Genetic algorithmy is something great from a theoretical perspective:
you get to supervise the evolution of given, simple "species" (of solutions in this context).
It makes some things about evolution quite obvious.
I have casually wondered, as many people probably, "why do I have to bear such an idiot?"
Well, something that will remain a revelation to me is the actual answer to that question, given by genetics and Darwin
Because evolution would not work without a "failure rate". People who get a Darwin award by removing theyre genes from the gene pool are actually causing
damage to the gene pool, because aside whatever makes them idiots, they "contain" information about unexplored paths of evolution.
And at some point, the handsome-clever Mr. Perfect, could turn out to be an error of evolution, and could just be "cancelled".
In these occurences, Mr. Bean himself could contain what Mr. Perfect lacked to be "the fittest" or rather, give birth to the next fittest individual.
So? Evolution needs defects, imperfections, different beauties, taller, shorter, fatter, uglier, thinner, brown, blue, orange, green people, evolution needs difference, it loathes
"perfect" dead-ends like Mr. Perfect.
I've learnt to really, really cherrish difference a whole lot more through this project. Thought I would share.
From: Happy Hunting Grounds... Insane since: Mar 2001
posted 06-05-2006 12:42
I would tend to say, that long-term, sustainable Evolution is dependant on change.
That being the one constant (hehe...if constant change is in itself not an oxymoron).
Therefore, you are right in that Evolution needs difference - because one cannot always foresee in what manner things may change (though certainly sometimes one can). What was well-adapted yesterday, struggles today, and may die out tomorrow if it is not capable of adapting. What was defective before, may be imperfect today, and beautiful tomorrow.
The real interesting thing for me is the term Pefection.
What is it?
What IS Evolutionary perfection?
WebShaman | The keenest sorrow (and greatest truth) is to recognize ourselves as the sole cause of all our adversities.
- Sophocles
Perfection doesn't exist, it's all a contextual thing A perfect being is nothing more, nothing less, than ideally suited to it's immediate environment right now.
At least, perfection doesn't exist in a quantum physics perspective (since nothing keeps a permanent state).
[side note] For GA, to explain how I came to this conclusion... perfection in the case of GA is "perfect suitability of a solution to a problem".
In Sudoku, it is the "perfect grid" that fully matches Sudoku rules.
Such a perfect grid can only be achieved if one of the evolutionary operators (one of the things that evolves the algorithm and changes genes and chromosomes, eg. mutation or crossover or...)
shifts elements by one unit at a time, maximum two (raise this and your GA ends up trying and failing forever).
So some options created by my GA are wrong enough to need to step back, one gene at a time, and restart on a new track, one gene at a time. Due to the randomness of mutations,
this happens very slowly, and returns on tracks that were not expected to produce satisfactory results.
[/side note]
Well, for human beings, we could prove evolutinary perfection doesn't exist by assuming it does exist.
Then, it would consist in "being adaptable to any environment contained in reality, anytime, anywhere"...
So, let me gather my thoughts here... This is all very interesting, and it lead me to ask myself the following questions:
1. What do you do of human evolution so far? Is human kind better suited to its "Context" than 2000 years ago or just differently suited? In other words, was human kind less adapted to its "Context" 2000 years ago?
2. If so, can we assume that human kind (and the other species, but this is not my point here) tend to perfection?
3. If it does tend to perfection, what is Perfection going to be like?
4. I agree with what you said about the need for weak links in the chain, but consider this: if human kind tends to perfection, does it mean it goes inversely exponential at some point? The more "sophisticated" or "perfect" (by lack of better term) we get, the less able to evolution? Therefore the less adaptive, and the less "perfect" - to quote your definition, which I find interesting:
quote:Perfection doesn't exist***, it's all a contextual thing A perfect being is nothing more, nothing less, than ideally suited to it's immediate environment right now.
Isn't perfection beyond being ideally suited to one's immediate environment at some point? Isn't it more global, like being ideally suited to any environment at any time?
5. Assuming the "balance" is somewhat ideal now (because it seems like it is, right here, right now, but is it an illusion?), is our effort to approach perfectness by modifying genes and selecting embryos actually going to lead us to some stagnant state?
