Evolution of altruistic cooperation and communication in robot societies

| 68 Comments | 2 TrackBacks

Discovery Magazine reports on a continuation of experiments involving evolvable robots, communication and concepts such as altruistic cooperation and lying.

By the 50th generation, the robots had learned to communicate—lighting up, in three out of four colonies, to alert the others when they’d found food or poison. The fourth colony sometimes evolved “cheater” robots instead, which would light up to tell the others that the poison was food, while they themselves rolled over to the food source and chowed down without emitting so much as a blink.

Some robots, though, were veritable heroes. They signaled danger and died to save other robots. “Sometimes,” Floreano says, “you see that in nature—an animal that emits a cry when it sees a predator; it gets eaten, and the others get away—but I never expected to see this in robots.”

Fascinating how simple processes of variation and selection can explain the evolution of altruism, cooperation as well as cheating. What has ID done recently that increases our understanding of how cooperation, cheating and altruism arose?

Nothing really

I do not have access to the paper yet but others have presented outlines of the work

The research of Floreano and colleagues is reported in the March 2007 issue of Current Biology. The researchers created four conditions for their experiments, varying the relatedness of the robots (how similar their ‘genes’ and programming were) and whether selection was on an individual level or colony level: “In the individual-level selection regime, the genomes of the 20% robots with the highest individual performance … were selected to form the next generation, whereas in the colony-level selection regime, we randomly selected all robots… from the 20% most efficient colonies” (p.514).

‘Deceptive’ communication only evolved when the robots were not closely ‘related’ to each other and selection was on an individual rather than a colony level. In this condition, “an analysis of individual behaviors revealed that … robots tended to emit blue light when far away from the food.” Despite this, and “contrary to what one would expect, the robots still tended to be attracted rather than repelled by blue light… ” (p.517).

Dario Floreano, Sara Mitri, Stéphane Magnenat and Laurent Keller (2007). Evolutionary Conditions for the Emergence of Communication in Robots. Current Biology 17(6), 514-519

2 TrackBacks

Panda’s Thumb has an interesting post on Evolution of altruistic cooperation and communication in robot societies. Here we have dumb robots exhibiting behaviour many would think is exclusively in the domain of humanity. This goes some way to &#... Read More

Evolution of a Moral Sense from Threads from Henry's Web on January 22, 2008 3:06 PM

One of the interesting things I’ve noticed over the years is that scientists who are also believers often tend to resolve theological issues in ways that make the theologians uncomfortable. I can’t call myself a professional theologian, be... Read More

68 Comments

Fascinating how simple processes of variation and selection can explain the evolution of altruism, cooperation as well as cheating.

yes, fascinating.

perhaps you should relay that article to Francis Collins?

Article on Collins and altruism by Gert Korthof.

Personally I see no problem with the idea that moral law evolved but I cannot speak for Collins.

Of course Behe will say that computer simulations of evolution are utterly meaningless (unless their *his* ID-supporting computer simulations),

jeh:

Of course Behe will say that computer simulations of evolution are utterly meaningless (unless their *his* ID-supporting computer simulations),

ROTFL, yes in general ID proponents tend to be a bit inconsistent about this. Perhaps they can take an ID perspective and show that it leads to better predictions?

How would ID go about studying the origin of altruism, especially in light of all the supporting research by scientists?

Ichtyic, do you really have to be so annoying. I agree that Collins arguments are flawed. As I said, I do not care really.. But your response is not appreciated,

If you want to keep trolling do it on the bathroom wall.

As I said, I do not care really

oh pathetic.

no wonder so few of the original contributers post here directly any more.

this obviously isn’t the place to discuss the SCIENCE involved with evolutionary biology any more.

good luck with your tactical battle.

seriously, you post an article discussing the evolution of altruism, yet feel perfectly happy to avoid debating the fact Collins’ argument is directly counter to this?

good luck avoiding that one anywhere you can’t control the flow.

Sorry to go off at a tangent but I’ve got a nit to pick with the linked article at www.talkreason.org/articles/Theistic.cfm.

I found the sentence “However, contrary to Dawkins’ selfish pitiless universe, the actual behaviour of animals towards genetically related individuals can certainly be altruistic and, famously, this has been theoretically derived from Darwinian principles by W. D. Hamilton.”

