MSU Devolab
8 views
Skip to first unread message
Bud
Jun 25, 2008, 4:53:51 PM6/25/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
In searching for information about how this challenge might be
approached in a practical manner, I found a group which seems to have
already qualified for the prize. Their web site is
http://devolab.cse.msu.edu/
They are associated with Lenski and descend from the Caltech Digital
Life Lab. There is a great deal of information on this site and its
links. The Discover article on the home page is a good start.
If this work does not qualify as a solution, could you please explain?
Thanks
approached in a practical manner, I found a group which seems to have
already qualified for the prize. Their web site is
http://devolab.cse.msu.edu/
They are associated with Lenski and descend from the Caltech Digital
Life Lab. There is a great deal of information on this site and its
links. The Discover article on the home page is a good start.
If this work does not qualify as a solution, could you please explain?
Thanks
bkl...@gmail.com
Jun 26, 2008, 11:48:16 AM6/26/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Bud, I looked at Avida long ago, and have spoken several times with
Chris Adami. I read the Discover article when it was new and reference
it in updates to my GECCO paper of 1999 (http://www.panspermia.org/
proof5.htm). Two years ago I met at length with several members of the
Avida team. I have no reason to think that OEEI-QS, by the published
criteria, has been demonstrated. The model does what it is programmed
to do, and that's all. Adami said, 4 June 2006, that if I insist on
Open-Ended, he could simply slow the process down to a crawl.
Seriously. That way, its limitations wouldn't become obvious before
we're all dead, I guess!
If anyone thinks this work qualifies, let him or her explain it. Sorry
to be grumpy. I think there is an important question here, and it gets
only superficial analysis, because people think they already know the
answer.
On Jun 25, 2:53 pm, Bud <bud.buttr...@mindspring.com> wrote:
> In searching for information about how this challenge might be
> approached in a practical manner, I found a group which seems to have
> already qualified for the prize. Their web site ishttp://devolab.cse.msu.edu/
Chris Adami. I read the Discover article when it was new and reference
it in updates to my GECCO paper of 1999 (http://www.panspermia.org/
proof5.htm). Two years ago I met at length with several members of the
Avida team. I have no reason to think that OEEI-QS, by the published
criteria, has been demonstrated. The model does what it is programmed
to do, and that's all. Adami said, 4 June 2006, that if I insist on
Open-Ended, he could simply slow the process down to a crawl.
Seriously. That way, its limitations wouldn't become obvious before
we're all dead, I guess!
If anyone thinks this work qualifies, let him or her explain it. Sorry
to be grumpy. I think there is an important question here, and it gets
only superficial analysis, because people think they already know the
answer.
On Jun 25, 2:53 pm, Bud <bud.buttr...@mindspring.com> wrote:
> In searching for information about how this challenge might be
> approached in a practical manner, I found a group which seems to have
Bud
Jun 27, 2008, 6:41:40 PM6/27/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Well, I also think that there are important questions here. Some of
them are probably unanswerable with our current knowledge. I think
whether panspermia occurred is one of those questions. But the
question of whether evolution can innovate, as contrasted to optimize,
is answerable. Furthermore, most people believe that the answer is
yes.
In what follows, I am going to try to explain how I view this problem
from the point of view of a scientist. To do that, I must first cover
a fair amount of background material and terminology.
First, how do we understand the occurrence of evolution? After all,
one view of evolution is that simple disorganized systems are becoming
more complex and more organized. This does not happen spontaneously!
So how is it possible? The answer is that the less successful
organisms die out more quickly, so that the better adapted ones
predominate. What does “better adapted” mean? It means better able
to survive the competition for resources. Without competition the
payoff for changing is minimal, and evolution slows down or stops.
An example from chemistry might make the process clearer. We know
that carbon is not converted to diamond except at very high
temperature and pressure, yet today people make good money growing
diamond films on machine tools everyday. How is this possible? The
answer is that the diamond film does not form spontaneously; it is
grown under conditions of extreme “natural selection”. In a low
pressure chamber, a carbon containing gas is decomposed on hot
filaments into very reactive species. One of these reacts with the
tool surface to deposit carbon. Most of the carbon is bound in a way
which will produce the thermodynamically favored graphite, but a small
percentage is bound in a way that can lead to the diamond structure.
Another species formed in the reactor reacts very, very rapidly with
the graphite like carbon on the surface, but reacts only very slowly
with the diamond like form. In this way, the unstable diamond film
builds up on the surface. The entire process is kinetically
controlled. The laws of thermodynamics are not defeated, but they are
circumvented.
So evolution is basically a kinetic process. Changes to the organism
occur at random, and different rates of dying relative to rates of
creation produce changes to the surviving population. And, of course,
the changes tend to reduce the rate of dying and/or increase the rate
of reproduction.
One aspect of this process which is very important to this topic is
that the changes observed occur in response to conditions which limit
the reproduction of the population. This is sometimes referred to as
evolutionary pressure. It is a fact that the environment in the
natural world changes frequently on the time scale of evolution. So
the traits and abilities which favor survival of a species change over
time, and it is to be expected that new innovations will evolve over
time in response to changes in the local environment. If the
environment does not change, then a stable population is likely to
develop which will have no need to innovate.
Now, as I understand it, you want to demonstrate a specific type of
evolution in an isolated system. This is termed open-ended
evolutionary innovation/quarantined system (OEEI/QS). Most of the
individual components of this name are susceptible to different
interpretations, making it difficult for people to understand
precisely what you want to demonstrate.
