Swiss Experimenter Breeds Swarm Intelligence 144
destinyland writes "Researchers simulated evolution with multiple generations of food-seeking robots in a new study of artificial swarm intelligence. 'Under some conditions, sophisticated communication evolved,' says one researcher. And in a more recent study, the swarms of bots didn't just evolve cooperative strategies — they also evolved the ability to deceive. ('Forget zombies,' joked one commenter. 'This is the real threat.') 'The study of artificial swarm intelligence provides insight into the nature of intelligence in general, and offers an interesting perspective on the nature of Darwinian selection, competition, and cooperation.' And there's also some cool video of the bots in action."
Of all the countries... (Score:5, Funny)
Re:Of all the countries... (Score:5, Funny)
Re:Of all the countries... (Score:4, Funny)
yeah but those clocks... you can set your watch by them!
Re: (Score:2)
Actually I think a friend from primary+high school was Swiss. He was pretty cool. Must be the exception that proves the rule, or some other strange turn of phrase.
Re: (Score:3, Funny)
Re: (Score:3, Insightful)
Re: (Score:2)
Skynet shows its face again (Score:2)
Re: (Score:2)
Re: (Score:2)
What makes you think life as we know it isn't nano swarm intelligence gone terribly wrong?
Re: (Score:2)
What makes you think life as we know it isn't nano swarm intelligence gone terribly wrong?
It has gone terribly wrong. But it isn't nano. One look at Rosie O'Donnell will tell you that. Intelligence is open to debate for similar reasons.
Re: (Score:2)
I know you're joking but this seems as relevant a place as any to point out some points about this, swarm intelligence is really about emergent properties of largely random systems.
Ants for example will spread out randomly from the nest looking for food, when they find it they will return with a piece of it to the hive leaving a pheromone trail behind, other ants that are moving about randomly may cross this trail and follow it, they will reach the food too by following it and will also leave a pheromone tr
to counteract (Score:5, Funny)
Re: (Score:1)
Re:to counteract (Score:5, Funny)
Actually, not a bad way to poison a bot's intelligence. Its been done before, with hilarious results (well, if you like 4chan-style humor), with a chatbot called Bucket. It was designed to pick up the basics of the English language and conversation techniques from random internet users.
Then 4chan found it.
http://encyclopediadramatica.com/Bucket [encycloped...matica.com] has the full story, along with quotes and screenshots.
Re: (Score:2)
you really can't balme 4chan for this.
If you go to the forums [jonanin.com] you will notice there are no capcha systems or even basic registration systems to allow proper bans to be imposed.
Every thread is quickly found by a forum spam bot and filled to the maximum page length with spam links.
This is a failure of basic security.
slashdot! (Score:2)
In other news, an experiment by SourceForge, using it's meatspace zombienet "Slashdot" proved that even Google-owned YouTube can be brought to it's knees by enough people trying to watch the same video at the same time.
replicators? (Score:1)
is this the beginning of replicators (from the Stargate universe)?
Re: (Score:2)
Re: (Score:2)
ummm...there's a difference between "Stargate Universe" and "Stargate universe".
One is the show (which has potential...but is not the same "stargate" like the others that have come and gone)
Re: (Score:2)
As far as potential... That plot has holes you could fly a Goa'uld Mothership through! Even ignoring the whole video game garbage and that cheese like that was only ever part of intentionally cheese episodes of previous series...there are more logical inconsistencies than I can count. At first glance it wasn't te
Re: (Score:2)
Didn't they read Prey!? (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
"Prey" is a pretty good scifi novel about this. It follows the tired cautionary-tale forumla, but like all of Crichton's novels has (some) basis in real research.
Not it's not; the formula is just scaremongering; and it's about as based in real research as Congo's gorilla hybrids, as Andromeda Strain's magical, energy-eating, crystal viruses, as Jurassic Park's spontaneous evolution of lysine synthesis genes in less generations than you can count on one hand, as State of Fear's wide-eyed acceptance of junk science that challenges the "religion" of global warming, and as Sphere's... whatever the f--- Sphere was supposed to be.
