Robots evolve communication, cooperation, and deception

[benjdm]

http://sciencenow.sciencemag.org/cgi/content/full/2007/222/1

[ QUOTE ]
Experts disagree over exactly when and how communication arose among social animals. Evolutionary biologists suspect that early communication may have developed as a way for closely related individuals to boost each other's chances for survival. Studying such evolution in the lab is practically impossible, however, because most socially sophisticated creatures, such as bees or monkeys, can take hundreds of generations to show substantial behavioral changes.
.


Enter the s-bots, robots fated to live, reproduce, and die within 2 minutes. Keller and company equipped these 15-centimeter-tall subjects with wheels, a camera, a ground sensor, and a virtual "genome"--a computer program that dictated their responses to their environment. Some of the robots also had blue lights they could turn on or off. The robots then entered a foraging environment consisting of a "food" source and a "poison" source. Robots that found food were "mated" with other successful robots: Their genomes were recombined into new programming for the next generation. Robots that didn't find food, or that found poison, saw their genomes vanish from the game...
.


During the course of 500 generations, or about a week, the robots evolved to use their blue lights to communicate. Some groups flashed them to tell others where the food was; other groups used them to warn of the presence of poison. As the tactic worked and the genomes of successful communicators survived, the robots became more and more efficient at foraging.
.


The researchers expected the lone bots to largely ignore each other. But they were surprised, says Sara Mitri, a graduate student involved in the experiment. Bots acting alone developed the same communication strategies, along with some strategies of deception. When surrounded by their kin, the incentive of trying to get their genome--or one similar to theirs--into the next round of the game kept the cooperation going. But when surrounded by "stranger" bots with dissimilar genomes, they flashed their blue lights far from food to sabotage the nonkin bots' chances for survival. "We did not expect that they would evolve such a sophisticated system of communication," says Keller.

[/ QUOTE ]

Extra periods added between paragraphs just to make it look spaced right.

[madnak]

Woo! I've always said that heuristics and emergent algorithms will be more effective than brain emulation.

[Skidoo]

A stunning confirmation of the viability of intelligent design.

[eurythmech]

[ QUOTE ]
A stunning confirmation of the viability of intelligent design.

[/ QUOTE ] /images/graemlins/confused.gif

[madnak]

[ QUOTE ]
A stunning confirmation of the viability of intelligent design.

[/ QUOTE ]

Fella, they weren't designed. They were given very simple instructions and then they "evolved" into communication as they randomly generated code and let it survive based on a standard of selection.

In other words, if this kind of process creates thinking beings, then evolution can create humans.

[chezlaw]

[ QUOTE ]
[ QUOTE ]
A stunning confirmation of the viability of intelligent design.

[/ QUOTE ]

Fella, they weren't designed. They were given very simple instructions and then they "evolved" into communication as they randomly generated code and let it survive based on a standard of selection.

In other words, if this kind of process creates thinking beings, then evolution can create humans.

[/ QUOTE ]
I think his tryin to point out that all this happened under the auspicious of intelligent designers and hence it confirms that an intelligent designer could be responsible for life on earth.

But of course no-one has ever suggested ID wasn't viable. Its problems are much worse than a simple lack of viability.

chez

[madnak]

Ah, I see. Yeah, it confirms that an unfalsifiable proposition has a possibility of being true. Woo!

[Borodog]

I for one welcome our new blue-lighted robotic overlords.

[Metric]

HAIL ROBOTS

[alphatmw]

why??? why was i programmed to feel pain???

[Skidoo]

[ QUOTE ]
why??? why was i programmed to feel pain???

[/ QUOTE ]

How many times does "the man" have to tell you?

The environmental adaptation process that certain wise ones have determined created you does not require a subjective dimension. Therefore, you don't feel pain.

Now stop thinking, and go into doggie submission.

[abridge]

I wonder how the reproduction process worked.

[Stu Pidasso]

I'd like to see how well these robots perform if you take away the wheel, the ground sensor, the camera, and the blue light. The experiment was designed to reach a foregone conclusion. In fact if the "researchers" had let this expiriment continue they would find their beings "evolve" to a point and then become stagnant even though there is still room to improve.

In all honesty I don't see how its any different from programming a computer that learns by its mistakes to play tic tac toe.

If you want to run a similar expiriment on your home computer you can. Visit this site (http://www.frams.alife.pl/). The program is pretty cool the first couple of days you run it.