The term “open-ended” is probably the most problematic because there
seems to be no mathematical or operational definition. Your writings
suggest that multiple innovations need to be produced by a system if
it is to be deemed “open-ended”, but there seems to be no way to
quantify how many are necessary. Of course, if the evolutionary
pressures are allowed to change, then we can expect to see changes in
the capabilities of the population in response.
Another difficult distinction is the one you make between “puzzle
solving” and “innovation”. In nature, I think that the current view
is that all evolution is basically changes in the DNA of the
organism. If the genetic code is changed, then the nature and
behavior of the organism changes, and sometimes these changes are
obvious or even dramatic. Most of the time they are difficult to
notice. Yet the genetic code contains just four “letters” or
elements. And all changes to the organism result from changes to the
length and sequence of this code. So evolution may be looked upon as
puzzle solving. The puzzle is how to arrange these four letters in a
sequence that will result in an improvement in the organism’s ability
to cope with its environment. If you randomly cut the Declaration of
Independence into 6 pieces, then putting them back together in the
correct order is a puzzle. But if you completely breakup the genetic
sequence of an organism into its constituent letters and put them back
together in a new order to yield an improved organism, that is
evolution. If the organism has a new capability, then we have an
innovation. The operations are very similar, and the distinction is
not clear. I personally would be inclined to argue that evolution is
just puzzle solving by brute force trial-and-error methods. It is
just a very, very large and complicated puzzle.
Now as discussed earlier in this forum, the only practical way of
establishing a quarantined system is on a computer. Real organisms
are just not sufficiently well characterized, but computer code can be
rigorously scanned and analyzed.
The Avida program seems to me to meet all of the requirements. First,
the code, or instructions, contained in the organism’s genome are
created for the model. They do not exist in the computer, but define
operations on a virtual computer built into the organism. Thus there
is no possibility of any “program” evolved in an organism getting
incorporated from outside, because the “genetic material” of the
organism only exists within the models of the organisms. So the
system is quarantined, rigorously and completely.
Second, in the Avida system, the ability to metabolize a particular
food source is modeled as the ability to perform logical operations on
numbers. However, to begin with, the initial organisms do not have
most of these capabilities in their program. Yet, using a fairly
standard, but well implemented genetic algorithm, Avida organisms
evolve the capabilities to produce the required operations and thus
benefit from the food source. One can examine the little internal
“programs” used by the organism to see how they are performing the
needed operation. Not only have these innovations appeared
spontaneously, but more than one method of reaching the desired result
appears.
Because the competition in Avida includes “living space”, evolution
continues indefinitely, because of crowding. Starting from the basic,
stupid organism, the ability to utilize the various “food sources”
evolves. Some organisms improve by learning to use many food sources,
others use a few sources very rapidly. The entire population evolves,
but maintains a surprising amount of diversity. You can download an
educational version of the software ready-to-run from their web site.
It is both educational and fun, and can easily take more of your time
than you intended.
So, I think that experiments run on Avida would meet the stated
objectives of this challenge. As it stands, one can observe about a
dozen innovations in the code of the organisms with only a few hours
of running on a large XP system. If one built in some constraints and
interactions into the “food supply” (some of which are already in the
research version), then the number of innovations over a long time
should be quite large.
Finally, I think that if you are seeking a computer based
demonstration, then you have an obligation to provide more rigorous
and consistent criteria for success. In particular, you should be
able to explain why the Avida system does not demonstrate OEEI. The
program which runs the model does what it is programmed to do, but the
results vary from run to run because the mutations are random. The
innovations do not result from the supervising program, but from the
evolution of the code in the organisms.
> > Thanks- Hide quoted text -
>
> - Show quoted text -
them are probably unanswerable with our current knowledge. I think
whether panspermia occurred is one of those questions. But the
question of whether evolution can innovate, as contrasted to optimize,
is answerable. Furthermore, most people believe that the answer is
yes.
In what follows, I am going to try to explain how I view this problem
from the point of view of a scientist. To do that, I must first cover
a fair amount of background material and terminology.
First, how do we understand the occurrence of evolution? After all,
one view of evolution is that simple disorganized systems are becoming
more complex and more organized. This does not happen spontaneously!
So how is it possible? The answer is that the less successful
organisms die out more quickly, so that the better adapted ones
predominate. What does “better adapted” mean? It means better able
to survive the competition for resources. Without competition the
payoff for changing is minimal, and evolution slows down or stops.
An example from chemistry might make the process clearer. We know
that carbon is not converted to diamond except at very high
temperature and pressure, yet today people make good money growing
diamond films on machine tools everyday. How is this possible? The
answer is that the diamond film does not form spontaneously; it is
grown under conditions of extreme “natural selection”. In a low
pressure chamber, a carbon containing gas is decomposed on hot
filaments into very reactive species. One of these reacts with the
tool surface to deposit carbon. Most of the carbon is bound in a way
which will produce the thermodynamically favored graphite, but a small
percentage is bound in a way that can lead to the diamond structure.
Another species formed in the reactor reacts very, very rapidly with
the graphite like carbon on the surface, but reacts only very slowly
with the diamond like form. In this way, the unstable diamond film
builds up on the surface. The entire process is kinetically
controlled. The laws of thermodynamics are not defeated, but they are
circumvented.
So evolution is basically a kinetic process. Changes to the organism
occur at random, and different rates of dying relative to rates of
creation produce changes to the surviving population. And, of course,
the changes tend to reduce the rate of dying and/or increase the rate
of reproduction.
