# Can humans compute better than computers? Dave Thomas’s design challenge

I wish to respond to some questions Dave Thomas posted at: Design Challenge Results: Ã¢â‚¬Å“Evolution is Smarter than You AreÃ¢â‚¬Â . This is a continuation of several threads:
Dave Thomas says, Ã¢â‚¬Å“CordovaÃ¢â‚¬â„¢s algorithm is remarkableÃ¢â‚¬Â

Kudos to George Atkinson, Bram de Beer, Paul Flocken, Virgil Keys, Alex Labram, Mike McCants, Ray Spurlin and Kim Van der Linde for nailing it [ideal solution], and showing us that the answer to this tricky problem can indeed be obtained via Intelligent Design!
….
Kudos for finding this solution [1st approximation] go to Roy Thearle, Ray Spurlin, Duncan Buell, Kevin Vicklund, Matthew Vonk, and Salvador Cordova (our official IDM respondent). At a length of 1596.3 units, this shape is only 0.6% longer than the [ideal] formal Steiner solution!

I’m actually pleasantly surprised mine came so close to being optimal!

Dave asks a question:

The first is, why did Salvador go the conventional route, finding web pages that discussed Steiner Trees or Fermat Points, and using trigonometry and algebra, like our other Designers? Why didnÃ¢â‚¬â„¢t Salvador instead simply go get the answer from the Fortran listing of my genetic algorithm,

The answer is complex software doesn’t give the answer without running it (unless of course one’s brain is equipped with a compiler and appropriate hardware), and since I don’t have a fortran compiler on my computer to run Dave’s software, I didn’t run his software to get the answer and instead chose to give it a shot on my own.

Why didn’t I derive the answer from the code snippet? The computer software outlines a strategy that entails many (possibly thousands) of computations. Because such an approach is tedious, the humans attempt to find shortcuts using far fewer calculations. Thus I wasn’t about to recreate the algorithmic process of his software by hand. That’s why I didn’t derive it from the code snippet.

And if Thomas boasts that he reversed engineered my software, that’s more because I chose something that can be succinctly described. I even gave a narrative describing how it works, plus the program boils down to a compact formula :

SUM (i)

where SUM(i) is the sum total of all integers, i = 1, 2, 3, …. 1000

His program can not be alternatively expressed in such compact form. Thus he is again being disingenuous by demanding that I describe his code in the same way he described mine, since his cannot achieve such a compact expression as mine.

Dave Thomas wrote:

What is it going to be like having to go to Bill Dembksi and admit that youÃ¢â‚¬â„¢ve learned the hard way of the true meaning of what Daniel Dennett terms Leslie OrgelÃ¢â‚¬â„¢s Second Law: Ã¢â‚¬Å“Evolution is smarter than you areÃ¢â‚¬Å“?

Thomas has proven no such thing, neither Dennett nor Orgel. Computers can compute certain things faster than us, that is why they exist to help us. For Thomas to argue that evolution is smarter than humans because computers can compute faster than humans is a non-sequitur.

Is an adding machine (or simple hand calculator) smarter than us because it can compute square roots in a flash? No. Intelligence creates adding machines, not the other way around.

Intelligence created the computers and software which enabled the creation of genetic algorithms on computers. Because computers came into existence via human intelligence, in the formal sense, they can’t be said to be smarter than us, because we made them, and never the other way around. Thus the “evolution” in the computer simulation, though faster at computing answers, is not smarter than us, since it was through human intelligence these evolutionary algorithms in the computer came into existence in the first place.

Even after a weekÃ¢â‚¬â„¢s worth of effort, you still couldnÃ¢â‚¬â„¢t find the correct Answer. You were close, but itÃ¢â‚¬â„¢s Charlie Darwin (along with some very Intelligent Designers) who got the cigar.