Stu

p.s. Just so its understood, I'm not slamming the theory of evolution. Just this half thought expiriment.

[NotReady]

[ QUOTE ]

In all honesty I don't see how its any different from programming a computer that learns by its mistakes to play tic tac toe.


[/ QUOTE ]

Scanning the article reminded me of Dawkins' weasel program. I don't know the details of either but it seems to me they both have one thing in common - an intelligent designer. Which proves that evolution can be designed.

[vhawk01]

[ QUOTE ]
[ QUOTE ]

In all honesty I don't see how its any different from programming a computer that learns by its mistakes to play tic tac toe.


[/ QUOTE ]

Scanning the article reminded me of Dawkins' weasel program. I don't know the details of either but it seems to me they both have one thing in common - an intelligent designer. Which proves that evolution can be designed.

[/ QUOTE ]

And for at least the second time in this thread, I will tell you, "No duh!" Of course evolution CAN be designed. Just look at the automobile industry.

This does nothing to show that biological evolution WAS designed.

[Phil153]

[ QUOTE ]
[ QUOTE ]

In all honesty I don't see how its any different from programming a computer that learns by its mistakes to play tic tac toe.


[/ QUOTE ]

Scanning the article reminded me of Dawkins' weasel program. I don't know the details of either but it seems to me they both have one thing in common - an intelligent designer. Which proves that evolution can be designed.

[/ QUOTE ]
Um...this was never in question. OF COURSE evolution, or any part of evolution, or all the animals on earth, could be designed by a potent being. WTF are you talking about?

edit: vhawk beat me to it /images/graemlins/smile.gif

[NotReady]

[ QUOTE ]

WTF are you talking about?


[/ QUOTE ]

My impression was that the two were supposed to somehow illustrate evolution. Maybe not. Like I said, I don't know the details of the two. So what were they for?

[Phil153]

I think I misread you. You obviously meant "which only means that evolution can be designed". Sorry.

The studies indicate that under certain conditions, lower order systems can self organize into higher order systems. Or purely random behavior with weak, simple feedback mechanisms can self select into systems that are highly specialized. It's kind of a simple proof of concept.

[Skidoo]

[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[Phil153]

[ QUOTE ]
[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[/ QUOTE ]
What else would call random blinking lights vs ones that act as signalling to increase survival - without any specific programming for that end?

I'm not saying this proves that it happens in biology (though it obviously does - egg + sperm -> adult brain shows this to be trivially true), but at the very least it shows that certain basic, lower complexity feedback mechanisms can add new, higher complexity functionality to an entity without specific design of those higher complexity features.

[Skidoo]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[/ QUOTE ]
What else would call random blinking lights vs ones that act as signalling to increase survival - without any specific programming for that end?

[/ QUOTE ]

A repeatable consequence of measurable starting conditions, and completely expressible in those terms. This is not what I would call a "higher order" of anything.

[ QUOTE ]
I'm not saying this proves that it happens in biology (though it obviously does - egg + sperm -> adult brain shows this to be trivially true), but at the very least it shows that certain basic, lower complexity feedback mechanisms can add new, higher complexity functionality to an entity without specific design of those higher complexity features.

[/ QUOTE ]

Increase in complexity is a trivial component of the hypothesized evolution of species.

[Phil153]

[ QUOTE ]
A repeatable consequence of measurable starting conditions, and completely expressible in those terms. This is not what I would call a "higher order" of anything.

[/ QUOTE ]
Interesting. What would you call single cell -> complex multi celled intelligent organisms? I would call it: A repeatable consequence of measurable starting conditions. You appear to be drawing an arbitrary line at the limits of what you can comprehend or what can be replicated in the timescale of a few years.

[ QUOTE ]
Increase in complexity is a trivial component of the hypothesized evolution of species.

[/ QUOTE ]

Skidoo, I'd love to start a separate thread to debate you tomorrow (I'm on my way out). Anyone who denies the wealth of evidence that proves single cell -> human evolution, or the idea that such would be likely solely through the laws of physics and biology, should be a fun adversary.

[Skidoo]

[ QUOTE ]
[ QUOTE ]
A repeatable consequence of measurable starting conditions, and completely expressible in those terms. This is not what I would call a "higher order" of anything.

[/ QUOTE ]
Interesting. What would you call single cell -> complex multi celled intelligent organisms? I would call it: A repeatable consequence of measurable starting conditions.