Crichton is a hack that you stop being imp
Re: (Score:2)
Russian Science Fiction (Score:2)
Hullabaloooo (Score:1)
Cool - but why use real robots for this? Seems like you'd be better off creating virtual robots in a simulated environment to develop the algorithms for something like this. You don't have to worry about dead batteries and hardware failures, and your simulations can run faster than real-time.
Then again, maybe that's what the researcher did, and we're just seeing the end product applied to real robots.
Re: (Score:2, Insightful)
Re: (Score:2)
From TFA: "First simulated in software before using actual bots, five hundred generations were evolved this way with different selective pressures by roboticists and biologists at the Ecole Polytechnique Fédérale de Lausanne in Switzerland in 2007."
So yes, that's exactly what they did.
Also, I'm sure this is at least a rehash of a previous /. article, because I remember discussing the deceptive behavior with the light-flashing. It's still interesting.
Re: (Score:2)
I did however do a text search and came across this line: "First simulated in software before using actual bots"
Real hardware is more information rich (Score:4, Interesting)
Real hardware can hold more states than a purely digital system.
I remember reading a paper (can't find it now though - darn it) about a guy who was doing neural net research with Xilinx chips. Same idea. Whenever an algorithm would do well he'd break it into "genomes" and pair them off with other successful programs.
The board was a bank of Xilinx chips, the genomes were the programming files (basically 1s and 0s fed into the configuration matrix), and the goal was to get the thing to turn on and off when you would speak "on" and "off" into a microphone.
It eventually started working. More interesting than that is what happened when he loaded the program into another board. It didn't work.
It turns out the algorithm had evolved to take advantage of the analog properties of the specific chips in that particular board. The algorithm didn't see the board as a digital thing. It saw it as a collection of opamps, amplifiers, and other analog parts. Move the program to a board that is identical digitally, and it failed because the chips weren't analog exact. You wouldn't have seen that behavior in a purely digital simulation.
Re:Real hardware is more information rich (Score:5, Interesting)
It turns out the algorithm had evolved to take advantage of the analog properties of the specific chips in that particular board. The algorithm didn't see the board as a digital thing. It saw it as a collection of opamps, amplifiers, and other analog parts. Move the program to a board that is identical digitally, and it failed because the chips weren't analog exact. You wouldn't have seen that behavior in a purely digital simulation.
Yeah, I remember that, but differently (or maybe it's a similar but different incident). What I recall is that he looked at the working design, and saw that it included a section that wasn't connected to anything else. Thinking this was just random waste, he removed it. Then it stopped working. Capacitive and inductive effects from the 'disconnected' section was affecting the main 'working' section and making a complicated analog circuit.
In either case (and both are certainly possible outcomes), this outlines what is so awesome about Genetic Algorithms and the natural evolution that inspires them -- no preconceived notions about what the solution should look like. Whatever works, works, and that's literally all that matters. Us humans very often start with a picture in mind of what the answer "should" be, and it limits our thinking. On the other hand, a lot of times we have those preconceived notions like "this circuit should be digital not analog" for very good reasons, and we simply fail to notify the GA of that requirement. Which also makes GAs fun. :)
Re: (Score:2)
Like a future with humans in it, you mean?
Indeed! And it's exactly that kind of unspoken assumption that can make/break an algorithm. Who knows -- maybe the only problem with Skynet was that somebody forgot to write a rule that wiping out humanity was not an acceptable solution to human conflict.
Re: (Score:2)
You would if you virtualized those analog parts.
That's what most virtual modeling does, whether it's stress analysis in AutoCAD or reproducing that "tube warmth" from a solid-state amplifier through massaging the wave.