For Thomas to suggest that I spent an entire week is highly disingenuous, as that is not what I was spending time on for an entire week. But even if it did take me a week because I wasn’t as quick as the others on his website, it’s highly uncharitable to dismiss someone who invested time attempting to respond to his formal request. I credit the others who gave speedy answers. I devoted a few hours on and off to reading his posts and finding information on the internet about this topic, and dusting off some of my trig skills. Thomas should be a little more courteous to people attempting to respond to his formal request and who take the time to study his material and the topic at hand in order to respond to his question. But then trying to belittle me by saying his computers were faster than me at figuring out an answer is bad form on his part, par for Panda behavior.

In any case it was intelligent agents or intelligent agents acting through intelligently designed systems that solved the problem. To claim that it was undesigned Darwinian evolution that solved it is doublespeak given the fact that the evolutionary algorithm was designed.

I predict thereÃ¢â‚¬â„¢ll be some Ã¢â‚¬ËœsplaininÃ¢â‚¬â„¢ to do backstage at Uncommon Descent.

No backstage explaining to do. The little puzzle Thomas offered was amusing. It only shows that computers can compute faster than humans and that some at Pandas thumb can connect dots with lines better than others. What he didn’t prove was mindless undesigned agents can create Genetic Algorithms from scratch.

## 28 Replies to “Can humans compute better than computers? Dave Thomas’s design challenge”

1. 1
Ekstasis says:

And even if real Artificial Intelligence, “learning”, software is developed someday, it still will not exceed us humans, since the design (there goes that word again) of the software itself will require all the logic and reasoning necessary for it to perform the learning. This parallels Mr. Dembski’s mathematical reasoning that shows that the information or logic component necessary to perform data searches must equal or exceed that of the search results application. But no doubt the NDE crowd will continue performing all sorts of slight-of-the-hand tricks using Artificial Intelligence to supposedly demonstrate the finer points of evolution.

2. 2
mike1962 says:

I posted this on Panda:

“How does your challenge relate to the bacteria flagellum, for example, with regards to the question of design? Nothing, as far as I can tell, since nobody knows if the flagellum was designed or not. (Duh.) In other words, your challenge doesnÃ¢â‚¬â„¢t help to answer whether a particular entity IS designed or not. If you set up a test where the fitness algorithms were randomly generated as well as the mutations, and could generate something on the order of a flagellum, that would be impressive….When somebody can find some randomly produced fitness algorithms that use random inputs to generate some complex virtual machines like the flagellum, please let me know. The war will be over…Nobody has ever demonstrated that itÃ¢â‚¬â„¢s even possible for a flagellum to have been assembled the way MET asserts. For that to be true, at least one precise (down the molecule) developmental pathway would have to be demonstrated. For all we know, the laws of nature might actually *forbid* any putative pathway that could be devised due to various chemical interactions. ItÃ¢â‚¬â„¢s an open question.”

So far, I’ve got plenty of flack, been told I’m an ignorant troll, and that my demand for this level of evidence like looking at the blue sky and denying it’s blue. One thing I have *not* received is an answer to my request.

3. 3
Mats says:

One thing that was clearly obvious in the debate Peter Ward had with Steven Meyer, is that Peter seems to fall into the belief that if scientists create life in the lab, that will be evidence that it happened in nature without any guiding intelligence.

One has to be careful with news titles that say “Scientists Create Life In Lab Showing How It Happened In Nature By Purely Unguided Processes”

The tragedy of this is that, for what I have seen, it apears that some people do believe that by creating life in a lab, Chemical Evolution (or Abiogenesis) becomes a fact.

Perhaps Darwinian OOL (Origen Of Life) theorists take advantage of that?

4. 4
scordova says:

Mike1962,

I would like to address your concerns regarding CSI. I appreciate your expertise and participation here. I hope you will be patient regarding the issue because rigor in discussion makes the issue tedious and long. Perhaps we can start semi-rigorously and work from there. I think you’ll be able to fill in the blanks.

To help me answer your questions, can you tell me how familiar you are with Dembski’s writings? I’ll try to respond apporpriately.

Specifically, have you read No Free Lunch which specifically deals with evolutionary algorithms?

5. 5
mike1962 says:

Scordova: “I would like to address your concerns regarding CSI.”

No, I have not read No Free Lunch, or any of his books, except for excerpts, and also discussions related to it. But I do believe I understand the gist of it, however, from an information theory angle.

