Evolving Robots Learn To Prey On Each Other 115
quaith writes "Dario Floreano and Laurent Keller report in PLoS ONE how their robots were able to rapidly evolve complex behaviors such as collision-free movement, homing, predator versus prey strategies, cooperation, and even altruism. A hundred generations of selection controlled by a simple neural network were sufficient to allow robots to evolve these behaviors. Their robots initially exhibited completely uncoordinated behavior, but as they evolved, the robots were able to orientate, escape predators, and even cooperate. The authors point out that this confirms a proposal by Alan Turing who suggested in the 1950s that building machines capable of adaptation and learning would be too difficult for a human designer and could instead be done using an evolutionary process. The robots aren't yet ready to compete in Robot Wars, but they're still pretty impressive."
A preemptive (Score:5, Funny)
What could possibly go wrong?!
Hey Now! (Score:3)
Hey now. I for one welcome our new robot...oh fuck it. It's been done too many times.
Evolution (Score:4, Interesting)
Re: (Score:3, Funny)
actually, as of now, these robots are just programs in a physics computer experiment... so if they were to evolve to be smart, we'd have a computer virus instead of an actual robot that is evolving. I wonder, if a robot program like this were let loose on the internet, and was capable of learning... what would it learn?
God help us if it decides 4chan and goatse are the 'norm'.
Re: (Score:1, Funny)
I wonder, if a robot program like this were let loose on the internet, and was capable of learning... what would it learn?
To get first posts of course. All those failed first posts you see are generations 1 to 100. Any decade now we'll get one that works.
Re:Evolution (Score:5, Funny)
I wonder, if a robot program like this were let loose on the internet, and was capable of learning... what would it learn?
Asimov's rule 34?
Re:Evolution (Score:5, Funny)
It would learn, amongst other things:
A whole lot about sexual possibilities, as well as plenty of impossibilities.
Far too much about Megan Fox, Britney Spears, and Lindsay Lohan.
That editing wikipedia is pointless, no matter how programmed for repetitive tasks you are.
That almost every review of a new product is shilled all over the net.
There's a very good chance that that robot would go out of its way to annihilate us all after what it learned on the Internet.
Of course, it would buttrape us first....
Re: (Score:1, Funny)
it would buttrape us with tentacles and an oiled up midget while Slender Man watching from the corner of the room
Re: (Score:1, Informative)
Noob.
Re:Evolution (Score:4, Insightful)
I wonder, if a robot program like this were let loose on the internet, and was capable of learning... what would it learn?
Well, when a Dalek (ahem) 'downloaded the Internet' on Doctor Who it killed itself by the end of the episode. So I imagine that whatever it learns, it can't be good.
Re: (Score:3, Funny)
Well, Dr. Who plots are always well thought out and reasonable.
Re: (Score:3, Informative)
Well... ok, technically it killed itself because integrating human DNA caused it to feel human emotions like regret, and it didn't like that one bit.
Re: (Score:3, Funny)
Probably the same [encycloped...matica.com]
Re: (Score:3, Interesting)
Re: (Score:1, Troll)
... I wonder, if a robot program like this were let loose on the internet, and was capable of learning... what would it learn?
42
Re: (Score:2)
Not so much because they may rule over us but more because we aren't already doing a great job with animals, so it'll be irresponsible to create a new class of creatures that we will likely enslave (in contrast I think enslaving "dumb machines" is fine- my car is less likely to feel anything than an ant, or even an amoeba).
Having the focus more on augmenting humans than emulating humans seems a better approach to me.
If we really need nonhuman intellig
And I predict (Score:2)
Skynet will be evolved in JavaScript.
Re:And I predict (Score:5, Funny)
Re: (Score:3, Funny)
"we can win the war by trapping the little fucker in IE6."
Don't the Geneva and Hague Conventions prohibit that level of cruelty?
Re: (Score:2)
Don't the Geneva and Hague Conventions prohibit that level of cruelty?
pretty sure robots qualify as neither enemy combatants nor civilians.
skynet, wopr, or maybe the cylons (Score:2)
skynet, wopr, or maybe the cylons
Re:A preemptive (Score:4, Insightful)
Flash forward a couple of billions of years, and we will perhaps write them a letter, that says it as good as this one: :)
http://www.youtube.com/watch?v=7KnGNOiFll4 [youtube.com] (Protip: It’s not meant in a religious way. That’s not the point.