One aspect of this process which is very important to this topic is
that the changes observed occur in response to conditions which limit
the reproduction of the population. This is sometimes referred to as
evolutionary pressure. It is a fact that the environment in the
natural world changes frequently on the time scale of evolution. So
the traits and abilities which favor survival of a species change over
time, and it is to be expected that new innovations will evolve over
time in response to changes in the local environment. If the
environment does not change, then a stable population is likely to
develop which will have no need to innovate.
Now, as I understand it, you want to demonstrate a specific type of
evolution in an isolated system. This is termed open-ended
evolutionary innovation/quarantined system (OEEI/QS). Most of the
individual components of this name are susceptible to different
interpretations, making it difficult for people to understand
precisely what you want to demonstrate.
The term “open-ended” is probably the most problematic because there
seems to be no mathematical or operational definition. Your writings
suggest that multiple innovations need to be produced by a system if
it is to be deemed “open-ended”, but there seems to be no way to
quantify how many are necessary. Of course, if the evolutionary
pressures are allowed to change, then we can expect to see changes in
the capabilities of the population in response.
Another difficult distinction is the one you make between “puzzle
solving” and “innovation”. In nature, I think that the current view
is that all evolution is basically changes in the DNA of the
organism. If the genetic code is changed, then the nature and
behavior of the organism changes, and sometimes these changes are
obvious or even dramatic. Most of the time they are difficult to
notice. Yet the genetic code contains just four “letters” or
elements. And all changes to the organism result from changes to the
length and sequence of this code. So evolution may be looked upon as
puzzle solving. The puzzle is how to arrange these four letters in a
sequence that will result in an improvement in the organism’s ability
to cope with its environment. If you randomly cut the Declaration of
Independence into 6 pieces, then putting them back together in the
correct order is a puzzle. But if you completely breakup the genetic
sequence of an organism into its constituent letters and put them back
together in a new order to yield an improved organism, that is
evolution. If the organism has a new capability, then we have an
innovation. The operations are very similar, and the distinction is
not clear. I personally would be inclined to argue that evolution is
just puzzle solving by brute force trial-and-error methods. It is
just a very, very large and complicated puzzle.
Now as discussed earlier in this forum, the only practical way of
establishing a quarantined system is on a computer. Real organisms
are just not sufficiently well characterized, but computer code can be
rigorously scanned and analyzed.
The Avida program seems to me to meet all of the requirements. First,
the code, or instructions, contained in the organism’s genome are
created for the model. They do not exist in the computer, but define
operations on a virtual computer built into the organism. Thus there
is no possibility of any “program” evolved in an organism getting
incorporated from outside, because the “genetic material” of the
organism only exists within the models of the organisms. So the
system is quarantined, rigorously and completely.
Second, in the Avida system, the ability to metabolize a particular
food source is modeled as the ability to perform logical operations on
numbers. However, to begin with, the initial organisms do not have
most of these capabilities in their program. Yet, using a fairly
standard, but well implemented genetic algorithm, Avida organisms
evolve the capabilities to produce the required operations and thus
benefit from the food source. One can examine the little internal
“programs” used by the organism to see how they are performing the
needed operation. Not only have these innovations appeared
spontaneously, but more than one method of reaching the desired result
appears.
Because the competition in Avida includes “living space”, evolution
continues indefinitely, because of crowding. Starting from the basic,
stupid organism, the ability to utilize the various “food sources”
evolves. Some organisms improve by learning to use many food sources,
others use a few sources very rapidly. The entire population evolves,
but maintains a surprising amount of diversity. You can download an
educational version of the software ready-to-run from their web site.
It is both educational and fun, and can easily take more of your time
than you intended.
So, I think that experiments run on Avida would meet the stated
objectives of this challenge. As it stands, one can observe about a
dozen innovations in the code of the organisms with only a few hours
of running on a large XP system. If one built in some constraints and
interactions into the “food supply” (some of which are already in the
research version), then the number of innovations over a long time
should be quite large.
Finally, I think that if you are seeking a computer based
demonstration, then you have an obligation to provide more rigorous
and consistent criteria for success. In particular, you should be
able to explain why the Avida system does not demonstrate OEEI. The
program which runs the model does what it is programmed to do, but the
results vary from run to run because the mutations are random. The
innovations do not result from the supervising program, but from the
evolution of the code in the organisms.
>
> - Show quoted text -
bkl...@gmail.com
Jun 28, 2008, 12:20:34 PM6/28/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Bud, thanks for spending time on this. The question is hard to make
both clear and approachable, no doubt. That's why I'm seeking help
from Innocentive's solvers. One thing that bothers me is this --
making the terms clear should be a problem whether one thinks OEEI-QS
happens, or one has doubts. For an example of unclear terms from the
no-doubt side, you write that "'better adapted' ...means better able
to survive the competition for resources." Nothing survives better
than bacteria, and nothing can utilize a wider variety of resources.
Yet some how "better adapted" is also allowed to mean: having complex
features like wings or lungs.
Suppose I asked-- "Can a single prokaryote evolve, ultimately, into
people, with no genetic input beyond the DNA in the initial
prokaryote?" This is perfectly clear. But it is not approachable ---
too much to manage. Tests would take way too long.
Suppose I ask, "Can encoded meaning come from non-meaning?" How do you
quantify meaning?