The problem with the CSI angle as I see it is that you still cannot detect design vs accident by mathematics alone if one’s philosophy is wrongheaded. The concept of information originator and information receiver is irrelevant when one accepts the idea that an *infinite* variety of combinations are possible within the universe, or multiverse. The problem with the MET-only crowd is not their science, even though it is limited, and doesn’t explain what they think it does. The problem with them is their philosophy, which is ridiculous. They embrace a methodological meterialism a priori (which usually turns out to be a philosophical materialism in disguise), and then skip along with the notion that matter *can* combine in just the right way to get the self-replicating process going, and then *can* reproduce and develop along the lines MET suggests. Talk about calling the blue sky a different color. They deny the sky is even possible, let alone blue. There’s no use trying to convince someone that the sky is blue if they don’t believe a sky is possible.

I am agnostic on design. I’m likewise agnostic about MET in many respects. MET fails to explain many aspects of the cellular world. And by explain, I mean explain in detail, of how something like the flagellum *could* (not *did) come together by unguided processes. I’m not saying it can’t. But they can’t show that it can. I’m the agnostic here. MET adherents are the ones making the positive claim about their theory. If it’s true, fine. I’d love to see real proof.

Anyway, CSI doesn’t seem to help a guy like me, since it’s *conceivable* (if particulars are of no importance) that ANYTIHNG can happen given enough universes and enough time. And the anti-ID MET adherents are simply not on the same playing field philosophically with the pro-ID guys.

6. 6
scordova says:

mike1962,

Rigor entails a lot of tedium. Also, what you see on the net both by critics and proponents alike is misleading regarding what CSI is and how it is measured. The most glaring things people don’t realize:

1. CSI is defined via dual binary representation, most conventional information theory is single unary

2. CSI deals with both single-target and multi-target systems

3. CSI avoids committing itself to a doctrine of intelligent agency

Most critiques of CSI misrepresent #1,#2, and/or #3. Added to that, there are some other restrictions dealing with how the space of possibilities is described, and what constitutes a bit of CSI versus what constitutes a bit in a computer.

Here is a discussion which will introduce the concepts:
CSI: Measuring the Paradox of Purpose, Basic and Intermediate Concepts

I highly encourage you to read it, because critics will often knock down a strawman of CSI rather than the real thing.

Your questions are those which I expect many of our readers have, thus I’m glad to entertain them as much as I have time to do.

Regarding multiple-universes, hypothetically it may be possible, but let us consder how science is normally practiced in our universe. If we see mount rushmore carved out, it is hypothetically possible we are in a universe where wind and rain did it, but that usually is not considered the best explanation. ID deals with trying to give what would be considered best explanations according to ordinary practice.

7. 7
bFast says:

One thing that was clearly obvious in the debate Peter Ward had with Steven Meyer, is that Peter seems to fall into the belief that if scientists create life in the lab, that will be evidence that it happened in nature without any guiding intelligence.

Certainly more than “creating life in the lab” is required to produce a serious case the abiogenesis “could have” happened spontaneously. (Note that even if it is demonstrated that life “could have” occurred spontaneously, this would not be proof positive that it did so. However, a strong “could have” would be enough for me.)

Other issues that would need to be addressed before the science of abiogenesis can make the “could have” claim:
1 – The artificially created life would have to be simple enough to have happened by chance. Dembsky has suggested UPB as a measure of “simple enough”, yet I really think that a reasonable simple enough is more like 1 in 10^40 that it happens in one try.
2 – The conditions necessary to attempt to produce life must have viably existed. Note of course that Denton would suggest that such conditions existing would inherently be evidence of the strong anthropic principle at work.
3 – The “artificially created life” would have to demonstrate a “reasonably” never ending pattern of increasing complexity producing increasing robustness. (Certainly the simplest life would be rather fragile.)
4 – A rather reasonable “just so story” leading from that artificially created life to DNA based life. Most logically, also, it should require demonstrating an RNA based life that is a minor devolution from current DNA based life.

If science ever gets to this point, I will accept scientific abiogenesis as reasonably “solved”.