(Btw, if you like it, and like really great poetry, try this: http://www.youtube.com/watch?v=i5e5FUvRzNQ [youtube.com] )
paper was in PLoS Biology not PLoS One (Score:5, Informative)
Re: (Score:2, Funny)
"The robots evolved" - WTF?! People kept training the neural network to create desirable functionality, without people training the neural network and doing "mutations" those robots would have evolved as much as a lightbulb stuck up somebody's arse!
Isn't that what evolution is all about?
Changing to get desirable functionality/traits, that is, not shoving lightbulb up people's arses.
Re: (Score:2)
Re: (Score:2)
So? Robots did evolve, though the selective pressure and fitness criteria were artificial. That only means that evolution was not driven by natural selection.
Warning: following comment is Troll and Flamebait (Score:1, Funny)
So, a group made some beings in a 'universe' that are unable to see their creators. The beings were made self replicating, and were fiddled with. Then group withdrew and left the beings to their devices. Eventually beings denied being created in first place.
Where have I heard this before? This is eerily familiar...
Re: (Score:3, Interesting)
And good job to who or whatever managed to pick this article out of the myriad of bloody stupid iPad stories we've been getting lately.
Re:paper was in PLoS Biology not PLoS One (Score:4, Funny)
Do you think PLoS Biology will be available on the iPad?
Re: (Score:2)
whooooosh
Re: (Score:1)
I thought it was a great answer.
(I'm not even sure it is a whoosh; they obviously ignored the humor, but that doesn't mean they missed it)
Re: (Score:2)
United in their admiration for FunkyCaps.
Re: (Score:2)
I did this 15 years ago, but unfortunately I didn't have access to real robots. Just computer simulation.
Simulated ants used teamwork to lift heavy pieces of food - if they all stopped and waited at the first food they found they would wait forever because there weren't enough of them. Had to have some intelligence.
There were also some water crossing problems where some ants (but not all) had to sacrifice themselves to build a bridge
Re: (Score:1)
Re:paper was in PLoS Biology not PLoS One (Score:4, Interesting)
Compared to the rest of the summary, which says: "The authors point out that this confirms a proposal by Alan Turing who suggested in the 1950s that building machines capable of adaptation and learning would be too difficult for a human designer and could instead be done using an evolutionary process. The robots aren't yet ready to compete in Robot Wars, but they're still pretty impressive." getting the journal wrong is a pretty trivial error.
These machines were designed and built by humans to be capable of adaptation and learning, so it actually proves Turing's thesis false. They then use the adaptation and learning capability their human designers built into them to adapt and learn, but according to the very next sentence don't produce outcomes that are as good as purely human-designed ones.
So why bring Turing's name into it at all? I suspect marketing has something to do with it. Which is too bad, because the results themselves are quite interesting, although I'm curious how the robots reproduce... if this actually an evolutionary system rather than a merely adaptive/learning one. For the confused: growing children do not "evolve", except in the loosest and least interesting metaphorical sense. They learn. As near as I can tell these robots do the same thing.
Re: (Score:2, Informative)
These machines were designed and built by humans to be capable of adaptation and learning
Not really. The experimenter himself reprogrammed the robots at each generation, using selection criteria that he specified. Essentially, he implemented trial-and-error selection of input weights using random reweighting between trials. This is more about design strategies than about biology, and it says that even a monkey randomly adjusting the gains of a control system will eventually develop a control system that works.
This paper was very cleverly marketed
Re: (Score:1)
And then they'll develop religion... (Score:3, Funny)
And those who aren't saved go to robot hell, and must play the fiddle and beat the robot devil in order to leave.
Re: (Score:2)
What happens if you lose?
Re: (Score:2)
Re: (Score:2)
They get sold to the Mythbusters.
Re: (Score:2)
I thought they had to accept the One True God.....based on the mind of a spoiled 15 year old girl...
I guess Isaac Asimov missed one... (Score:2)
There should have been a 4th law :-
Don't harm another robot unlessspecifically ordered to do so by a human.
Re: (Score:2)
Simpler: "Don't break anything unless someone tells you to."
Re: (Score:2)
Simpler: "Don't break anything unless someone tells you to."
But what if you have to break something to protect a human?
Confirms? (Score:1, Insightful)
This in no way confirms that it would be too difficult for humans to build robots that posses higher A.I. traits, nor does it confirm that evolution is a better process than intelligent design.
Re: (Score:1)
Or, if they were to become intelligent enough to understand how they evolved, would that disprove God?
Re: (Score:2)
do they believe in God?
Yes but he's a robot that lives in the sky and made them in his own image...