[/ QUOTE ]

People have been known to do all sorts of strange things. What repeatableness?

[ QUOTE ]
[ QUOTE ]
Increase in complexity is a trivial component of the hypothesized evolution of species.

[/ QUOTE ]

Skidoo, I'd love to start a separate thread to debate you tomorrow (I'm on my way out). Anyone who denies the wealth of evidence that proves single cell -> human evolution, or the idea that such would be likely solely through the laws of physics and biology, should be a fun adversary.

[/ QUOTE ]

Prepare to be vanquished.

[NotReady]

[ QUOTE ]

I think I misread you. You obviously meant "which only means that evolution can be designed". Sorry.


[/ QUOTE ]

What I meant was they prove that evolution can be designed - they don't prove that evolution can happen without design, nor do they prove that evolution did happen without design.

[ QUOTE ]

The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.


[/ QUOTE ]

I'm not convinced of that. From a layman's pov it appears the organizing principle was programmed into the event. At any rate you're starting with a lower order system that was designed.

[ QUOTE ]

Or purely random behavior with weak, simple feedback mechanisms can self select into systems that are highly specialized. It's kind of a simple proof of concept.


[/ QUOTE ]

It seems to me the randomness was programmed for and the "random" events were certain to happen. There's no random mutation that I can see. A simple example would be saying if it's heads I win if it's tails you win - one of the two will happen and the "program" accomodates both possibilities. If that's evolution, fine, but I don't see that either weasel or the robots add anything new - I think we all knew that can happen.

I'm not a scientist so all this is just a layman's take on it. I don't deny that some microevolution has occurred. I also don't see anything in that concept that excludes the necessity of God. Certainly neither of these two "events" prove Godless Darwinism.

[Neuge]

[ QUOTE ]
[ QUOTE ]

I think I misread you. You obviously meant "which only means that evolution can be designed". Sorry.

[/ QUOTE ]

What I meant was they prove that evolution can be designed - they don't prove that evolution can happen without design, nor do they prove that evolution did happen without design.

[/ QUOTE ]

I hate this quote, but "the absence of evidence is not evidence of absence." Just because this doesn't prove that evolution can happen without design, doesn't prove evolution cannot happen without design. I feel most theists here don't appreciate the difference between proving a hypothesis and disproving that hypothesis.

[NotReady]

[ QUOTE ]

I feel most theists here don't appreciate the difference between proving a hypothesis and disproving that hypothesis.


[/ QUOTE ]

What hypothesis is proved?

[Neuge]

[ QUOTE ]
[ QUOTE ]

I feel most theists here don't appreciate the difference between proving a hypothesis and disproving that hypothesis.


[/ QUOTE ]

What hypothesis is proved?

[/ QUOTE ]

Huh? Nothing, nor did I claim there was. All I'm saying is that because this experiment demonstrates a form of evolution was designed, does not prove evolution was designed a priori.

[tame_deuces]

In this study they are just studying the development of communication using evolutionary principles as part of their system.

It's only interesting in the sense that the scientific explanation for the development of communication and language works in a simplified evolutionary setting.

So its just like a practical application study of allready existing scientific principles. I frankly don't think they are trying hard to prove or disprove of anything. Just trying to use it.

Kinda like building a rocket car and trying to break the landspeed record.

It's awesome though. /images/graemlins/smile.gif

[MaxWeiss]

[ QUOTE ]
[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[/ QUOTE ]

Skidoo, once again you are flat out wrong. It has been demonstrated numerous times that complex systems can easily and do often arise from simple systems. For starters, look at James Gleick's "Chaos" and for a simple overview, chapter 2 of "God: The Failed Hypothesis" by Victor Stenger, along with some of Stegner's earlier books.

[Skidoo]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[/ QUOTE ]

Skidoo, once again you are flat out wrong. It has been demonstrated numerous times that complex systems can easily and do often arise from simple systems. For starters, look at James Gleick's "Chaos" and for a simple overview, chapter 2 of "God: The Failed Hypothesis" by Victor Stenger, along with some of Stegner's earlier books.

[/ QUOTE ]

If I was wrong you would have given a concrete example and used less words.

[MaxWeiss]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
The studies indicate that under certain conditions, lower order systems can self organize into higher order systems.

[/ QUOTE ]

Sorry, but that's utter nonsense.