However, I have had my left eye on this topic for many years. I have heard reports that this issue is licked, or nearly licked coming from the scientific community dozens of times. Each time, as I look more closely into what is being claimed, I find the claims to be radically exaggerated. Because of the false claims of success in this matter, I am solidly skeptical every time I hear another claim that pre-DNA life has been created in the lab.

Evolutionists please note, I for one can be convinced to give up on agency — just show me the evidence!

8. 8
John A. Davison says:

I know of not a single subcellular organelle that has ever been known to be present in a form other than its present structure. In that sense they did not evolve but were simply “made that way.” These include the centriole, the centromere, the basal granule of both the cilium and the flagellum, both the eukaryotic and prokaryotic flagella, the ribosomes, the various mitochondria and chloroplasts and other plastids, etc, etc.

We should remember that the original meaning of the word evolution derives from the Latin word evolvo, meaning to unfold as the pages of a book. Books have been written don’t you know. The word “gradualism” should be purged frum the evolutionary lexicon as it has no place there now as in the past. No biological process has ever been gradual. It is known in physiology as the “all-or-none law.”

The proper term to describe both ontogeny and phylogeny is not “evolution” but “emergence.” Both have “emerged” through the controlled release of contained “prescribed” information, a process in which the environment played but a trivial role.

I say down with “evolution” and up with “emergence!”

“Neither in the one nor in the other is there room for chance.”
Leo Berg, Nomogenesis, page 406.

“A past evolution is undeniable, a present evolution undemonstrable.”
John A. Davison

9. 9
Tom English says:

Salvador: “Specifically, have you read No Free Lunch which specifically deals with evolutionary algorithms?”

Bill’s treatment of the Blondie24 study of Fogel and Chellapilla is quite weak. Some here may not be familiar with Blondie24. The researchers set up an evolutionary program in which the population was a collection of checkers strategies (static board evaluators implemented as artificial neural networks, to be precise). The strategies in the initial population were chosen randomly, and they were so bad that most humans could beat them.

In each generation, the fitness of each strategy in the population was assessed. But here’s the key point: There was no fitness function. Instead, each strategy played games against randomly selected opponents in the population. For each game, both players were awarded points at the end. The payoff was 3 points for a win, -2 points for a loss, and 0 points for a draw. After all of the games were played, the strategies with fewer points were culled from the population, and those with more points engaged in reproduction-with-variation.

After close to one thousand generations, the researchers took the best strategy from the population and began using it to play against humans on the Internet. The evolved strategy played under the name Blondie24, and its opponents did not know they were playing against a computer. The strategy established an expert rating, playing better than 99% of human players.

I want to make it clear that the abstract environment in which the program evaluated strategies was the game of checkers. Those strategies that did better in competition were more likely to reproduce than those that did worse. There was emphatically not a fitness function. How could there have been? How can you inspect a single strategy and say how well it will do in play? You have to play strategies against one another to know which are better than others.

The form of evolution implemented by the evolutionary program is often referred to as coevolution. The program is independent of checkers. What does Bill Dembski identify as a source of CSI? He identifies the asymmetry of the payoff function — the fact that a win gains a strategy more (3 points) than a loss costs it (2 points). Does he compute the CSI derived from that asymmetry? No. Does he give any formal argument? No.

Coevolution is a bigger problem for you to deal with than Dave’s program. What are you going to do when I present results obtained with a symmetric payoff function?

By the way, Gil Dodgen has said some nice things about David Fogel’s book Blondie24: Playing at the Edge of AI. Check it out at Amazon:

http://www.amazon.com/gp/produ.....1558607838

10. 10
todd says:

Evolutionists please note, I for one can be convinced to give up on agency Ã¢â‚¬â€ just show me the evidence! – bFast

IDers remind me of the old lady in the Wendy’s commmercials of old…

WHERE’S THE BEEF?

🙂

11. 11
tribune7 says:

Mike1962

information receiver is irrelevant when one accepts the idea that an *infinite* variety of combinations are possible within the universe,

But that is not possible in a finite universe.

or multiverse.