*** [edit] I think perfection does exist; at least, I think we can reasonably assume the concept of perfection is realistic; it is just not measurable.
quote:The parallel existence of two concepts of perfection, one strict ("perfection," as such) and the other loose ("excellence"), has given rise ? perhaps since antiquity but certainly since the Renaissance ? to a singular paradox: that the greatest perfection is imperfection. This was formulated by Lucilio Vanini (ca. 1585 ? 1619), who had a precursor in the 16th-century writer Joseph Juste Scaliger, and they in turn referred to the ancient philosopher Empedocles. Their argument, as given by the first two, was that if the world were perfect, it could not improve and so would lack "true perfection," which depends on progress.
quote:
does it mean it goes inversely exponential at some point? The more "sophisticated" or "perfect" (by lack of better term) we get, the less able to evolution? Therefore the less adaptive
Hmmmm. This holds true for Sudoku and most genetic algorithms at least. The closer you get to the solution, the slower the resolution progresses, and it is exponential.
In a population of Sudokus, what happens, because what determines theyre "fitness" is based on a lot of fixed input and very strict rules, is that "excellent beings" tend to look alike
(since they all tend towards the same unique solution).
So the reproduction makes for a lot lense genes exchange... and mutation becomes the only source for evolution.
On to adressing this:
quote:
perfection is realistic; it is just not measurable.
I disagree, and this sentence echoes what I meant about quantum physics perspective.
In quantum physics, what exists exists because it is measurable (and the other way round), the rest is "undefined", or better yet, is subject to the Paradox defined by Schroedinger, and is "everything it could be at the same time"
(the cat is dead and alive at the same time).
In other words, while classic physics (physics being taken, here, as the science which most accurately describes reality) could accept what you've said as perfectly true,
it doesn't make sense at all in quantum physics: "translated" to that perspective, it would lose sense on the "not measurable" bit because then...
it would not exist, it would exist in many different possible states depending on how you observe it. Eg. a perfect object would either not exist, or would
be perfect only for the purpose for which you observe it.
scratches head....
quote:
Isn't perfection beyond being ideally suited to one's immediate environment at some point? Isn't it more global, like being ideally suited to any environment at any time?
This one is difficult and interesting too.
Being suited to any environment anytime is what I want to prove wrong... and it can be proven wrong: there are parts of the universe where nothing, regardless of it's form, can keep existing,
or can keep existing in a controlled state. Simple examples: the inside of a Sun where matter is transformed all the time and physic laws bend, or the inside of a black hole,
where matter breaks through space and time. A couple of places in the universe simply don't allow a living being to keep on living, and the same applies for anything material you put in such conditions.
But what you asked makes a lot of sense anyway:
being ideally suited to one's environment will "fail" as the rest evolves.
That says a lot about the purpose of death and reproduction: evolution would not work without them.
The human model has managed, I think, to reach a far better adaptation than any other living specie, one good indicator is that we have one of the broadest populations on this planet
(insects and bacterias do quite well too, and are, indeed, assumed to be potential replacers if our species disappears).
And what makes the human model a "bit better" is that while we can't evolve our genetic structure over time,
we can, during a lifetime, evolve our behavioral / psychic structure (it is intended to evolve), making us able to use our "deterministic" bodies in very different ways depending on the needs.
....that's all I can come up with as an answer to your questions, it leaves ground for plenty of other questions and reactions.
I think that this conversation really misses the point behind what evolution is.
It seems that there is an assumption here that evolution has a goal....that evolution is something that is on a path from one place to another....
Evolution is not goal driven.
Evolution happens because mutation happens.
Mutations that prove suitable are passed on because the mutated specimen survives. Mutations that are unfavorable are not passed on because the specimen does not survive.
This is clearly not a strict system, since clearly plenty of unfavorable traits can be passed on.
but it is essential to keep in mind that evolution is not about acheiving perfection - it has no goal, no end, no ideal.
We can subjectively define what we feel would be an 'evolutionary ideal', but that will no effect on what actually happens, so it is purely academic anyway.
quote: kimson said:
If so, can we assume that human kind (and the other species, but this is not my point here) tend to perfection?
Hardly, and this is the biggest stumbling block to understanding evolution, IMO.
In the big picture, our existence here has been short and insignificant. To say that we as a species represent perfection in any way is pure folly.