My best effort to date is, Is Open-Ended Evolutionary Innovation in a
Quarantined System Possible? I have defined each term carefully. And I
suggest that, if the answer is Yes, a demonstration would be the
surest way to prove it.
However the underlying question is best posed, I am convinced that it
is unanswered and profoundly important. Dismissing it because it is
not precise enough to answer easily is, I believe, a mistake that
prolongs our ignorance.
You write, "If you randomly cut the Declaration of Independence into 6
might be a fruitful distinction to pursue.
Meanwhile, Avida advocates may submit their case to Innocentive for
consideration. I think it only does what it is programmed to do.
> ...
>
> read more »
both clear and approachable, no doubt. That's why I'm seeking help
from Innocentive's solvers. One thing that bothers me is this --
making the terms clear should be a problem whether one thinks OEEI-QS
happens, or one has doubts. For an example of unclear terms from the
no-doubt side, you write that "'better adapted' ...means better able
to survive the competition for resources." Nothing survives better
than bacteria, and nothing can utilize a wider variety of resources.
Yet some how "better adapted" is also allowed to mean: having complex
features like wings or lungs.
Suppose I asked-- "Can a single prokaryote evolve, ultimately, into
people, with no genetic input beyond the DNA in the initial
prokaryote?" This is perfectly clear. But it is not approachable ---
too much to manage. Tests would take way too long.
Suppose I ask, "Can encoded meaning come from non-meaning?" How do you
quantify meaning?
My best effort to date is, Is Open-Ended Evolutionary Innovation in a
Quarantined System Possible? I have defined each term carefully. And I
suggest that, if the answer is Yes, a demonstration would be the
surest way to prove it.
However the underlying question is best posed, I am convinced that it
is unanswered and profoundly important. Dismissing it because it is
not precise enough to answer easily is, I believe, a mistake that
prolongs our ignorance.
You write, "If you randomly cut the Declaration of Independence into 6
pieces, then putting them back together in the correct order is a
puzzle. But if you completely breakup the genetic sequence of an
organism into its constituent letters and put them back together in a
new order to yield an improved organism, that is evolution." This
puzzle. But if you completely breakup the genetic sequence of an
organism into its constituent letters and put them back together in a
might be a fruitful distinction to pursue.
Meanwhile, Avida advocates may submit their case to Innocentive for
consideration. I think it only does what it is programmed to do.
>
> read more »
Bud
Jun 28, 2008, 3:34:26 PM6/28/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Two points:
First, you write, "Meanwhile, Avida advocates may submit their case to
Innocentive for
consideration. I think it only does what it is programmed to do. " If
you have already decided that an aproach like Avida does not qualify,
would it not be a complete waste of time to submit such a solution to
Innocentive? Who is judging this competition?
Second, you write, "However the underlying question is best posed, I
clearly is the essence of good science. Working on vague questions is
akin to exploring aimlessly. You may encounter interesting things,
but you do not progress towards a goal. It is fun but not
productive.
And a final comment: I think that the best analogy between the
biological genetic code and the Avida code is the 4 letter
combinations rather than the individual bases. Each 4 letter sequence
codes for a particular amino acid. There are approximately the same
number of Avida "instructions" as amino acids, so the number of
possibilities in a fixed length of code is about the same. In nature,
evolution occurs by means of random mutations of the genetic code
followed by survival testing of the resulting organism. In Avida,
evolution occurs by means of random mutations of the organism's
program followed by survival testing of the resulting organism.
> ...
>
> read more »- Hide quoted text -
First, you write, "Meanwhile, Avida advocates may submit their case to
Innocentive for
consideration. I think it only does what it is programmed to do. " If
you have already decided that an aproach like Avida does not qualify,
would it not be a complete waste of time to submit such a solution to
Innocentive? Who is judging this competition?
Second, you write, "However the underlying question is best posed, I
am convinced that it
is unanswered and profoundly important. Dismissing it because it is
not precise enough to answer easily is, I believe, a mistake that
prolongs our ignorance." Choosing questions which can be answered
is unanswered and profoundly important. Dismissing it because it is
not precise enough to answer easily is, I believe, a mistake that
clearly is the essence of good science. Working on vague questions is
akin to exploring aimlessly. You may encounter interesting things,
but you do not progress towards a goal. It is fun but not
productive.
And a final comment: I think that the best analogy between the
biological genetic code and the Avida code is the 4 letter
combinations rather than the individual bases. Each 4 letter sequence
codes for a particular amino acid. There are approximately the same
number of Avida "instructions" as amino acids, so the number of
possibilities in a fixed length of code is about the same. In nature,
evolution occurs by means of random mutations of the genetic code
followed by survival testing of the resulting organism. In Avida,
evolution occurs by means of random mutations of the organism's
program followed by survival testing of the resulting organism.
>
> read more »- Hide quoted text -
Keith C
Jun 28, 2008, 5:47:11 PM6/28/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
It is some time since I looked into the details of Avida, but I think
it satisfies all of the requirements of the challenge, except that in
the interests of time etc, the range of possible features which could
evolve was limited.
Only certain mathematical operations were checked for and rewarded if
present. The program would be completely open-ended if all possible
correct mathematical operations or calculations were checked for and
each received an appropriate reward if present.
I think there is an obvious difficulty in trying to program such a
general evaluation program, and I do not see any real conceptual
advantage over specifying only a limited number of tests.
I think the most important question at present is:- Who is going to
evaluate the challenge entries and do they share your opinions?
> ...