And that is a claim of faith.

12. 12
tribune7 says:

Tom English: The researchers set up an evolutionary program in which the population was a collection of checkers strategies

And they remained checker strategies. You couldn’t use them to win at poker.

Nobody here denies that natural selection is real. The debate is about how much it explains.

13. 13
trrll says:

And that is a claim of faith.

Currently, faith is the only basis upon which one can assume that the universe is singular or multiple, or finite or infinite. We have no method of testing which of these assumptions is true, and there are logically consistent models that go one way or another.

So any conclusion that is dependent upon any assumption regarding the size or multiplicity of the universe should be properly regarded as faith-based, rather than scientific.

At the present time, any scientific statement that bears on the issue would have to be contingent in nature: i.e. IF the universe is singular THEN …

14. 14
scordova says:

Tom mentioned Gil Dodgen, one of the software-engineers and authors at Uncommon Descent. Gil is in the history books of Checkers:

A brief history of computer checkers

There was Checkers by Gil Dodgen, and Colossus by Martin Bryant. Bryant traded his opening book against Chinook’s endgame database, and in the early 90’s, Colossus was probably the best PC program, but it was never developed further. Gil Dodgen went on to create World Championship Checkers with Ed Trice, ….
Murray Cash and the Trice/Dodgen team both computed the 8-piece database by the end of 2001, and after I also finished computing the 8-piece database in early 2002, Schaeffer finally released the 8-piece Chinook database to the public. Nemesis is the current computer world champion.
….
On high-end machines, it should be no big problem to compute the 10-piece database, and there are rumours that the Trice/Dodgen team have already done this.
….

15. 15
scordova says:

Tom wrote of Blondie24:

In each generation, the fitness of each strategy in the population was assessed. But hereÃ¢â‚¬â„¢s the key point: There was no fitness function. Instead, each strategy played games against randomly selected opponents in the population. For each game, both players were awarded points at the end. The payoff was 3 points for a win, -2 points for a loss, and 0 points for a draw. After all of the games were played, the strategies with fewer points were culled from the population, and those with more points engaged in reproduction-with-variation.

Thank you for offering your expertise, Tom, however, I must disagree with the bolded. There was still selection being defined in terms of a teleological goal, thus there was an implicit fitness function as far as I can tell….

There is very very crudely speaking something of a fitness filter in nature, but as Allen Orr unwittingly points out, its not a matter of whether the fittest survive (that is somewhat tautologically true), but whether “selection trades in the currency of Design”. See: Dennett’s Strange Idea.

That is, to this day, we don’t know if natural selection in the wild exists such that it will favor complexity over simplicity or even life itself over non-life! The high extinction rate of complex sexually reproducing creatures casts empirical doubt on whether selection favors CSI over simplicity as a general principle. I’m deeply skeptical that favoring CSI would be the norm for nature. Nature seems more likely to kill life, or at best sustain it. I don’t see it evolving complexity from scratch unless nature itself were designed to do so, but Chaitin’s work on the simplicity of the physical law casts doubt on this. (Chaitin is no an IDer, and he is still holding out hope for a Darwinian solution, albeit he is very skeptical of the prospects).

16. 16
Zachriel says:

scordova: “There was still selection being defined in terms of a teleological goal, thus there was an implicit fitness function as far as I can tellÃ¢â‚¬Â¦.

There are many ways an organism can live to reproduce when in competition with others. Checkers programs live to reproduce by winning more games than the others. It’s as simple as that. And it is important to note, that they can use widely differing collections of strategies to reach the same ends.

17. 17
tribune7 says:

Currently, faith is the only basis upon which one can assume that the universe is singular or multiple, or finite or infinite.

By what we can see and measure, the universe is singular and finite. If you claim otherwise you are either invoking the supernatural (not that there is anything wrong with that) or faith that there are physical truths yet to be discovered (not that there is anything wrong with that either.)

A devout believer in a multiverse is in in the same boat as a YECer, albeit the YECer would be significantly more rational.