Are we better suited to our environment now than 2,000 years ago? Hardly
2,000 years is far too short a time to measure the effects of evolution on a species such as ours anyway.
quote: _Mauro said:
The closer you get to the solution, the slower the resolution progresses, and it is exponential.
Once again, this assumes a goal
It also assumes that humanity is headed directly for that goal.
Truly: this line of thought, while it can be interesting to follow, is pure folly.
I think people forget about adaptation when speaking of evolution. I personally don't believe in evolution but I do believe that all humans will grow to adapt to environments and situations. Mr. Perfect is only perfect in the situations he's involved in. Sure he can program a VCR and punch all those buttons on his cell phone but his dad or grandfather probably can't. Doesn't mean they aren't as good as he, they have the same genes. If they were put into situations where this might be a useful thing then they would adapt. Then again there are some people who are exceptionally good at some things for no apparent reason. And why do I put my socks on just like my grandfather that passed away before I was born?
I think people forget about adaptation when speaking of evolution.
and what is it that leads you to think this?
quote: At0mic_PC said:
I personally don't believe in evolution
what does belief have to do with it?
Evolution is observable scientific fact.
The particular course a given species may have taken throughout the evolutionary process is quite open for debate. That evolution happens really isn't.
quote: At0mic_PC said:
Then again there are some people who are exceptionally good at some things for no apparent reason.
Which....once again has nothing to do with anything whatsoever. Evolution is not about you acquiring a particular talent
Again: evolution does not have a goal. There is no generational progression toward talent...
Your manner of putting your socks on, unless you have developed a 3rd foot on which to place them, has nothing to do with evolution.
Roflmao... well, while I have kept referring to the Sudoku example, and indeed, to the idea of an "evolutionary" goal,
which, as you said, does not exist per se, there are a couple more things that can be said about evolution.
Humans have their "objective function", eg. the evolution of species on earth is determined by "fitness functions" as in code somehow.
That fitness function is, in turn, defined by available ecological niches.
Eg. earth a certain "life capacity". It can be represented as a n dimensional space.
For example, most species live "best" between 20 degrees and 40 degrees. Draw an axis that goes from 0 to 40 degrees,
and you can use this to draw a "curve of ideal temperature" for a given species, the curve having a peak around 20 degrees for most mammals.
Then draw another axis, dimension, with a similar curve and values for pressure conditions, you get a "curved plane" which is the niche for the observed species
in given conditions of temperature and pressure.
Then draw a third curve for another parameter that impacts the ability of a species to survive (moisture, sun exposure, whatever..)
And so on... the n axis diagram you'll get is the ecological space where your species belongs, and the bizarre n dimension volumes elaborated from all the curves
will be the ecological niche occupied by that species.
This representation is assumed to be finite: earth is one finite ecosystem.
So when a species disappears, it's niche is left to be filled by the others, and this will "push" evolution somehow, acting a bit like my fitness function above.
Obviously, due to the complexity of this object function, the fact it relies on pretty uncontrollable factors and all,
there is no final evolutionary goal to species on earth, just sporadic gaps to fill in the ecological niche (and implicitely, billions of human on earth closed
niches for the many species that disappeared during our reign).
It's difficult for me to draw conclusions from this, namely because it is pure folly
I am merely trying to gather the few facts to expose about the topic.
A biologist and friend of mine mentionned the fact evolution has no clear goal, that's granted, there is no "ideal" to real evolution,
just casual, and temporary goals.
Still, it's everything but a random process, and that's what makes it usable in computer science.
It's not a permanent lottery with luck-a-lot all the way, it's more like a "well designed fitness machine" pushing beings to adapt over generations based on some controlled, limited random factors.
Hey, one more fact. I discussed all this with a biologist and friend of mine.
While she agreed with DL that "hey, stop, there is no final goal to evolution", she also totally seconded the concept of evolution as
a "clever process" rather than a casual miracle, and told me the following story.
Some species seem to keep an intact copy of given genes, while a second copy is left free to mutate over generations, at some stages.
This echoes what I said about required regression.
However, often, when the species has evolved, the mutated gene serves a purpose, and the original gene does as well, but in many cases, it's completely different of it's original purpose.