More on that (Score:2)
Yes, you would, but you'd also take a hit to simulation throughput, I'm guessing a pretty significant one, too. I'm not sure you'd gain anything specifically more useful than you would in a pure digital approach without this kind of low level detail, either. More interesting to add something a bit more "macro" in the sense that it's a high level behavior / feature you can see and evaluate by simple observation.
My qualifications to guess? I'm the author
Re: (Score:2)
It's true that the simulation will not simulate analog properties, but then again, that's not your desired behaviour. You want to be able to copy your boards, so your evolved "solution" can be manufactured after you've reached it.
I read a similar article in Scientific American in the early 90's. The problem was recognizing 1000Hz signal on an input. The chips also learned to recognize it using their analog, instead of their digital, properties, and the evolved program could not be copied to a different chip
The ability to deceive? (Score:5, Funny)
they also evolved the ability to deceive.
Obviously, once you've proved the entity has the ability to deceive, you must distrust any further results.
Re: (Score:1)
Re: (Score:2)
That's funny, but also a very interesting point.
They are "deceiving" each other, not the researchers : " By the 50th generation, some bots eventually learned not to flash their blue light as frequently when they were near the food so they wouldn’t draw the attention of other robots." I don't know if deception is really accurate in this case since to me it suggests intent while that's not the case here. Maybe natural "camouflage" like you see in animals is a better analogy.
Re: (Score:3, Funny)
They are "deceiving" each other, not the researchers...
That's just what they want you to think...
Re: (Score:2, Insightful)
And you'll have to go back to your earlier results and wonder, when did it start deceiving?
Re: (Score:2, Informative)
Robotic Evolution (Score:3, Interesting)
"You fool! We were created in our present form by the great nerd in the sky! Shun the non-believer!"
Re: (Score:2)
Evolving their AIs, yes, not their physical capabilities. Genetic Algorithms have been in use for AI programming for quite some time now.
On an unrelated topic, have you heard the good news of Robot Jesus?
Re: (Score:1)
Re: (Score:2)
Save your soul! Also remember to make regular off-site backups!
Re: (Score:2)
CmdrTaco looked deep into our souls, and assigned us a number based on the order in which we joined.
Re: (Score:1)
Re: (Score:2)
"You fool! We were created in our present form by the great nerd in the sky! Shun the non-believer!"
Sounds very much like the scenario in "Saturn's Children". All the humans have died off, and only the sentient artificial servants are left. The weird (well, one of them) is that they all have heard of "Evolution", but view it as some crazy old ancient religion that only the simple-minded would believe.
i think this was covered already... (Score:3, Informative)
Forget Zombies? (Score:1)
Putting the cart in front of the horse IMO (Score:1)
Re: (Score:2)
I don't know where you live but around here we realized swarm stupidity a long time ago. Then again, I doubt the swarm has figured it out yet.
Re: (Score:2)
We have not even realized swarm stupidity yet, how can they claim swarm intelligence?
"Stupidity" can't exist without intelligence. "Stupidity" is what you call it when one intelligence rates the performance of another intelligence, and it's usually measured against a background of the subject species' average intelligence. (i.e. A "smart dog" is "smart for a dog," not smart compared to a human.)
Until the robot swarm has identifiable intelligence to begin with, there's no more point in claiming stupidity than there is to claim stupidity for an amoeba or a chair. Therefore, it's not putti
What does it tell about the intelligent designer? (Score:3, Insightful)
Re: (Score:2)
With infinite wisdom and omnipotence and infinite resources, the Designer (or Designers) should have been able to create much more cooperative human beings. No wars. all peace.
Well, by Norse mythology, Odin, Vili, and Ve created the humans to fight in the final battle of Ragnarok, which wouldn't be much of a battle if humans just sit around all day and post to slashdot. The world is supposed to end in flames, perhaps Ragnarok will be started by a vi vs emacs flamewar on slashdot. Certainly the Norse mythology fits the human condition much more closely than the Christian mythology. Which would imply...