18. 18
trrll says:

And even if real Artificial Intelligence, Ã¢â‚¬Å“learningÃ¢â‚¬Â, software is developed someday, it still will not exceed us humans, since the design (there goes that word again) of the software itself will require all the logic and reasoning necessary for it to perform the learning.

Exceed in what respect? This trivial evolutionary algorithm did a better job of designing a short network than multiple human designers. Artificially evolved checkers-playing neural nets play better checkers than 99% of human players. Isn’t this a bit like arguing that a human will never be able to build a machine that travels faster than a man can run?

19. 19
trrll says:

By what we can see and measure, the universe is singular and finite.

Sorry, no. That is like arguing that “by what I can see and measure,” there is no such thing as a platypus, because I’ve looked around my neighborhood, and I have never seen a platypus. The fallacy should be obvious: to logically make such an assertion, I must be able to show that I have looked thoroughly enough that I would have found platypuses if they exist. As the adage goes, absence of evidence is not evidence of absence.

So the claim “By what we can see and measure, the universe is singular and finite” necessarily includes and implies the following claims:

1) We could see and measure other universes if they exist.
2) We can see and measure entire extent of the universe that exists.

Both are false.

A devout believer in a multiverse is in in the same boat as a YECer

I agree, but would extend it to say that such a person is also in the same boat as a devout believer in a universe that is singular or finite.

20. 20
tribune7 says:

by what I can see and measure,Ã¢â‚¬Â there is no such thing as a platypus,

You can see a platypus.

Ã¢â‚¬Å“By what we can see and measure, the universe is singular and finiteÃ¢â‚¬Â necessarily includes and implies the following claims:

No, it doesn’t. What it states is that “by what we can see and measure, the universe is singular and finite.” The physical evidence indicates the universe is singular and finite. If you want to believe otherwise you have to believe via faith.

21. 21
trrll says:

You can see a platypus.

I’ve looked all around my neighborhood. I can assure you that I cannot see a platypus anywhere. So by what I can see and measure, there is no such thing.

hat it states is that Ã¢â‚¬Å“by what we can see and measure, the universe is singular and finite.Ã¢â‚¬Â The physical evidence indicates the universe is singular and finite. If you want to believe otherwise you have to believe via faith.

This sentence is meaningful only if you by what you can “see and measure” you can distinguish whether the universe is singular or multiple, or whether it is finite or infinite.

Can you state one piece of physical evidence that would be different if the universe is infinite? Or multiple? Or both? If you cannot, then it is equally rational to say “By what I can see and measure, the universe is infinite and multiple.” Clearly, an observation that is equally consistent with either of two conclusions does not constitute evidence either way.

22. 22
tribune7 says:

So by what I can see and measure, there is no such thing.

Oky dokey. There’s not such thing as a platypus but there might be multiple universes.

This sentence is meaningful only if you by what you can Ã¢â‚¬Å“see and measureÃ¢â‚¬Â you can distinguish whether the universe is singular or multiple, or whether it is finite or infinite.

Or maybe we are just batteries powering a super computer controlling our reality. Take the red pill to find out.

23. 23
scordova says:

Gil was kind enough to lend his expertise regarding Blondie at:
Artificial Intelligence and the Game of Checkers

Thanks Gil!

24. 24
DaveScot says:

A couple questions for Tom English about the checkers code to see if he was seriously trying to emulate RM+NS.

First of all – exactly how were the variations accomplished on the variants selected for reproduction? If you didn’t use a random number generator to pick a bit in the variant’s code or data space to flip then you weren’t emulating RM+NS.

Secondly, did you emulate random factors in nature that happen to kill the fittest? Did you emulate forest fires and arbitrarily kill whole familes of variants? Did you send in any floods to drown them, volcanoes to bury them, predators to eat them, etcetera?

Third, did you make the fitness landscape more complex than just win – lose – draw at checkers? In nature when an antelope gets faster he also requires more food which is a disadvantage in some situations. Did you try emulate the complexity of the environment that RM+NS in nature must confront?