Just as if evolution had even more subtle mechanisms to try, succeed, then move on, and recycle old mechanisms into brand new purposes.
One more, weirder fact.
In organic chemistry, eg. in the study of all molecules considered "organic" (carbon/oxygene/hydrogene/nytrogene compounds), there is a tendency to chain and mix and
connect and build up more elaborate organic molecules.
A sort of tendency of atoms of life to get together.
This has been reported to happen sporadically in space, even: I remember a cloud, several light years-wide, of alcohol that had been observed
a few years ago
So while blakc holes or Suns aren't place that could hold life, the rest of the universe, including space, could.
quote: _Mauro said:
there is no final evolutionary goal to species on earth, just sporadic gaps to fill in the ecological niche
That's a very good way to put it.
You do raise a good point in regard to temporary ideals. It is essential to keep in mind that temporary nature though. It is also important to note the idea that species capable of surviving in certain conditions survive in those conditions. Those that are not, don't. So while it's easy to get caught up in the fact that species live in areas they are well suited for, and and exclaim the miracle (whether religious or scientific in nature) of their extraordinary adaptation, or of the extrordinary genius of thei rdesigner , it must be remembered that the problem is approached on all sides by nature.
Species become more suited to their environment by the combination of fortuitous mutation which is passed on, and by unsuitable mutations which are not.
Only those species suited to their environments survive - which does not sounds as miraculous as it does when only the well suitedness of the survivors is discussed.
From: Happy Hunting Grounds... Insane since: Mar 2001
posted 06-08-2006 10:53
quote:Species become more suited to their environment by the combination of fortuitous mutation which is passed on, and by unsuitable mutations which are not.
Only those species suited to their environments survive - which does not sounds as miraculous as it does when only the well suitedness of the survivors is discussed.
Actually, this is not true.
Unsuitable mutations can be passed on if they are dormant. In fact, we see that there IS some sort of process to Evolution - it builds on the past adaptions, producing more and more variables - and these variables can be both beneficial and unsuitable, active or inactive, dependingly.
Compare the DNA of a single-celled microbe from 1 billion years ago to the Human DNA of today.
We ARE seeing some sort of simple-to-complexity process going on. It doesn't really matter why and how it is happening - it is happening. We can actually see it.
This doesn't necessarily mean that Evolution is moving towards or away from something. But it does indicate that things start simple and then tend to become more complex. We see this for example in the way that the simple eye has evolved into a very complex, very capable eye (and there are examples where the eye has degenerated and almost disappeared, like in blind cave fish, for example - but the DNA for the eye is still there, just mostly dormant).
On a large scale, and viewed over time, there does seem to be a process of simple-to-complexity going on in the Evolution that we can witness and measure.
WebShaman | The keenest sorrow (and greatest truth) is to recognize ourselves as the sole cause of all our adversities.
- Sophocles
Exactly what I failed to express above, thanks WS. We can witness it, measure it, AND apply it to information science, and it works. It's not a loto or totally random thing at all,
it's exactly what you put very, very well: a mechanism of nature that, aimless or not, is extremely clever rather than random.
Lil' summary, so far, the most valid conclusions are:
1) perfection doesn't exist in real world
2) evolution is not random, but is a natural process that changes beings adapted to one context, to beings adapted to another context
3) difference does good to a gene pool: uniformity kills evolution or slows it down a lot.
These are the only facts, so far, that can be proved to a certain extent, and stand out of the philoso-brainstorming.
For human beings, a LOT of what makes our minds and behaviors is genetic. A lot more than anyone would think, so, to give another answer to the following question...
quote:
1. What do you do of human evolution so far? Is human kind better suited to its "Context" than 2000 years ago or just differently suited? In other words, was human kind less adapted to its "Context" 2000 years ago?
Simple example: are you ever aware of the ongoing temperature regulation in your body? Well, your body, nervous system, every cell *knows* what's going on in terms of temperature,
but we never have to think about it. It just happens.
I think our minds and therefore our societies and cultures have evolved a lot in the past thousands years, a whole lot.
But I think our genes and *inner* behavioral patterns didn't follow at the same speed, nope.
In other words, we ARE better suited to our context in our minds and culture, somehow (proof: demography, the race grew a lot within the last thousand years).