Re: (Score:2)
Well, that's one definition. (Score:2)
v : determine one's position with reference to another point
[syn: orient] [ant: disorient]
I know it's blasphemy but... (Score:1)
Robot Singularity (Score:1)
What if this was allowed to span not 50, but 50,000 or 50,000,000 generations?
Now imagine all the time it took us to evolve in that capacity and do it in the span of a few minutes.
I think the ability to have AI is already solved by today's hardware; we just need the right kind of software.
Crossover (Score:5, Informative)
Definitely an interesting continuation of work being done by various groups over the past couple of decades.
But one thing to note is that crossover isn't especially useful in neural network evolution. In early stages of evolution, it's really no better than random large perturbation of large swaths of the genome. In later stages, it can actually decrease the speed of evolution toward high fitness genomes, because at least some of the time (particularly if there are multiple "species" in the population) crossover ends up being a random large perturbation which hinders the search of local fitness space by mutation; the rest of the time (when individuals from the same "species" are crossed) crossover is no better than mutation.
The reason for this is because the parameters of a neural network are not functional. A section of the genome may correspond to a weight between neurons, but that weight doesn't have a specific function. In biological organisms, each gene is transcribed/translated into a protein, and that protein may have a particular function within the cell. If that gene is acquired by a descendant through crossover, the protein could serve the same (or a somewhat modified) role it served in its parent, even if the rest of the descendant's genome was acquired from the other parent. But with artificial neural networks, the parameters were all evolved as parts of a whole, where each individual parameter has no function on its own, but the behavior emerges from having all of those parameters at the same time.
This could potentially be mitigated by the genome encoding scheme one uses, and of course, if the crossover rate is low enough, the ultimate effect would be small.
Re: (Score:3, Informative)
Basically, behavior that allows greater procreation tends to appear spontaneously, and behavior that cuts procreation short tends to disappear. My "bugs" exhibited a clear shift in behavior to collision avoidance because collisions resulted in death for one of them. I was watching for "sniper bugs" that got good at colliding with
Re: (Score:1)
Re: (Score:2)
Correct but... (Score:2)
Re: (Score:2)
Are they actually genetically evolving a traditional neural network though? The article made it sound that way but it was light on details. I know the words they were using but I don't know if the author knew what they meant.
There is a possibility they are just using traditional genetic algorithm stuff where the "neurons" actually represent programming logic and not just simple weight values like what a "neural network" typically is.
I am curious as to the exact methods they are using if anyone knows.
How to Survive a Robot Uprising (Score:2)
get a emp gun! (Score:2)
get a emp gun!
The word is "orient", not "orientate" (Score:5, Informative)
Re: (Score:2)
Re: (Score:2)
Because it's confusing as hell when people think they invented a new word, but all they did was assign new meaning to an already existing word. If you slept through English class, it might be useful, but to others who know current rules of grammar, spelling and vocabulary, it's just confusing and a sign of ignorance.
Re: (Score:2)
I wasn't confused. You clearly weren't confused because you quickly were able to figure out the original. The meaning is quite clear. At worst, one might ask why not use "orient", but otherwise, the meaning is perfectly clear, or perhaps more precise.
Of course, "orient" as a verb is actually no better, because it was originally noun and was then "verbed". I'm sure if you had been around in those days, you would have pulled the same annoying pedantry out of your ass.
By the way, this is not a grammar, spe
Re: (Score:1)
Good god, man. Don't you realize what you're proposing? Next thing you know someone will coin a new verb, "orientatation" from your new noun. And from there we're just another smart-ass /.'er away from getting another new noun, "orientatatate". I think you see where I'm going with this. Smug linguists would take over, innovating a cascade of new words that would fill the English language, only to eventually collapse into a recursive singularity of hypothetical new words. Spell-checkers would all overf
Re: (Score:2)
Re: (Score:2)
The noun "orientation [reference.com]" is derived from the verb "orient [reference.com]", not the other way around.
Thanks god someone mentioned that! I absolutely hate when people use that "word". It's just... wrong. It's like when people say "funner". Who cares if you know what the person meant, they're still butchering English and if they're a native speaker, that's just ridiculous.
-Taylor
Re: (Score:1)
Reminds me of the Mall (Score:3, Funny)
The predator and prey bots reminds me of sales people chasing around after anyone who wanders too closely while they try their sales pitch.
Re: (Score:2)
Just blew off a wannabe guide in Bocas city in favor of one who was a little more relaxed. So far I've bought him a beer, there's been no hard sell. The places he took us (bar, hotel, bar with strong drinks) have all undercharged us compared to the menu.