It’s been a few years since I read the book but as I remember it a large part of The Selfish Gene is *about* Hamilton’s work on altruism, not something that is contradicted by it. What am I missing here, are you playing to popular misconceptions about Dawkins work or is my memory just really bad?

Perhaps they can take an ID perspective and show that it leads to better predictions?

Of course they can do it. God er, cough, uh, cough, the Intelligent Designer, designed us to be altruistic. Simple really.

Oh, that’s not a prediction you say. Oh really. Well of course neither is any other ID prediction.

It was all snuck in by the programmer I tells you. Intelligent front loading by the intelligent robot designer! That proves that altruism couldn’t possibly evolve naturally. Evolution can only produce really mean bad things and everything good must come from God.

Well, I guess I really can’t find the code where the front loading happened. But I’m just sure it must be there. Otherwise these results must be legit. Man that programmer must have been really smart to do this and not leave a trace. Oh well, I just know I’m right anyway, even if I have no evidence. It’s all an evil “Darwinist” conspiracy.

“Danger, Will Robinson!”

(Well, somebody had to say that.)

I think you mean _Discover_ magazine: http://discovermagazine.com/2008/ja[…]-how-to-lie/

Your Magnenat and Keller URL at the end of your article was broken when I tried to click through just now.

this obviously isn’t the place to discuss the SCIENCE involved with evolutionary biology any more.

If you were interested in discussing the science, how come that you went for the religious disagreements with Collins?

Hmm…

Fixed the link to the PDF.

In the trackback to this article, That Humanist gives us some interesting pointers, including the work of Pascal Boyer. Well worth reading through Boyer’s extensive collection of papers.

Matt, you’re correct. There’s a chapter in The Selfish Gene called “Nice guys finish first”. In fact, he remarked that the title of his book was misleading and that he might have called it The Cooperative Gene among others.

Discovery Institute/Discover Magazine, what an unfortunate combination.

Matt: It’s been a few years since I read the book but as I remember it a large part of The Selfish Gene is *about* Hamilton’s work on altruism, not something that is contradicted by it. What am I missing here, are you playing to popular misconceptions about Dawkins work or is my memory just really bad?

No, Your memory is not bad. The second edition of The Selfish Gene has the chapter (12?) Nice Guys Finish First. Talks about iPD game, Axelrod’s seminal paper on the tournament of strategies, explains why it makes sense to cooperate etc. Gets into the correlation of the level of cooperation and the genetic distance between the players. Makes a connection that an insect colony is just like an animal body. With different castes of insects essentially doing what each organ is doing in a body. And once you get the difference between soma line cells and germ line cells, it makes perfect sense. Comes up with curious cases where the genetic distance between sisters to be 0.75 (which is greater than distance between mother/daughter 0.5) and there it makes more sense to make sisters than daughters. Shows that the Queen of the insect colony is not exploiting the workers but really the workers are using the queen as a sister making machine under these circumstances.

Yeah, Dawkins concludes in that chapter, the best way to win is to be non jealous, non greedy, not holding a grouse, but never excusing bad behavior until it changes. Not too different from the advice you would get from any religious scripture. But Dawkins is a favorite whipping boy for the religious fanatics, who have never read his book, but just make up their own conclusions about what the book must be saying based on the title.

What has ID done recently that increases our understanding of how cooperation, cheating and altruism arose?

OTOH, IDiots have taught us plenty about cooperation, cheating and the constraints on altruism. If they only wouldn’t continue to refuse to put their inherent knowledge to a test…

“What has ID done recently that increases our understanding of how cooperation, cheating and altruism arose?”

Well, they are a shining example of just that.

Stuart Weinstein:

“What has ID done recently that increases our understanding of how cooperation, cheating and altruism arose?”

Well, they are a shining example of just that.

Yes, but only because they were designed that way…

Ichthyic.

Yeah, I’ve pretty much given up on pandas thumb. It used to be an interesting place for educated laymen like me to learn and lurk. I can’t quite put my finger on it, but I find myself skipping almost every new entry.

paul fcd

Yes, but only because they were designed that way…

Ergo, not all designs are intelligent…

(Did I say that?)

Henry

What works, works. What doesn’t, don’t.