>
> read more »
it satisfies all of the requirements of the challenge, except that in
the interests of time etc, the range of possible features which could
evolve was limited.
Only certain mathematical operations were checked for and rewarded if
present. The program would be completely open-ended if all possible
correct mathematical operations or calculations were checked for and
each received an appropriate reward if present.
I think there is an obvious difficulty in trying to program such a
general evaluation program, and I do not see any real conceptual
advantage over specifying only a limited number of tests.
I think the most important question at present is:- Who is going to
evaluate the challenge entries and do they share your opinions?
>
> read more »
ctyankee
Jun 28, 2008, 6:24:20 PM6/28/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Just a few thoughts...
On Jun 28, 12:20 pm, bkl...@gmail.com wrote:
> from Innocentive's solvers. One thing that bothers me is this --
> making the terms clear should be a problem whether one thinks OEEI-QS
> happens, or one has doubts. For an example of unclear terms from the
Yes, it's a problem to make the issues clear. A problem the OP hasn't
solved.
>
> Suppose I asked-- "Can a single prokaryote evolve, ultimately, into
> people, with no genetic input beyond the DNA in the initial
> prokaryote?" This is perfectly clear. But it is not approachable ---
> too much to manage. Tests would take way too long.
>
The test has occurred. look around. I claim success and request the
award.
>
> Quarantined System Possible? I have defined each term carefully. And I
> suggest that, if the answer is Yes, a demonstration would be the
> surest way to prove it.
>
Since you cannot prove the quarantine was compromised, nor can anyone
prove the negative, until someone positively identifies genetic
material from space, in support of the pan-spermia (a nonsense
theory), I submit that you're surrounded by the demonstration.
>
> Meanwhile, Avida advocates may submit their case to Innocentive for
> consideration. I think it only does what it is programmed to do.
>
I agree, and if Innocentive rules against the award, I would tend to
believe that the whole process is fixed or fraudulent at the very
least.
I thought that this was a game in philosophical inquiry. At least if
it was labeled as such, it would have retained an element of
amusement.
Good luck folks, I think I can safely say I've had my fill of this
'going-nowhere' challenge.
On Jun 28, 12:20 pm, bkl...@gmail.com wrote:
> from Innocentive's solvers. One thing that bothers me is this --
> making the terms clear should be a problem whether one thinks OEEI-QS
> happens, or one has doubts. For an example of unclear terms from the
solved.
>
> Suppose I asked-- "Can a single prokaryote evolve, ultimately, into
> people, with no genetic input beyond the DNA in the initial
> prokaryote?" This is perfectly clear. But it is not approachable ---
> too much to manage. Tests would take way too long.
>
award.
>
> Quarantined System Possible? I have defined each term carefully. And I
> suggest that, if the answer is Yes, a demonstration would be the
> surest way to prove it.
>
prove the negative, until someone positively identifies genetic
material from space, in support of the pan-spermia (a nonsense
theory), I submit that you're surrounded by the demonstration.
>
> Meanwhile, Avida advocates may submit their case to Innocentive for
> consideration. I think it only does what it is programmed to do.
>
believe that the whole process is fixed or fraudulent at the very
least.
I thought that this was a game in philosophical inquiry. At least if
it was labeled as such, it would have retained an element of
amusement.
Good luck folks, I think I can safely say I've had my fill of this
'going-nowhere' challenge.
mike markey
Jun 30, 2008, 10:59:04 AM6/30/08
to OEE...@googlegroups.com
On Sat, Jun 28, 2008 at 6:24 PM, ctyankee <sol...@gmail.com> wrote:
>Good luck folks, I think I can safely say I've had my fill of this
'going-nowhere' challenge.
>Good luck folks, I think I can safely say I've had my fill of this
'going-nowhere' challenge.
I have to concur with ctyankee. The more I read of this challenge,
the more absurd I find it. I'm more and more convinced that the asker
is really after some political goal (his blog is panspermia.org , so
that should give you a clue) and is modeling this challenge (poorly)
after James Randi's "Million Dollar Paranormal Challenge".
http://www.randi.org/joom/challenge-info.html I say "poorly" because
there's no way to win. This project proposes two methods to arise at
a solution (computationally and biologically) and then disqualifies
both in the fine print. In the near future look for the existence of
this challenge to be held up as "proof" that evolution does not
innovate, and therefore life comes from space. Meanwhile the
creationists will be chiming in with "this is proof that evolution is
a theory in crisis" or other such rubish.
Check out the asker's conclusions in his article "Neo-Darwinism"
(http://www.panspermia.org/neodarw.htm):
Summary
...A considerable part of Darwinism is not of the nature of an
empirical theory, but is a logical truism. — Karl R. Popper, 1972 (27)
+ Artificial selection never produces wholly new characteristics.
Without the input of new genes, there is no evidence that natural
selection does either.
+ The notion that mutation and recombination can compose new genes is
implausible.
+ There is scant evidence that mutation and recombination can compose
functional new genes that differ from any known predecessor by more
than, say, a dozen essential nucleotides.
+ The evolution of antifreeze glycoproteins in Antarctic cod presents
problems for both Darwinism and Cosmic Ancestry.
+ Evolution does not appear to be gradual, contrary to Darwin's firm prediction.
+ The standard theory cannot explain why the coordinating genes that
control the development of embryos and major features are often very
similar across totally different species.
+ Convergent evolution is a surprise not well-explained by neo-Darwinism.