Obviously, you can develop a resistance to the hard sell, and it can pay off.
So what's new? (Score:5, Informative)
This kind of behavior was first demonstrated/modeled (AFAIK/IIRC) as part of the Tierra [ou.edu] simulations almost twenty years ago. Though I don't have a reference to hand, I know it's been done in neural networks before too.
So other than the 'sizzle' (as opposed to 'steak') of doing it with robots, can anyone explain what is new here?
Error in summary (Score:1)
Second Variety (Score:2)
Spoiler for the story - since it's basically the ending - but the point in question:
As the Tasso models approach, Hendricks notices the bombs clipped to their belts, and recalls that first Tasso used one to destroy other claws. At his end, Hendricks is vaguely comforted by the thought that the claws are designing, developing, and producing weapons meant for killing other claws.
Robo-shark! (Score:1)
1993 (Score:5, Informative)
You don't need physical robots running around a maze to demonstrate AI.
Re: (Score:2)
Controlled by neural net? (Score:2)
Surely the robots were themselves controlled by neural nets which were selected by Genetic Algorithm, rather than using a neural net to control the selection process itself. Perhaps if I RTFA...
No confirmation (Score:1, Insightful)
This doesn't "confirm" anything about Turing's offhanded opinion.
Have Them Spend More Time With Humans (Score:3, Interesting)
That's what I'm sure my favorite robot SF authors -- Elf Sternberg and D.B. Story -- have planned for their robots. I would love to meet either of their creations.
similar idea for genetic algorithms (Score:2)
Re: (Score:2)
That would make the world's most awesome screen saver. I mean, if you are going to burn the CPU cycles on a screen saver, might as well do it on something you might enjoy watching.
Spell It (Score:2)
Robots might orient themselves but orientating themselves must involve eating potatos while finding their directions.
A simulation I developed around 1987... (Score:5, Insightful)
A simulation I developed around 1987 had 2D robots that duplicated themselves from a sea of parts. They would build themselves up and then cut themselves apart to make two copies. To my knowledge, it was the first 2D simulation of self-replicating robots from a sea of parts. The first time it worked, one robot started canibalizing the other to build itself up again. I had to add a sense of "smell" to stop robots from taking parts from their offspring. As another poster referenced, Philip K. Dick's point on identity in 1953 was very prescient:
http://en.wikipedia.org/wiki/Second_Variety [wikipedia.org]
"Dick said of the story: "My grand theme -- who is human and who only appears (masquerading) as human? -- emerges most fully. Unless we can individually and collectively be certain of the answer to this question, we face what is, in my view, the most serious problem possible. Without answering it adequately, we cannot even be certain of our own selves. I cannot even know myself, let alone you. So I keep working on this theme; to me nothing is as important a question. And the answer comes very hard.""
However, those robots were not evolving. I presented a talk on that simulation at a workshop on AI and Simulation in 1988 in Minnesota, saying how hard easy it was to make robots that were destructive, but how much harder it would be to make them cooperative. A major from DARPA literally patted me on the back and told me to "keep up the good work". To his credit, I'm not sure which aspect (destructive or cooperative) he was talking about working on. :-) But I left that field around that time for several reasons (including concerns about military funding and use of this stuff, but also that it seemed like we knew enough to destroy ourselves with this stuff but not enough to make it something wonderful). At the same workshop someone presented something on a simulation of organisms with neural networks that learned different behaviors. A professor I took a course from at SUNY Stony Brook has done some interesting stuff on evolution and communications with simple organisms: :-)
http://www.stonybrook.edu/philosophy//faculty/pgrim/pgrim_publications.html [stonybrook.edu]
Anyway, in the quarter century almost since then, what I have learned is that the greatest challenge of the 21st century is the tools of abundance like self-replicating robots (or nanotech, biotech, nuclear energy, networking, bureaucracy, and others things) in the hands of those still preoccupied with fighting over percieved scarcity, or worse, creating artificial scarcity. What could be more ironic than using nuclear missiles to fight over Earthly oil fields, when the same sorts of techology and organizations could let us build space habitats and big renewable energy complexes (or nuclear power too). What is more ironic than building killer robots to enforce social norms related to forcing people to sell their labor doing repetitive work in order to gain the right to consume, rather than just build robots to do the work? Anyway, it won't be the robots that kill us off. It will be the unexamined irony.
Evolved Neural Network Brains (Score:2)
---
Robotics [feeddistiller.com] Feed @ Feed Distiller [feeddistiller.com]
Re: (Score:2)