Most of the time any predictions are overshadowed by the reality, like this morning’s stock market. What is surprising in nowsight is familiar, sometimes boring, in hindsight. That is, when, and immediately after, something happens it has a sense of urgency that is lacking in later recall. This is sometimes offset when later recall, and gleaning through memory, can glibly restate the very genesis of the event.

This particular process is similar to what little I do know of biology and the story of life as we can reasonably say so far. In biochemical processes, and now in lines of code, running bots, it is shown that there are usually more ways to get the job done than the one most recently discovered. We already know that there are many ways to make and inject venom, there are many ways to make a living inside a host, there are many ways to make a buck. Oh, yeah, there are many, many ways to make babies.

Interesting how some common human inclinations, like making an honest living, (or stealing one) have analogs in artificial simulations. Oh, wait. These simulations were based (at least in some part, I assume) on the principles and predictions of genetics derived from the Theory of Evolution. Well, then. This all should come as no surprise.

This is, for some, a source of great awe and surprise, followed later by a sense of belonging, inclusiveness, welcome. They are related to the entire universe on the most intimate level, made of the same stuff, carrying its imprint, giving birth to its great-grandchildren. And one day they will return all of their parts, fundamentally, to the fundament. And rejoin a grand cycle.

Not woo. Just a notion borrowed from others and polished by experience. And wonder. Pedestrian. Common. Everyman. Everyone wrestles with the miraculous, the numinous. The best bet is to beet feet away from belief and go contemplate a clear, moonless sky.

Give Kids Science Books!

If you were interested in discussing the science, how come that you went for the religious disagreements with Collins?

The evolution of altruism is a scientific topic, and to the extent that Collins rejects the science, it is a scientific disagreement. That his motives are religious doesn’t alter that fact. Else we may as well just shut up shop here.

Syntax Error: not well-formed (invalid token) at line 1, column 59, byte 59 at /usr/local/lib/perl5/site_perl/5.16/mach/XML/Parser.pm line 187.

Yes, it seems weird to have a post about the science of the evolution of altruistic behaviour, and then cry foul when it’s pointed out that Collins has claimed that altruism must have come from God. This sort of study makes it pretty clear that Collins is factually wrong on that point. If people are going to claim that theistic evolution is scientific then it needs to face up to this sort of issue.

In Nature two weeks ago there was a very good paper on the coevolution of choosiness and cooperativity in a population.

The coevolution of choosiness and cooperation p189

John M. McNamara, Zoltan Barta, Lutz Fromhage & Alasdair I. Houston

doi:10.1038/nature06455

http://www.nature.com/nature/journa[…]re06455.html

Has Collins actually made strong statements that altruistic behaviors cannot evolve? Has he made a direct comment on this particular research, or on other actual research directly studying altruistic behavior?

Or has he just made vague statements along the lines that Jesus makes people love each other?

I know that Collins is guilty of publicly endorsing the “anthropic principle” or whatever it’s called (*the logically flawed idea that we can know whether or not the universal constants we observe were “improbable” at some prior non-deterministic point in time before they existed, that it would mean anything if they were (even if we could know that), and that we have any clue whether any other combination of constants could have allowed sentient or even human life*).

However, that particular piece of illogic has been accepted by a fair number of brilliant, productive scientists. Unlike creationism, it doesn’t deny science (*arguably it denies or at least misunderstands probability theory to some degree, if math is science, I guess*) - doesn’t deny the history of the universe or the measurement of the constants.

It’s merely a demonstrably wrong religious idea that makes reference to science, but in a non-denying, non-destructive way.

Likewise, if Collins is going around saying that Jesus is one of the reasons that people are sometimes kind to each other, so what?

Has he actually denied scientific explanations for altruism, or altruism research?

For the record, I don’t share Collins’ religion, but I think it’s his own business, and I don’t see why he even bothers to layer on illogical arguments.

However, if his illogical arguments are not associated with actual denial or dismissal of any legitimate branch of scientific research, then addressing them does construe a personal religious argument with Collins, rather than a relevant comment on the evolution of altruism.

Has he denied any altruism research, or stated openly that altruism can “only” be brought about by magic? If not, who cares if he thinks that sometimes it’s Jesus?

Has Collins actually made strong statements that altruistic behaviors cannot evolve?