+ Macroevolutionary progress is not accounted for by neo-Darwinian
microevolution.
No bias there. Nope.
bkl...@gmail.com
Jun 30, 2008, 11:15:28 AM6/30/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
The question is not vague. This was already clear in the previous
post. The goal is to reexamine an assumption in science, that OEEI-QS
is possible and happens. How to do that is not obvious. That's why I'm
asking for help.
What Avida has done so far (last I heard about) does not qualify.
Doesn't mean that an approach like that can't work.
Codons are 3 nucleotides.
Innocentive is consulting for The Astrobiology Research Trust. I am
the Trustee of ART. There are 2 other board memebers. We will listen
to Innocentive's advice and make our own judgement. The challenge is
to design a method of demonstrating something, that's all.
> ...
>
> read more »
post. The goal is to reexamine an assumption in science, that OEEI-QS
is possible and happens. How to do that is not obvious. That's why I'm
asking for help.
What Avida has done so far (last I heard about) does not qualify.
Doesn't mean that an approach like that can't work.
Codons are 3 nucleotides.
Innocentive is consulting for The Astrobiology Research Trust. I am
the Trustee of ART. There are 2 other board memebers. We will listen
to Innocentive's advice and make our own judgement. The challenge is
to design a method of demonstrating something, that's all.
>
> read more »
bkl...@gmail.com
Jun 30, 2008, 8:49:31 PM6/30/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
[My reply of earlier today crossed in cyberspace with Mike Markey's.]
I often hear that OEEI-QS must be logically possible, because, without
miracles, there's no other way to account for life on Earth. Primarily
to dispatch that objection, I advocate and promote another account on
my panspermia website. I'm glad if you have looked at it. And, yes, if
OEEI-QS is not possible, my bet is on panspermia. But the OEEI-QS
question needs answering no matter what the scientific alternatives
are. If you can meet the demonstration-design-challenge, or even
provide a demonstration, go ahead. If not, let someone else do it.
If OEEI-QS is possible, there must be a way to demonstrate it. First
there needs to be an accepted test. I have been reluctant to suggest
specific goals for computer models because that invites a kind of man-
made teleology which would amount to cheating. (Such models would fail
the OE part of OEEI. (Avida is in this category.)) Nevertheless, in
response to earlier posts I have suggested, as an example, at least
one very specific way.
I admit, contamination is a more serious problem for biological media
than I knew until recently. But clever lab-techs may solve that one.
And other media or methods besides computers or biology may be
suitable for the job. The current challenge can be won by any design
of a suitable test.
Finally, the deep discussion on Innocentive.com mentions that the
eventual prize would be offered for a proof of either answerto the
OEEI-QS question, yes or no. So the eventual prize is definitely
winnable. I welcome your considered thoughts about ways to answer the
question.
On Jun 30, 8:59 am, "mike markey" <mikepmar...@gmail.com> wrote:
I often hear that OEEI-QS must be logically possible, because, without
miracles, there's no other way to account for life on Earth. Primarily
to dispatch that objection, I advocate and promote another account on
my panspermia website. I'm glad if you have looked at it. And, yes, if
OEEI-QS is not possible, my bet is on panspermia. But the OEEI-QS
question needs answering no matter what the scientific alternatives
are. If you can meet the demonstration-design-challenge, or even
provide a demonstration, go ahead. If not, let someone else do it.
If OEEI-QS is possible, there must be a way to demonstrate it. First
there needs to be an accepted test. I have been reluctant to suggest
specific goals for computer models because that invites a kind of man-
made teleology which would amount to cheating. (Such models would fail
the OE part of OEEI. (Avida is in this category.)) Nevertheless, in
response to earlier posts I have suggested, as an example, at least
one very specific way.
I admit, contamination is a more serious problem for biological media
than I knew until recently. But clever lab-techs may solve that one.
And other media or methods besides computers or biology may be
suitable for the job. The current challenge can be won by any design
of a suitable test.
Finally, the deep discussion on Innocentive.com mentions that the
eventual prize would be offered for a proof of either answerto the
OEEI-QS question, yes or no. So the eventual prize is definitely
winnable. I welcome your considered thoughts about ways to answer the
question.
On Jun 30, 8:59 am, "mike markey" <mikepmar...@gmail.com> wrote:
> On Sat, Jun 28, 2008 at 6:24 PM, ctyankee <sola...@gmail.com> wrote:
> >Good luck folks, I think I can safely say I've had my fill of this
>
> 'going-nowhere' challenge.
>
> I have to concur with ctyankee. The more I read of this challenge,
> the more absurd I find it. I'm more and more convinced that the asker
> is really after some political goal (his blog is panspermia.org , so
> that should give you a clue) and is modeling this challenge (poorly)
> after James Randi's "Million Dollar Paranormal Challenge".http://www.randi.org/joom/challenge-info.html I say "poorly" because
> >Good luck folks, I think I can safely say I've had my fill of this
>
> 'going-nowhere' challenge.
>
> I have to concur with ctyankee. The more I read of this challenge,
> the more absurd I find it. I'm more and more convinced that the asker
> is really after some political goal (his blog is panspermia.org , so
> that should give you a clue) and is modeling this challenge (poorly)
MangoCats
Aug 11, 2008, 5:50:33 PM8/11/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
I think the primary problem with Avida and other computer simulations
of life is the scale of resources issue.