I suspect what you did was set out with a goal of using an interative process, essentially trial and error with a feedback mechanism to select improved starting points for each successive trial (making slight modifications to the starters each iteration in a blind search for serendipitous improvement), and when it worked you would claim this is just like nature does it. I could have told you 25 years ago that method will work. It’s a bloody obvious method, Tom. Thousands of programmers have recreated this algorithm without ever being taught how. I used to review patent abstracts at Dell and if someone came to me with this algorithm I’d reject it as being obvious to an expert in the field even if I thought it was novel, which it ain’t. Even so, it isn’t really how nature does it. It’s how intelligent agents sometimes look for solutions.

25. 25
Tom English says:

tribune7: “And they remained checker strategies. You couldnÃ¢â‚¬â„¢t use them to win at poker.”

The issue is conservation of CSI. The fact that the strategies remained checkers strategies does not imply that there was no gain in CSI. In particular, I will point out that the specification for all strategies in the initial generation was

“sub-novice checkers player”

and the specification of most or all in the final population was

“expert checkers player.”

Does that sound like conservation of CSI to you?

26. 26
Tom English says:

DaveScot,

“A couple questions for Tom English about the checkers code to see if he was seriously trying to emulate RM+NS.”

I have no idea why anyone would emulate evolution in software. I have written repeatedly that genetic algorithms are abstract simulation models of evolution. (I include the word “abstract” in this forum only to try to get people to understand that modeling is abstract. The modifier is redundant.) A simulation model does not emulate.

“First of all – exactly how were the variations accomplished on the variants selected for reproduction? If you didnÃ¢â‚¬â„¢t use a random number generator to pick a bit in the variantÃ¢â‚¬â„¢s code or data space to flip then you werenÃ¢â‚¬â„¢t emulating RM+NS.”

The term I used above, “reproduction with variation,” is confusing you, I suppose. “Random mutation” is a misnomer, and “reproduction with variation” is an alternative some people are using. The details of the random variation operator are unimportant. The principle is that random variation in offspring is inevitable.

“Secondly, did you emulate random factors in nature that happen to kill the fittest? Did you emulate forest fires and arbitrarily kill whole familes of variants? Did you send in any floods to drown them, volcanoes to bury them, predators to eat them, etcetera?”

Again, modeling is abstraction. You seem to be asking for artificial life here, not an evolutionary algorithm. If you don’t know the difference between the two, Wiki will help set you straight. But I will mention that selection of parents had a strong random component. The strong strategies could die and the weak ones could live.

“Third, did you make the fitness landscape more complex than just win – lose – draw at checkers? In nature when an antelope gets faster he also requires more food which is a disadvantage in some situations. Did you try emulate the complexity of the environment that RM+NS in nature must confront?”

RM+NS does not confront anything in nature. It IS nature. There is no fitness function, and there is no fitness landscape. The system is not mine, and the abstract environment in which individuals were evaluated was essentially Checkers World. The fact that it is not an “emulation” does not detract from the study a bit.

“ItÃ¢â‚¬â„¢s a bloody obvious method, Tom.”

Ever since Darwin.

“Even so, it isnÃ¢â‚¬â„¢t really how nature does it.”

Please tell me, then, the abstract properties of evolution my simulation model should capture.

27. 27
trrll says:

Ã¢â‚¬Å“How does your challenge relate to the bacteria flagellum, for example, with regards to the question of design? Nothing, as far as I can tell, since nobody knows if the flagellum was designed or not. (Duh.) In other words, your challenge doesnÃ¢â‚¬â„¢t help to answer whether a particular entity IS designed or not. If you set up a test where the fitness algorithms were randomly generated as well as the mutations, and could generate something on the order of a flagellum, that would be impressiveÃ¢â‚¬Â¦.When somebody can find some randomly produced fitness algorithms that use random inputs to generate some complex virtual machines like the flagellum, please let me know. The war will be overÃ¢â‚¬Â¦Nobody has ever demonstrated that itÃ¢â‚¬â„¢s even possible for a flagellum to have been assembled the way MET asserts. For that to be true, at least one precise (down the molecule) developmental pathway would have to be demonstrated. For all we know, the laws of nature might actually *forbid* any putative pathway that could be devised due to various chemical interactions. ItÃ¢â‚¬â„¢s an open question.Ã¢â‚¬Â