Certainly seems that way:

I had to admit that the science I loved so much was powerless to answer questions such as “What is the meaning of life?” “Why am I here?” “Why does mathematics work, anyway?” “If the universe had a beginning, who created it?” “Why are the physical constants in the universe so finely tuned to allow the possibility of complex life forms?” “Why do humans have a moral sense?” “What happens after we die?”

[quote] I had to admit that the science I loved so much was powerless to answer questions such as “What is the meaning of life?” “Why am I here?” “Why does mathematics work, anyway?” “If the universe had a beginning, who created it?” “Why are the physical constants in the universe so finely tuned to allow the possibility of complex life forms?” “Why do humans have a moral sense?” “What happens after we die?” [/quote]

I suppose it might make a difference what it is meant by a “moral sense” and the explanatory power of the evolution of altruism. If altruistic behaviors can evolve, does that basically spell the end of that? For instance, they do say that when selection was on an individual level and people were “unrelated” to each other cheating robots did exist. So its not like altruism is a necessary component of evolution.

One could say that we are altruistic because we evolved that way. Ok. But now what? Why is it that a person cannot say something along the lines of “I’ve evolved this way to feel that I should be altruistic, however I realize I can benefit myself more if I am not, if I’m a cheater. After all, I’m only around for one generation. Why should I care about my children? I don’t even HAVE children!”

Hamlet:

Societies would enforce it because under conditions of general altruism, majority of individuals fare better then otherwise.

Hitler was evil because he caused the death of millions, which, as most people think, he didn’t have right to do.

What if Hitler’s actions really would have benefited the “majority”? Or at least, he acted in that way? It’s one thing to say “Hitler was an evil mass murdering a-hole”… another entirely to say “He was a decent guy (as he tried to act to have the majority fare better than otherwise), but just wrong”.

I have a problem with “pragmatic” morality. Moral codes which judge whether or not an action was “right” or not purely based on its benefit/detriment to a “group”.

My problem is this: If Hitler would have won - would he still be evil? Wouldn’t Jews eventually go the way of all those Old Testament genocides which some people still can’t see anything wrong with?

We consider this appaling now, but it’s just as appaling thought that the trend might reverse in the coming centuries and slavery might be moral again. (Are you going to bet it can’t happen?)

No, I wouldn’t bet against it.

So this is another question: Is there a morality valid once and for all? Morality that you could use to judge all ages? I tend to believe that there’s not.

Here is where we disagree. I think.

I think the disagreement may come from the following: You seem (I think) to think that morality is a feeling that we have about an action, not a property of the action itself. Yet, when I see morality used in discourse, it is as if morality is a property of the action. That is, just as a liquid may be acidic or basic, an action may be “good” or “bad”. Now, while this may depend on the entire situation the action is in (If I kill someone it may be self-defense, it may be murder… and so on). I find the idea that morality is a feeling we have about the action to “eliminate” morality (in a sense). While we can perhaps, have common agreement to have the same feelings on the same actions, its just a “contract” we have, of sorts.

Yes, a social contract. This would imply that morality standards can differ between cultures, and that’s what I see. Not only cultures in geographical sense, but also in temporal sense. 21st century Europe. Ancient Greece. Aztec civilization in America. Middle East in Biblical times. India at the height of caste system. Early 19th-century America. All had moral codes that were, to a certain point, incompatible. And I’d wager that every one of these societies would consider their own code the best possible, and all the other abhorrent. 21st century Europe is appalled by the standing of women and slaves in Ancient Greece. Greeks would be horrified of Aztec human sacrifice. Aztecs would be angry at Jews that they are not sacrificing any of their citizens, thus forcing Aztecs to take the whole brunt of appeasing the Sun God on themselves. Jews would abhor the idea that Indians eat pigs, but not cattle. Caste Indians might find 19th century America dangerously oversimplified. And, as I said, 19th century America would find us appalingly immoral in areas of sexuality and human rights.

After all, wouldn’t the slave-holders find us appalingly immoral? Wouldn’t they deplore the sexual permissiveness of our society where noone enforces or even CARES whether a bride is a virgin or not? Wouldn’t they be puzzled by the idea that you can kill a black man and be actually PROSECUTED for it as if he killed a human being?

Of course they would. It seems then we have several possibilities.

The slave-holders were completely right, we are wrong.