To develop obvious appendages like wings or lungs requires millions
(billions?) of cells. Even with Gigabytes of memory, if the
"universe" contains a 1000x1000 resource grid, and each grid can hold
a "creature" that is busily evolving, your Gigabyte of memory now has
only 1000 bytes for each creature. This is perfectly acceptable when
working on the scale of single celled creatures, but becomes
intractable if you are expecting this little petri dish to sprout legs
and walk away. You may say that a less than a million evolving
creatures should do the job, but one tooth in your mouth is covered
with millions of plaque cells, and they plaque hasn't been observed
demonstrating any interesting evolution of complexity in a very long
time (longer than several human lifetimes....)
Large scale evolution takes a large number of generations, and a
relatively vast resource pool.
I have read elsewhere that demonstrations of symbiosis are not
acceptable as "innovation" - if not, then what level of cooperative
"cells" would be required? Three? Three million? The same problem
exists for biological demonstrations, the biosphere needs to be larger
than practical if a significant demonstration of evolution of
complexity is desired.
The fossil record suggests that slime didn't evolve into more complex
structures for a _very_ long time. The ocean contains 3.612 x 10²º
gallons of water, expecting an experiment that is 24 or so orders of
magnitude smaller (a computer would have a tough time simulating more
than a milliliter of water), to execute 8 orders of magnitude faster
(say, in 10 years, instead of a billion) when the computer is running
calculations in series, or with a very small degree of parallelisim,
and the water "executes" in full-parallel, might be asking a bit
much. The experimental conditions can be primed and made more ideal,
but overcoming 30+ orders of magnitude is a slanted battle to say the
least.
>
> read more »
bkl...@gmail.com
Aug 12, 2008, 11:18:00 AM8/12/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Hod Lipson & Jordan B. Pollack, who are well-respected among ALifers,
wrote in 2001, "After accumulating several Million CPU hours on this
project and reviewing many evolved creatures we have concluded that
merely more CPU is not sufficient to evolve complexity...." And given
that mammalian populations, for example, are small, it is not dificult
to believe that sufficiently many generations of them could be
modelled on today's computers.
> ...
>
> read more »
wrote in 2001, "After accumulating several Million CPU hours on this
project and reviewing many evolved creatures we have concluded that
merely more CPU is not sufficient to evolve complexity...." And given
that mammalian populations, for example, are small, it is not dificult
to believe that sufficiently many generations of them could be
modelled on today's computers.
>
> read more »
Keith C
Aug 13, 2008, 9:37:05 AM8/13/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
bkl,
If you actually read something about what the Golem project was doing,
eg
http://helen.cs-i.brandeis.edu/golem/design.html
you will find that they use very simple components, only a few 100
generations,
and fitness is only measured as distance moved.
I suspect that the millions of CPU-hours has been devoted to repeated
re-runs of
the same very early stages of evolution in this system. If more
interesting results
are possible, they will only appear much later.
I also suspect that really effective evolution depends on having
something similar
to HOX genes which can control development of the body form and allow
large
changes in morphology. I do not think the Golem project had anything
like that.
> ...
>
> read more »
If you actually read something about what the Golem project was doing,
eg
http://helen.cs-i.brandeis.edu/golem/design.html
you will find that they use very simple components, only a few 100
generations,
and fitness is only measured as distance moved.
I suspect that the millions of CPU-hours has been devoted to repeated
re-runs of
the same very early stages of evolution in this system. If more
interesting results
are possible, they will only appear much later.
I also suspect that really effective evolution depends on having
something similar
to HOX genes which can control development of the body form and allow
large
changes in morphology. I do not think the Golem project had anything
like that.
>
> read more »
MangoCats
Aug 13, 2008, 9:57:25 AM8/13/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Lipson & Pollack may have been a bit short-sighted in terms of "merely
more CPU" - the Earth itself can be viewed as a large parallel
processing engine for the evaluation of physical rules. As I pointed
out, the evolution of life in the ocean took roughly 30 orders of
magnitude more CPU than a reasonable A-Life simulation can manage this
year.
Moore's law (as corrupted over the last 40 years by the press and
industry) postulates a doubling of CPU power every two years - if you
take the leap of faith that this will continue indefinitely (seems
unlikely, unless you consider options like nanotechnology reworking an
asteroid into a computing engine accessible as a "cloud service..."),
then the billion year evolution of life in earth's ocean could be
simulated in less than a year's time by the 23rd century (100
doublings roughly equals 10^30).
My feeling is that Moore's law is driven in large part by economic
forces, its occasional breakdowns have more often been faster progress
instead of slower (such as the quantum leap in hard drive capacity
several years back.) Industry can double computing capacity per
dollar every 2 years with reasonable assurances of success, and they
can feed consumers a steady stream of improved products. If they
"swing for the fences," they would likely make faster overall
progress, at the expense of a few bad quarters when things didn't turn
out as planned. In the world's economic ecology, steady progress of
technology breeds a stronger, more successful corporate entity - the
advantages of innovating vast leaps in technology are overshadowed by
the cost of short term failures. And if you want to talk about a
small population - high technology mega-corporations in a given field
are counted on fingers, usually of one hand.
So, I suppose the point of the ramble is that (assuming continued
exponential technological innovation over the next 200 years similar
to that observed for the last 100 years), by the beginning of the 23rd
century, the average person working in a menial service role in a
"first world country" should be able to purchase, in exchange for
roughly 100 hours labor, enough computing capacity to simulate a
billion years of earth's entire ocean chemistry in less than a year.