I wonder how they (the IDists) are able to square their ability ti "infer design" with the obvious "deficiencies of design".
If you really want to mess with the heads of IDers, ask them what they'd do i
Re: (Score:2)
soooooo... Destiny vs. free will & self-determination?
Re: (Score:2)
I have an even better question for ID'ers. What does THIS say about their so-called intelligent designer?: http://en.wikipedia.org/wiki/Penis_plant [wikipedia.org]
Were we created by Beavis and Butthead? I can imagine the scene on Day 3 or thereabouts of the Creation:
[God] Huh-huhuhuh-huh-huh. Hey, Lucifer. Check this out, dude. *zap!* It's a schlong cactus. ...what's a schlong?
[Lucifer] Heh-m-heh-heh. Yeah, that's pretty cool, m-heheh. Schlong.
Re: (Score:2)
Waste of time (Score:2)
Why even bother with robots? So it looks more real and tangible than just a computer simulation? Maybe, but other than that it's a waste of time and resources. Anything you could learn you could learn from a simulation of those robots, since this is entirely an algorithmic problem. I guess these guys just like to play with robots.
Re:Waste of time (Score:4, Insightful)
I imagine that there might be interesting results that come from putting objects into an environment where you don't control all the variables. I've heard of cases where the robots end up using features of their own hardware (which is generally cobbled together from off the shelf parts) that the researchers never anticipated.
Re: (Score:2)
Re:Waste of time (Score:4, Insightful)
Because intelligence isn't just a software thing. At least not in humans.
I recall reading about field programmable gate arrays being used in an experiment with genetic algorithms. They wanted to force the FPGAs to evolve to tell the difference between two different frequency sounds. Eventually they wound up with chips that accomplished the task in a variety of ways - ways that worked but for no explicable reason, some of them being ways that took advantage of tiny differences in the individual (identical, at least from a manufacturing perspective) chips, and even that required slight differences in the room's environment. This was years ago.
Simulations won't have those little idiosyncraces between individual units and thus might miss a huge component. Variation among individuals that is only in software misses the whole concept of variation between individuals that comes about from hardware, and also from the interaction between the two.
Re: (Score:2)
Re: (Score:2)
Clearly you missed the point of the Great Movie "Short Circuit"
Re: (Score:2)
It is far more efficient to use real devices, although simulations can be very useful also.
Re: (Score:2)
How many things do you need to simulate when you're exploring AI algorithms? Besides, you control things better when things are simulated, you can speed things up tremendously, you can get as many units as you need (no need to build them) and then you don't have to work out the kinks of making a robot that interacts with its environment.
I think that outweighs any of the bullshit effects another poster mentioned.
worst spaghetti code ever (Score:3, Insightful)
One question that intrigues me is just how human-readable the code produced by such genetic algorithms is. Some of the practical promise of this work is that it produces problem-solving code in ways very difficult from that of human programmers -- but how can such code be maintained by humans? It's a bit like making an engineer try to figure out how your lower intestine works.
Re: (Score:1)
Maybe we'll need a new breed of biological progammers. Maybe eSurgeons?
Re: (Score:3, Informative)
From what I've read on the subject of machine evolution (mostly articles for the layperson), the end results are often completely baffling. It works, but the reason why isn't very obvious. In a few cases, I recall reading about evolved antenna schematics & shapes that worked REALLY well, but made absolutely no sense, or took advantage of things that engineers normally consider flaws/problems to be overcome in design.