We are completely right, and the slave-holders were wrong (say on caring about the virginity of the bride).

Ain’t none of us perfect (We both hold some moral beliefs, and some immoral beliefs).

We were both “right” as morality has changed by the situation/times.

We were both “right” as morality is what you make of it.

The last one seems to be moral relativism. The second situational ethics. The third from the last is what I personally believe (and I’ll admit, is similar to the Christian idea of mankind being fundamentally “corrupt”. We’ve never quite “gotten it” so to speak). The first two seem to me to be endorsing objective morality, but at the same time, I can’t say that either our culture, or the slave holders, were really “completely” right.

There’s one more possibility. Morality might not be perfectly ordered system. There might not be a single numerical value that you assign to morality, and claim that this action is more, less, or the same moral as another. Maybe that some actions are simply incomparable, the same way like when you can’t order complex numbers by size.

Could that be a possibility?

Marek 14:

My problem is this: If Hitler would have won - would he still be evil? […]

Hitler would not be evil, had he won, at least I presume so. On the other hand, dropping nuclear bombs in towns that had no military (including flak guns) might be considered a major crime against humanity.

Regards

Eric

The New York Times recently had an excellent article about the function of morality in human society that I think is very apropos to this discussion. Simply put, there appear to be five basic moral functions; the relative combinations of which can describe pretty much all morality-related behavior in human society. And it’s not hard to see how each of these five functions could have evolved naturally.

The Moral Instinct

Altair IV — Thanks for the link!

Altair IV - Yes, very interesting article!

So no, there’s no rational, self-interested reason why you should ever sacrifice anything for anyone else if you can help it, but hopefully your brain is already wired to do so anyways.

So altruism is not rational, but its just biological? Why heed my “wiring”? I may not heed it in other cases, why do we priviledge our “altruism” wiring?

Moreover, it may be that far, far in the future we will have progressed to the point where we can change our “wiring” to however we want. Should we keep altruism?

My problem is this: If Hitler would have won - would he still be evil?

Hitler may not have been perceived as evil had he won. Now, if morality is all in our perception then he would not have in fact, been evil. But then again, if morality is all in our perception, can we really say that ANY moral perception is “wrong”?

And I’d wager that every one of these societies would consider their own code the best possible, and all the other abhorrent.

True. That however, just speaks of our perceptions. Not reality. We know that there is a difference in other cases, why is it that in regards to morality we consider them one in the same?

Could that be a possibility?

That could indeed be a possibility, however I would say that it ends up being the same as moral relativism. Since we cannot compare, we cannot judge and morality just ends up being what an individual person makes of it.

Hamlet:

So no, there’s no rational, self-interested reason why you should ever sacrifice anything for anyone else if you can help it, but hopefully your brain is already wired to do so anyways.

So altruism is not rational, but its just biological? Why heed my “wiring”? I may not heed it in other cases, why do we priviledge our “altruism” wiring?

Moreover, it may be that far, far in the future we will have progressed to the point where we can change our “wiring” to however we want. Should we keep altruism?

My personal vision of the future toys with the idea of increasing our “personal power” (what each of us can, personally, achieve without relying on others). Basically, with enough sophisticated nanotechnology, plus maybe an ability to augment our minds and increase our inteligence beyond our natural limits, a person could become completely self-sufficient, both physically and mentally.

If this would come true, then we would no longer need other people, and society would effectively cease to exist. In such a world, altruism wouldn’t serve any purpose and could be abandoned.

I think Isaac Asimov wrote about similar non-society in Naked Sun. I think this would be pretty natural state for interstellar travelers - you won’t worry about returning to your planet thousands years after you left if you had no bonds to it in the first place.

Notice how no new information has been added to the robots. The robots are still robots. Evolutionism would require the robots evolve into something higher like human beings or perhaps HAL 9000. This is not evidence of the possibility that humans evolved from monkeys or monkeys evolved from amoebas or amoebas evolved from rocks!

About this Entry

This page contains a single entry by PvM published on January 21, 2008 11:39 PM.

FSU prof: Gibbs memo is ‘hollow threat’ was the previous entry in this blog.

Anti-Evolution Gains Momentum in Florida is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Categories

Archives

Author Archives

Powered by Movable Type 4.381

Site Meter