By the end of the 23rd century, that simulation time should drop from
one year to 28 nanoseconds. If this sounds outrageous, think of the
computing speed available in 1908 (say, a hand cranked mechanical
calculator) and compare it to today (say, a 3GHz dual core box from
Dell) - then compare it to what is available for the same cost in 2
years (probably a 3GHz quad core box from Dell).
OEEI-QS is a BIG question, and it may take an extensive simulation to
answer it. There are philosophical theories that the observable
universe is actually a simulation inside some other reality - this
isn't unimaginable if you continue exponential growth in computing
capacity from the 23rd century to the 30th... We may be some 30th
century fry-cook's kid's elementary science experiment, just getting
ready for the school fair. The 30th century is closer than the 1st,
and past civilizations have run for 1000 years and more before
breaking down.
Of course, getting off the lofty impracticalities - the biggest thing
that determines what it takes to demonstrate OEEI-QS is the definition
of the question itself, the poser (of question) hasn't jumped up and
said "yes, that's what I'm looking for" to anything demonstrated so
far - even though existing A-life simulations have shown some
remarkable properties in their present state. Given the relatively
vast level of research and resources that already go into A-Life, I
think the poser is going to have to get a lot more focused on near-
term achievable goals if they expect a $100,000 incentive to have any
impact on the field.
>
> read more »
Message has been deleted
i borisen
Aug 20, 2008, 5:09:15 PM8/20/08
to Open-Ended Evolutionary Innovation / Quarantined Syst.
Hello there,
May I join your discussion? My background is in molecular biology and
I've only recently been reading up about this challenge. I'm still not
yet certain exactly what parameters are involved in satisfying the
challenge. The way I understand it is this:
In a quarantined system/environment, is it possible for the contained
system (which may be DNA or any biological construct, or even a
computer system) to innovate?
If the question above is stated correctly based on the challenge of
open-ended evolutionary innovation/quarantined system (OEEI/QS), may I
further enquire into more specific quantifying and qualifying
parameters in the challenge. Such as:
1. Does "Quarantined" mean that nothing goes into the system and
nothing goes out of the system once the experiment begins.
A system such as this can be very hard if not impossible to create in
the lab. Aseptic techniques will not be enough and contamination will
be always be a possibility not only from other biologicals but as
simple as the act of observation by scientists, which in itself
already disturbs the quarantined system. To solve this problem, can I
utilize controls instead of a perfectly enclosed and quarantined
system?
2. How does one qualify "innovation", for example, in DNA or
organisms? Does this pertain to a mutation that permits suitable
survival of the organism? Can any mutation in the DNA qualify as
innovation? Even nonsense mutation, a base pair deletion or insertion
which may not reveal any obvious phenotype?
I ask this because we grow cells in incubators all the time and after
certain number of passages cells can manifest mutations, which may be
due to age, or whatever. Random mutations occur everytime. 99 % of the
time, however, the mutation induces death rather than advantage. Do I
need to capture the 1% survival mutation to win this challenge?
The most straightforward solution I can think of with this challenge
(discarding ALL practical measures to achieve it, only theoretical) is
thus:
Create a Quarantined perfectly enclosed space, nothing goes in,
nothing goes out which contains nucleotides, enzymes, etc, which has
no capacity to replicate. Keep it for a year or so. Afterwards open it
and hopefully find an "organism" or something that can replicate. It
can be as simple as a few base pairs that are able to unwind and copy
itself.
Thanks for reading. Hope to read your comments.
I B.
May I join your discussion? My background is in molecular biology and
I've only recently been reading up about this challenge. I'm still not
yet certain exactly what parameters are involved in satisfying the
challenge. The way I understand it is this:
In a quarantined system/environment, is it possible for the contained
system (which may be DNA or any biological construct, or even a
computer system) to innovate?
If the question above is stated correctly based on the challenge of
open-ended evolutionary innovation/quarantined system (OEEI/QS), may I
further enquire into more specific quantifying and qualifying
parameters in the challenge. Such as:
1. Does "Quarantined" mean that nothing goes into the system and
nothing goes out of the system once the experiment begins.
A system such as this can be very hard if not impossible to create in
the lab. Aseptic techniques will not be enough and contamination will
be always be a possibility not only from other biologicals but as
simple as the act of observation by scientists, which in itself
already disturbs the quarantined system. To solve this problem, can I
utilize controls instead of a perfectly enclosed and quarantined
system?
2. How does one qualify "innovation", for example, in DNA or
organisms? Does this pertain to a mutation that permits suitable
survival of the organism? Can any mutation in the DNA qualify as
innovation? Even nonsense mutation, a base pair deletion or insertion
which may not reveal any obvious phenotype?
I ask this because we grow cells in incubators all the time and after
certain number of passages cells can manifest mutations, which may be
due to age, or whatever. Random mutations occur everytime. 99 % of the
time, however, the mutation induces death rather than advantage. Do I
need to capture the 1% survival mutation to win this challenge?
The most straightforward solution I can think of with this challenge
(discarding ALL practical measures to achieve it, only theoretical) is
thus:
Create a Quarantined perfectly enclosed space, nothing goes in,
nothing goes out which contains nucleotides, enzymes, etc, which has
no capacity to replicate. Keep it for a year or so. Afterwards open it
and hopefully find an "organism" or something that can replicate. It
can be as simple as a few base pairs that are able to unwind and copy
itself.
Thanks for reading. Hope to read your comments.
I B.
Reply all
Reply to author
Forward
0 new messages