So yeah, it'd probably come up with code & designs that are pretty difficult to pars
Re: (Score:2)
See the work of philosopher Patrick Grim from 2000 (Score:2)
"Evolution of Communication in Perfect and Imperfect Worlds "
http://sunysb.edu/philosophy//faculty/pgrim/pgrim_publications.html [sunysb.edu]
http://www.sunysb.edu/philosophy/faculty/pgrim/evolution.htm [sunysb.edu]
"We extend previous work on cooperation to some related questions regarding the evolution of simple forms of communication. The evolution of cooperation within the iterated Prisoner's Dilemma has been shown to follow different patterns, with significantly different outcomes, depending on whether the features of the model a
Dup dup! (Score:3, Informative)
And offtopic: $&^@%! Taco, what's up with the popups that sneak past Firefox popup blocks? I've dutifully allowed advertising to continue, despite having that checkbox I could click to turn ads off for good behavior. Do I really need to turn on adblock and noscript for
Re: (Score:2)
No script for you, come back more than one year!
What's new here? (Score:2)
Ok, what did this study teach us that wasn't learned years ago in (for example) Boids [red3d.com] (1987), Core War [corewars.org] (1984), and Tierra [wikipedia.org] (1991)? I mean, it's cool having little bots running around a tabletop and all, but I was simulating the same behaviors on my '286 back in the mid 90's.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Didn't bother with bots - I concentrated on the steak, not the sizzle, and stayed in software. Their being hardware robots rather than software agents doesn't really change the underlying behavior. My agent's behavior was evolved, though they didn't evolve the same behaviors as those in the experiment, they did evolve unique behaviors of their own.
Which is my point, they discovered specific new behaviors that arose because of the specific features of their environment - something long known to occu
Apple sponsored? (Score:2)
That apple looks like a familiar sticker one gets when buying a certain computer.
tried evolution in AI (Score:2)
Re: (Score:2)
conspicuously absent is any explanation of what is meant by "learned" in this context or how the algorithms "evolved"
Um, it's in the first couple paragraphs of the article. You did read that, yes?
Re: (Score:2, Insightful)
I am confused as to how it was possible to understand the claims it made. Robots don't have genomes and don't eat food. Their "genomes" cannot be recombined.
Robots have code and programmers. What does a random change mean in this context? Do they use faulty dram or mess with the voltage?
Re: (Score:2)
Well, I did not read the article, but maybe they're talking about genetic algorithms?
Re: (Score:2)
The "food" is just something that they're supposed to move towards. I have heard of similar (hobbyist) setups where it's actually a charging station so the "food" aspect is more literal, but all that's necessary is tha
Re: (Score:1)
You're agreeing that the terms are not what a normal reader would construe them to mean.
if the experiment wants to show anything, the methodology has to be more transparent so that we can know whether to consider its "genome" as really a genome or are something more banal.
if the semi-random is really just someone going through and changing parameters in a config file (or using a script to do it), then it's not really random at all.
Re: (Score:3, Informative)
it sounds like you are now accepting that the article is misleading.
You're agreeing that the terms are not what a normal reader would construe them to mean.
The terms are a (very good) metaphor, and the article is not at all misleading. I would have thought this would be obvious.
if the experiment wants to show anything, the methodology has to be more transparent so that we can know whether to consider its "genome" as really a genome or are something more banal.
The entire point of this sort of research is that the "genome" in the bots is analogous to, but far simpler than, a biological genome, and the means of selecting which "genomes" to generate the next "generation" from is analogous to how genomes are selected in biology (either "natural selection" like you find in nature or "artificial selection" like you get with farmed crops or dog bree
Re: (Score:1)
The terms are a (very good) metaphor, and the article is not at all misleading. I would have thought this would be obvious.
unless you're the author of the underlying study, I am unclear as to how you have knowledge of the methods and science behind what they are doing.
I would have that this would be obvious
The entire point of this sort of research is that the "genome" in the bots is analogous to, but far simpler than, a biological genome, and the means of selecting which "genomes" to generate the next "generation" from is analogous to how genomes are selected in biology (either "natural selection" like you find in nature or "artificial selection" like you get with farmed crops or dog breeding).
the entire failing is that it's not clear that the simplified model in any way duplicates the more complicated model.
oddly, when you simplify something, you often
Re: (Score:2)
in summary, while you have marshaled an interesting array of wikipedia articles, the original article in question remains a piece of hype-mongering.
it has in no way connected itself to any of what you have stated.
instead, it has merely used (or possibly abused) the terms of biology to describe what might otherwise be a rather boring high school science fair experiment.