[/ QUOTE ]

Skidoo, once again you are flat out wrong. It has been demonstrated numerous times that complex systems can easily and do often arise from simple systems. For starters, look at James Gleick's "Chaos" and for a simple overview, chapter 2 of "God: The Failed Hypothesis" by Victor Stenger, along with some of Stegner's earlier books.

[/ QUOTE ]

If I was wrong you would have given a concrete example and used less words.

[/ QUOTE ]

You are wrong, many more words are needed to properly convey an example, and finally, that argument is worse than I thought even you would try to use. A few examples are:

1. Minimizing potential energy in a system often creates the double spiral pattern, for example in sunflowers.

2. Minimizing pressure in the distribution of fluids through a pipe creates a specific branching structure, which is found in among other places, the veins of intestines of many animals.

3. And the robots above, with a randomly blinking light, found a way to control the blinking light, then use it to survive, and finally to create a social order around survival and like robots vs. unlike robots.

These are a few examples where one "instruction," or one variable, or one physical process which optimizes energy, can create a complex system from simple beginnings--this happens because of the compounding and the evolution of competition which naturally arises--such as survival of a specific code or compounded sequence or survival of a physical process which happens from its ability to use less energy and thereby reproduce faster.

The fact that you're too lazy to do any research and to try and actually know what you're talking about doesn't make me magically wrong, it just makes you look like the [censored] idiot that you are.

Even I don't know a lot about the topic, but I make it a point to gain enough knowledge to be able to have a coherent discussion about it. I wish I could say the same for the people like you---except of course learning anything would be a threat to your faith. I just wish that your faith wasn't such a threat to others.

[Stu Pidasso]

[ QUOTE ]
3.And the robots above, with a randomly blinking light, found a way to control the blinking light, then use it to survive, and finally to create a social order around survival and like robots vs. unlike robots.

[/ QUOTE ]

Your giving these robots way too much credit. The robots did not have randomly blinking lights. They "had blue lights they could turn on or off". Different algorythms are created to tell the robot when to blink the light and when not to blink it. They could be as follows

find food blink light
find poison blink light
blink light every 3 seconds
blink light every 10 seconds
blink light until food is found
blink light until poison is found

A blinking light algorythm is randomly selected for the first generation of robots. A similiar process happens for other behaviors(i.e. see blinking light move towards it...see blinking light move away from it...etc).

The robots are set loose and the ones that perform the best are "mated" to created the programing next generation.

Its a bunk a experiment though. I could program a group of robots to play russian roulette and only pass on the "genes" of the robots which survives the longest. The robots can choose how many bullits are in the gun. They initial group of robots randomly recieves a gun choosing algorythm from the following set.

Choose the gun with 5 bullits loaded
Choose the gun with 4 bullits loaded
Choose the gun with 3 bullits loaded
Choose the gun with 2 bullits loaded
Choose the gun with 1 bullit loaded

Let the robots choose thier guns and play russin roulette until dead. The robots who last the longest get there "genes" passed on to the next generation. Its not going to take long at all until only "Choose the gun with 1 bullit loaded" robots exist. As slick as that seems it doesn't prove evolved intelligence.

Stu

[AWoodside]

[ QUOTE ]
Let the robots choose thier guns and play russin roulette until dead. The robots who last the longest get there "genes" passed on to the next generation. Its not going to take long at all until only "Choose the gun with 1 bullit loaded" robots exist. As slick as that seems it doesn't prove evolved intelligence.

[/ QUOTE ]

This post is related to what could be an interesting point. Your roulette robots, and from what I can tell the robots in the OP, are certainly being selected against in a way that in many ways is analogous to the evolutionary process, but their 'fitness landscape' if you will, is pre-determined and finite. It seems like in the real world organisms wouldn't be choosing from an already defined landscape, but actively expanding and modifying their available landscape. This can be seen simply in the fact that different organisms have drastically different #s of base-pairs in their DNA strands. Longer strands = more potential information encoding = bigger fitness landscape. I'm sure there are countless other examples of ways in which landscapes are modified. Perhaps to evolve intelligence the robots would need a way not only to navigate a given set of options/parameters, but be able to add options/parameters that aren't covered by the initial principles of their programming.

[PairTheBoard]

You could do essentially the same thing a lot easier with virtual robots being run by computer programs. I'm suprised there aren't a lot a lone wolf guys sitting in front of their computers tinkering around with such experiments. If there were it seems like there should be some coming up with Truly astounding results.

PairTheBoard