But not in our guts, we're the same cavemen our ancestors were, or almost.
Proof: mating rituals. Did they really change that much I don't think. Females are still attracted to dominant males "regardless of what common sense would tell them", some of them being abusive
jerks (I am all against the *abusive bit), but still...
And males are essentially attracted to *nameless hotties (Playboy magazines make a living of this).
This is controversial theory, but such mating rituals (as opposed to dating and pseudo-romantic bullshit that does not work when it comes to seduction)
suggest the following:
- males are supposed to play a protective role towards the family, hence the fact dominant males are perceived as better reproductors to the woman's instinct
- women are supposed to give birth and nurture, hence the fact women that look good look like "good genes and care for the children" to a male's instinct
This still works 100% for most mammal species, and to an extent, behind society, morals and ethics, I strongly think it still is what determines mating in humans.
FWIW.
Oh, lil' warning... I expect many males to totally disagree with this theory, find it ridiculous, etc.
I expect kimmy, or girls, to understand what I am talking about a bit better
(simply because society, morals and ethics taught us that a good male/father is someone kind and romantic. Go tell that to a woman's instinct )
2) evolution is not random, but is a natural process that changes beings adapted to one context, to beings adapted to another context
Have to disagree with this one. This is a big part of what I was saying above. The mutations that are most useful to a species in adapting to their environment are the ones which allow them to survive. The random aspect of evolution seems less random when you only look at the survivors, but when you figure in any unsuccessful variations which led to less suitable versions of the species (which as a result died and did not pass on the mutation), you see the lack of a goal far more clearly, and the lack of purpose far more clearly.
The point:
Evolution does not cause species to become more adapted to their environment. Natural selection is the name for the process through which those species more suited to their environment survive, and those less suited perish. This leaves an array of species well suited to their environments. Not because evolution made them so, but because their suitibility allowed them to survive.
It may sound like semantics but it is actually an essential point.
FWIW.
quote: _Mauro said:
I think our minds and therefore our societies and cultures have evolved a lot in the past thousands years, a whole lot.
But I think our genes and *inner* behavioral patterns didn't follow at the same speed, nope.
I *strongly* disagree with this one.
Our societies are little different in the modern world than in any time in the past. Some minor differences, and of course the technology which makes things *seem* so different, but which in fact is mostly just variations on themes...
Our minds do not function differently on any real level than they did 1, 2, 5, 10 thousand years ago. it is estimated that it takes around 20,000 years for any significant change in mindset (if I recall correctly from the text books).
We like to think, as a species, that we have come a long way becuase of the trappings we are able to wrap around our redundant acheivements. It's the same line of thought that leads us to believe we are the chosen science fiar project of any number of gods
We aren't as special as we like to think...
quote: _Mauro said:
This is controversial theory
aside from not seeing much in the way of an actual theory there, I see nothing controversial. There are basic facts out there, based in our biology and genetics. ~shrug~
quote:
It may sound like semantics but it is actually an essential point.
It even is the crux of it all, but really, selection is treated as a part of the process called evolution in my understanding, only a part of it, not something distinct.
In terms of AG, it's an "evolutionary operator" along crossover and mutation namely (like + or - or * or / for integer numbers).
I do look at it in terms of "how it impacts a gene pool": in the continuity of what I said so far, a niche appears in an ecosystem, some beings of a given species
are close to fitting that new niche, then reproduction, including sub-processes of evolution like mutation and cross-over, will cause one of the new individuals to be "really close" to fitting that niche,
which will eventually lead the species, through selection, to create more individuals that fit that niche.
The point? Disagreement with this:
quote:
Evolution does not cause species to become more adapted to their environment.
It doesn't seem to, it seems "accidental" but is a lot less random than what it appears to be. Once a niche, a gap is open, evolution *aims* at least at the short term goal
of filling it (which doesn't contradict the fact it doesn't make sense to look for a final and absolute goal to evolution).
How, who and why are to determine, and it's the job of evolution - wether such an adaptation seems predictable or not.
What I mean is that as random as the steps seem to be, as casual as natural selection seems to be, it seems to be so because we don't have means to draw the big picture in details,
but it flawlessy *wants* and *works* at filling niches: regardless of what will work for that niche, evolution will build beings that can fill it.