It would certainly be nice to have a detailed history of the evolution of the flagellum. On the other hand, given the number of degrees of freedom of protein folding and interaction, resulting in an extremely high dimensionality of the Ã¢â‚¬Å“search spaceÃ¢â‚¬Â traversed by evolution, I believe that we can with confidence make the following prediction from evolutionary theory:

The detailed, mutation-by-mutation pathway of evolution of the flagellum (and indeed, or most biological structures) will never be known

(Of course, that doesnÃ¢â‚¬â„¢t mean that we wonÃ¢â‚¬â„¢t be able to come up with models, which have been suggested by multiple researchers e.g.

but they will always contain a substantial degree of speculation. Actually testing every one of the millions of possible intermediate for fitness is an impossible project)

So does this meant that we must reject evolutionary theory because it refuses to provide us with an answer that we would like to have? If so, it will be in good company, because we will also have to discard such things as quantum physics, which perversely refuses to provide us with the exact trajectory of a photon through a slit. And Newtonian physics, which perversely refuses to provide us with an exact solution of the n-body problem, enabling us to accurately extrapolate planetary positions for any time.

But the scientific criterion for rejection of a theory is not failure to answer a question that we would like it to answerÃ¢â‚¬â€it is experimental refutation based on the theory’s predictions. Here, of course, evolutionary theory has no problem, because it serves up a wealth of predictions, testable by methods ranging from computer simulations of genetic algorithms to genomic sequence comparisons (and so far has accumulated a remarkable track record of success).

ID on the other hand has a big problem. Predictions, after all, arise from the limitations of a model: the things that it cannot do. And since the ID crowd, perhaps for political reasons, is unable to get specific about the natureÃ¢â‚¬â€and particularly the limitationsÃ¢â‚¬â€of their hypothetical designer, they have been unable to make any testable predictions. The closest they can get to a prediction is Ã¢â‚¬Å“No pathway for evolution of the flagellum exists.Ã¢â‚¬Â Unfortunately, this kind of prediction is next to worthless, because to test it, they would have to show that they have examined every possible sequence of mutations from every possible set of protein precursors, and determined whether all of them are blocked by low-fitness Ã¢â‚¬Å“chasmsÃ¢â‚¬Â in the fitness landscape. Of course, they have no desire to take on this impossible task themselves. So instead, they turn to biologists, and sayÃ¢â‚¬â€Ã¢â‚¬Å“You must carry out the impossible studies to test our theory; if you canÃ¢â‚¬â„¢t disprove it, then our theory must be right.Ã¢â‚¬Â

But of course, biologists have no interest in trying to carry out this impossible study, because they have a rich theory that makes lots of testable predictions, and are busy testing those predictions and making discoveries. And all the ID guys seem to do is stand on the sidelines and heckle, Ã¢â‚¬Å“Hey, what about the flagellum?Ã¢â‚¬Â

28. 28
trrll says:

Secondly, did you emulate random factors in nature that happen to kill the fittest? Did you emulate forest fires and arbitrarily kill whole familes of variants? Did you send in any floods to drown them, volcanoes to bury them, predators to eat them, etcetera?

Introduction of random factors will slow down the convergence of a genetic algorithm on optimized strategies, but it will not prevent it from converging unless you make the random factors so large that the entire population is likely to go extinct from random factors before it has time to evolve.

Third, did you make the fitness landscape more complex than just win – lose – draw at checkers? In nature when an antelope gets faster he also requires more food which is a disadvantage in some situations. Did you try emulate the complexity of the environment that RM+NS in nature must confront?

Ultimately, natural selection boils down to win-lose-draw in terms of producing more descendants. Of course, there are multiple subgoals required to achieve that, just as a successful checkers playing simulated organism needs to succeed in tactical subgoals (protecting one’s pieces, crowning pieces, jumping opposing pieces) in order to achieve the strategic end of winning the game. And of course, these subgoals must be balanced, e.g. a player that is too aggressive in trying to capture pieces may place itself at greater risk.