Oh, I see. You're not complaining about this specific article, you're trying to claim that that entire field is a crock.
Have a nice day.
Re: (Score:2)
unless you're the author of the underlying study, I am unclear as to how you have knowledge of the methods and science behind what they are doing.
Electron microscopes have been around for decades. So long, in fact, that you do NOT have to explain how an electron microscope works every single time you show a picture taken with an electron microscope. Instead, you publish an article and you say "figure three was taken with an electron microscope" and anybody unclear on the subject can go and read up on how that works.
In the exact same way genetic algorithms have been around for decades. And in the exact same way, you do not have to spell out the pre
Genetic Algorithms (Score:3, Informative)
I did. Did you understand the meaning behind the paragraphs you cite?
I did! Whenever you hear about a robot/AI "evolving", you should immediately think Genetic Algorithms [wikipedia.org]. Most frequently in association with Neural Networks [wikipedia.org] as is the case here (mentioned lower in TFA).
I am confused as to how it was possible to understand the claims it made. Robots don't have genomes and don't eat food. Their "genomes" cannot be recombined. Robots have code and programmers. What does a random change mean in this context?
Re: (Score:1)
Re: (Score:2)
the robots do not deceive; they do not see food. the article is a misconstrual of what is happening.
Only in the same sense that you are a figment of your own imagination, and any discussion of there being a "you" or "me" is also a misconstrual.
Re: (Score:1)
Only in the same sense that you are a figment of your own imagination, and any discussion of there being a "you" or "me" is also a misconstrual.
how so?
the one is clearly a construction that we can fully comprehend because we generated it.
the other has yet to be shown to be merely a construction (whether or not it can ever be shown as such).
maybe to make it more clearly, robots do not survive on the basis of said "food" so it's not the same as our "food" even if both deserve the quotes.
the further difficulty with your claim is that you state "Only in the same sense that you are a figment of your own imagination". But then it seems that we need t
Re: (Score:3, Informative)
oh, so you're saying we just make up new meanings for old terms and act like their the same! oh now i get it!
Redefine? They're direct analogues of their biological terms. It's not like they're using the terms for something completely different.
If you have a problem with the specification of a bit pattern that defines the robot's behavior as a "genome", state it. The choice to refer to it as such is not simply a made-up meaning. It works perfectly in theory and in practice.
see that's the problem. the ro
Re: (Score:1)
That was actually helpful.
I see your point about the usefulness and prevalence of these analogies.
I just still question how well they fit the biological model.
but i do appreciate your efforts to help me see.
Re: (Score:2)
I see your point about the usefulness and prevalence of these analogies.
Yeah, they are essentially analogies -- how could they be otherwise, when comparing robots to life forms? They're just pretty good ones. :)
I just still question how well they fit the biological model.
It's a fair question.
One thing that may help is to remember that these robots are evolving behaviors comparable to those of very stupid bugs. "Deceive" may seem to imply some intentional act of subterfuge, but for many animals the concept
Re: (Score:2)
[...] and they do see the thing they are rewarded for acquiring.
They aren't "rewarded". They don't need to be "rewarded". The genome of units that do NOT seek food is removed from the pool of available genomes. That is all. You evolve this for a couple generations and after a while every single genome in the pool will seek out food. It doesn't have to "know" that it does and it doesn't have to know "why" it does it, it just has to do it. And whenever, wherever any given gene sequence happens to stumble upon the innovation of NOT seeking food it is removed from the pool.
Re: (Score:2)
I guess I should'nt try to provide food for thought to someone too retarded to parse the phrase "food for thought".