That's why GA works: what you define in GA is the exact specification of the niche you want to be filled (eg. the rules of Sudoku), and an evolutionary simulation
(that includes selection) will flawlessly reach that state -without your program *knowing* it's aiming at that solution- the fittest will invariably be the solution (and it willl invariably occur in a certain amount of generations).
Go figure, but it's really a clever process, not a permanent chain of accidents: it's a "fitness machine" that balances nature and the ecosystem all the time.
That's the whole magic. Give evolution some beings and an ecosystem, and it will balance them evenly no matter what.
quote:
I *strongly* disagree with this one.
Our societies are little different in the modern world than in any time in the past. Some minor differences, and of course the technology which makes things *seem* so different, but which in fact is mostly just variations on themes...
Hey, what about the invention of language?
What about religion, and it's early purpose as a "families consolidator", which, to an extent, was favourable to the development of our societies in the first centuries after J.C. (before becoming a way to abuse and control masses).
What about the invetion of money to rule commercial exchange?
What about all those processes, social structures, school, law and a legal system, politics, what about moving from a few hundred cavemen to billions of humans getting so much power
they can impact (and potentially destroy) the work of nature?
I don't think these are only variations on a theme, I really think these are major non-biological changes.
And I wasn't meaning to say we were so special,
but merely pointing out the fact that we don't go hunting anymore, we don't ride horses anymore, etc.
quote:
aside from not seeing much in the way of an actual theory there, I see nothing controversial. There are basic facts out there, based in our biology and genetics. ~shrug~
Theory: mating is not based on "formal communication" but essentially on informal communication (eye contact, looks, gestures, voice..) and instinct (and therefore, hasn't changed much since cavemen came out of theyre caves).
And yes, this can be defended as a scientific theory.
Science accepts what can be proved as *fact* and the rest as theory, worthy an investigation or not.
To biologists, this is not strict scientific fact. Because while it applies to all mammal species and can be experimented, it's a lot harder to experiment it on human beings.
It is assumed true, but untested.
To psychologists, this is highly controversial. It will irk some, because they do attach a strong power to the mind, while in reality, conscious thinking has little to do with mating.
To most male humans, society teaches the opposite of what it takes to seduce, and that's what I am trying to point out when stating a "gap" between culture and reality.
Mommy tells her babies to be kind, gentle.
Romantic, I mean, twisted and wrongly named romantic cliches instruct males to "buy gifts", set "rrrrrromantic dates in restaurants",
be servant knights to theyre princess, etc. The crappy dating classics.
95% of movies and modern books tell us tales that look like this. Example: Mel Gibson in "what women want". Women do want a girly guy according to that movie.
There is a whole buzz on those wacky romantic practices.
You do know it, I do know it, but society teaches the opposite: women want dominant males, not wussbags or servant knights.
Try to be like Mel Gibson in "what women.." and you'll land somewhere between enjoying a curious dry feeling where your legs meet, or meeting an abusive woman that will drain and dump you.
Ok, it *is* obvious to you DL, it is not to 9/10 mens I meet.
quote: _Mauro said:
Once a niche, a gap is open, evolution *aims* at
i guess for now we'll have to agree to disagree. this line of thought is counter to everything I have studied in regard to evolution, and to common sense even, IMO.
quote:What about the invetion of money to rule commercial exchange?
What about all those processes, social structures, school, law and a legal system, politics, what about moving from a few hundred cavemen to billions of humans
You said thousands of years.
What you are talking about there has taken far more than "thousands" of years...
quote:
Theory: mating is not based on "formal communication" but essentially on informal communication (eye contact, looks, gestures, voice..) and instinct (and therefore, hasn't changed much since cavemen came out of theyre caves).
I agree fully on this point. I've said it many times: we *are* animals, and we do behave as such.
Ok, point 2 and 3 clarified. I am all for "agreeing on disagreeing", still... maybe I am simply failing at making my point of view clear, or understanding yours?
Let's try to oversimplify things.
Do you think and learnt that evolution is totally random and accidental?
OZONE Asylum
posted 06-05-2006 11:20
thanks kimmy.
, it must be remembered that the problem is approached on all sides by nature.