Some Questions on Genetic Algorithms

vjtorley:

I was very struck by Glenn Williamson’s [vjt meant GlenDavidson] remark that creativity is not the same thing as complexity. Very deep. Glenn seems to think that people are good at the former, but the blind processes can outdo them in the latter. That’s an interesting view, but I’d want to see evidence that blind processes are actually capable of producing systems with a high degree of functional complexity, of the kind Axe described in his book. Even a computer simulation would be something.

What with all the experts in writing GA’s here at TSZ I was hoping VJT would have elicited more of a response.

Which brings me to my second point. Tom English remarked above:

“What’s more important in responding to Axe, I suspect, is the issue of knowledge. Do you get only what you know how to make? The answer to that, coming from evolutionary computation, is a top-of-the-lungs NO!!!. I’ve set up a number of evolutionary systems that ended up knowing, in clear operational terms, how to do what I hadn’t a clue how to do.”

I’d like to ask Tom: in terms of building functional coherence, what’s the best your algorithms are capable of doing? I’m interested in finding out, and if you can point me to a good place to start familiarizing myself with GAs, I’d be grateful.

Why doesn’t someone walk VJT through designing a GA that does what he is asking? Is it because a generic GA doesn’t do anything of the sort of thing he is asking for?

For VJT, do you have any experience at all in any programming language?

268 thoughts on “Some Questions on Genetic Algorithms

  1. Allan Miller: And … how does Natural Selection only occur to an individual?

    Since evolution is about populations and not individuals, it should be pointed out that any GA purporting to model some aspect of evolution must have a population that changes, and not just an individual.

    Weasel is a rather special case in that the population bottlenecks to one in every generation. The mathematics of this is important if you are claiming to model evolution, but nothing is illuminated by word lawyering.

  2. petrushka,

    Since evolution is about populations and not individuals, it should be pointed out that any GA purporting to model some aspect of evolution must have a population that changes, and not just an individual.

    Indeed, pointing that out has been the thrust of my last couple of dozen posts! 😀

  3. Mayr spent a lot of time harping on “population thinking.” Apparently not enough.

    I am a bit tired of creationists who don’t take the trouble to understand evolution, but who feel qualified to criticize it.

  4. EAs ad GAs don’t have to model biology. This has been pointed out many times, but apparently not enough.

    If you want to solve a problem, you are going to optimize the program for solving the problem. It may not look like biology at all.

    Since we can’t yet model biochemistry in real time, a GA that models some aspect of evolution will not be able to solve any practical problem. It will not, for example, be able to design pharmaceuticals or design enhancements to real genomes.

    A good model of evolution will mimic population genetics on a purely abstract level.

    It would be helpful if critics of evolution would take the trouble to understand what is being modeled before spewing crap.

  5. Mung,

    Besides failing in your attempt to code a Weasel and contradicting yourself regarding your intent, you also failed to demonstrate the power of cumulative selection in your program.

    1) Your program doesn’t evolve a phrase; it evolves individual letters, one after the other, latching each one when it matches.

    2) There is a separate fitness function for each letter.

    3) The fitness functions don’t reward proximity to the target — they only reward an exact match for a single character.

    The only thing your program demonstrates the “power” of is latching, not cumulative selection.

    It’s a remarkable display of incompetence.

  6. keiths: The only thing your program demonstrates the “power” of is latching, not cumulative selection.

    Latching is a mechanism whereby they project their real (mis)understanding of how things are. When KF was presented with a video of the original Dawkins Weasel changing letters from the correct character he just invented quasi-latching instead. I think their problem is an inability to separate out how they think things are from how things actually are. Poor scientists, the lot of them!

  7. Creationist: “Oh, I accept microevolution. It would be foolish to deny.”

    Programmer: “Here’s a GA. It’s a problem-solver utilising mechanisms inspired by microevolutionary processes. Here’s another one. It models microevolutionary processes”.

    Creationist: “Oh, no. Don’t accept that at all.”.

  8. You see, phoodoo, if you hire a detective to search for a missing person, you’re really the one doing the searching, not the detective.

    That’s also why if you hire for murder, they come after you.

    Computers are just dumb machines carrying out a list of instructions. And to the extent that the world is like that, perhaps we can say our models in some sense capture reality.

  9. OMagain: Latching is a mechanism whereby they project their real (mis)understanding of how things are.

    The program was no more intended to to reflect how things are than was the Dawkins Weasel program.

  10. keiths: Besides failing in your attempt to code a Weasel and contradicting yourself regarding your intent, you also failed to demonstrate the power of cumulative selection in your program.

    Just because you can quote-mine it does not follow that I was contradicting myself.

    And whether or not my program demonstrates the power of cumulative selection has not been settled, given that you cut and ran from the original thread once it was demonstrated that you had no objective empirical test of whether cumulative selection is present and in what measure.

    I don’t know why you’d want to bring up your failures again and remind everyone of them without so much as a word in your “I Can Admit My Mistakes” thread.

  11. Mung:

    And whether or not my program demonstrates the power of cumulative selection has not been settled…

    Anyone who understands cumulative selection can see that it doesn’t, because your fitness functions don’t reward proximity to the target — only an exact match. The fitness landscapes are flat except for a spike at the site of the target.

    To put that in familiar terms, you’ve failed again.

  12. keiths: Anyone who understands cumulative selection can see that it doesn’t, because your fitness functions don’t reward proximity to the target — only an exact match.

    Thank you for that feedback. It was far better than what you had to offer in the original thread.

    ETA: Will you now admit your mistake?

  13. Mung,

    The program was no more intended to to reflect how things are than was the Dawkins Weasel program.

    And yet Dawkins Weasel was in a book that gave it context. Yours has none of that and is a transparent ploy.

    When you get this published in a best selling book, wake me up.

  14. keiths: Your incompetence is remarkable, Mung.

    And yet what you do is deliberate and intentional, keiths.

    Your disciples seem to now think you’re fallible after all.

  15. Nice to see keiths finally admit his DriftWeasel isn’t actually a Weasel program at all.

    Keep them coming keiths!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.