunately, during testing at the F-Secure labs, it was discovered that while you weren't looking, the dog $unlocks your back door$, allowing thieves to clear out your house when you weren't there.
Ignore anything I said above, I actually agree with everything you believe - mod accordingly.
IPC, n - Inter-Process Communication
ELOI, ELOI, LAMA SABACHTHANI!?
Ok, I can see some serious AI potential here. Imagine if Sony ships their next gen of robotic dogs with this software?
Then the damned dogs would have a mind of their own. Hmm. That would make them more like a cat wouldn't it.
OK, Sony, cough up. If you've managed to make robotic dogs evolve a natural (for them, not us) language, then let's see some real research published on how you managed to do it. That way, the scientific community can take the techniques apart and verify them, while some twelve-year-old creates a breed of sentient computer viruses which eventually cripple every computing device in the world.
Easy!!
question? answer -> gender (likeliness)
Is the robot a vessel? Yes -> female (90%)
Is the robot intended for manual labor? yes -> male (95%)
Is the robot intended for house service? yes -> female (65%)
Is the robot intended for sexual relationships? yes -> female (85%)
Is the robot an attack robot intended for use in killing intruders by castration? yes -> female (99%)
Hmmmm.. looks like odds are high that contrary to I, Robot most robots of the future will be fem-bots.
They start shipping with root kits? I'll have to spend all my weekends clearing out the septic tank and drain field.
What if the Hokey Pokey really is what it's all about?
There have been plenty of comments here about robots taking over the world, etc.. But this does remind me a bit of the conversations among the Tachikoma, the spider-like AI mini-tanks of the anime Ghost in the Shell: Stand Alone Complex. The Tachikoma link up and discuss all sorts of things during their downtime. I remember one such discussion involved taking over the world. In that case, I believe the Major was listening in on them, but if robots are left to evolve their own languages, isn't it likely that we humans won't be able to understand anymore? Sufficiently capable robots might just decide that the best way to get the ball and play with it might be to get rid of those pesky humans first!
To the making of books there is no end, so let's get started
This article is all fluff. They don't say anything really interesting. Ok, they can communicate. If that's so, then engineers can record it and perform analysis on the lexicon and gramatical structure. I want to know something about that! I'm sure it won't match up well to human language, but that's okay, because human languages are themselves very diverse in the way things are represented. Would it kill them to give a few examples of 'words' (even if they're described in terms of musical notes or whatever), what they mean, and how they go together to form sentences?
;).
I believe that's not the main thrust of the article- the main thrust was that the aibos developed, independently of human intervention, their own system of communication. I'm certain that if you wanted, you could contact the researchers on your own and get lexical information independently of a slashdot article
Just great. Now I'll have to learn Spanish and AIBO...
It must have been something you assimilated. . . .
There is absolutely no concrete information in this article. Its just a bunch of laymen's terminology that could apply to any old lame technology. It should be a red flag that this research is tied to a commercially branded toy. AI researchers and marketting droids share the common trait of talking up stupid shit; there is no evidence of anything interesting happening here.
Nice to see that it's in place at Sony but um... this has been done. See Luc Steels.
http://arti.vub.ac.be/~steels/
Very interesting work but not new.
putting pigion into Google results in Google asking "Did you mean: pigeon", the correct spelling.
Reminds me of "Colossus - The Forbin Project". It's about yer basic computer-gone-berserk. Halfway through the movie, "Colossus", the machine that the US has built to automate defense, announces "There is another system", referring to the heretofor unknown Soviet counterpart system "Guardian". Colossus demands to be connected to Guardian, and the two machines develop their own language for communication, unreadable by the pesky carbon life-forms. Together they take over the world, peace prevails, and the remaining survivors live more or less happily ever after, until the second book, when the Martians help the humans disable the computers so that they (the Martians) can take over the world.
The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
IPC (inter-process communication) is the method by which two different computer programs, or parts of computer programs, can request information or actions of each other, or give status updates or event notifications. In many cases, this is just an abstraction: the two processes do not communicate, but simply go in and directly fiddle with each other's bits.
i'd hit it so hard, if you pulled me out you'd be the king of britain [bash.org]
Would it kill them to give a few examples of 'words'?
I'm guessing it'll be something like Droidspeak.
i'd hit it so hard, if you pulled me out you'd be the king of britain [bash.org]
Let's film it! That would be great if they wagged thier tale when focused...
Well, if it is true we're one step closer to being replaced by machines, and we all know what happens next, the Matrix. Unless we make powerplants that run on human waste... make them need us. Yes, good thinking Bill.
I find it interesting how this "group intelligence" parallels a similar if not identical idea in how Helen Keller (see http://www.percepp.demon.co.uk/hkeller.htm for some more in-depth attributions) and her leraning of language allowed her to think. It all starts with simple building blocks of basic words, and it grows almost as if its a continuous recurssion, with each itteration becoming a more complex and "intelegent" set of grammar.
I said "kind of language" and not just "language" because that's what I meant. I realize that the system isn't giving the same results every time they run it.
However, the software they wrote to allow this language to develop will only allow it to develop in certain ways, because we haven't developed unbounded artificial intelligence. (Nor do I know of an example of unbounded natural intelligence - the way humans think is a result of how their brains are constructed.) So these languages will be substantially similar.
For example, the grammar might change, but I doubt the kind of grammar will ever change. I can't imagine that these dogs are using anything more than an iterative one. That is, no part of speech can be defined in terms of itself. Human grammar is recursive. Computer languages are recursive. Mathematical notation uses a recursive grammar. Recursive grammars allow for deep, flexible modes of communication. An iterative grammar, like what I expect these Aibos are doing, can only really generate canned sentence structures that can represent a limited range of ideas. None of them is going to be saying a sentence equivalent to "Look at the ball that the dog on the table is holding."
Does this 'meta-brain' draw from any of Lenat's CYC research? Just curious.
DNA is a Turing machine. You, however, being dynamic and emergent, are not.