Lethean
Posts: 292
Joined: Jan. 2014
|
I know jack all about programming, I'm not even a biologist, but I've always been pretty good at visualizing concepts and being fairly able to differentiate and also relate between a model or map and the landscape itself. In Mung's latest Weasel thread it's apparent he's trying not to see the bigger picture and he completely misses how it's related to the very demonstrable reality we all live in. Which to be honest if one were trying to "get it" it should be pretty evident and light bulb would come on.
In a post with a number of queries, petrushka asks...
| Quote | | Is there any conceptual problem with replacing a static target with a moving target? |
Mung responds...
| Quote | | No. Does that help or hinder when trying to demonstrate the power of cumulative selection? Perhaps a moving target Weasel will be next up, now that I finally have the hang of this Weasel programming. |
keiths tries once again to clear up Mung's confusion that while Weasel has a target, cumulative selection does not require one. Or in relation to reality perhaps it's not a hard target but a fluid one, as in survival in an ever changing environment.
| Quote | This is just Mungish confusion. Weasel has a target, and so its fitness function rewards proximity to that target. That does not mean that cumulative selection requires a target, as is obvious to those of us who actually understand the concept. The target and the proximity-rewarding fitness function are characteristics of Weasel, not of cumulative selection generally.
Another example was your challenge regarding what would happen if we changed Weasel’s target phrase in the middle of a run. Biological evolution operates in changing fitness landscapes, so if you could somehow show that cumulative selection was thwarted by such changes, you could argue that it’s insufficient as a selective mechanism for biological evolution.
But as anyone who understands Weasel could predict, it simply starts tracking toward the new target, and I even provided a feature in my Weasel that allows you to experiment with this. Your weird single-character latching program, by contrast, can’t handle changes to the target phrase, because any characters it has already latched are latched for good.
Since the target gambit failed, you are now dicking around with weird fitness functions that are designed to impede convergence while still (you hope) qualifying as instances of cumulative selection. As Joe and I have pointed out, that tactic also fails. |
A bit earlier in the thread Keiths tries to explain the logic problem Mung is having.
| Quote | What is the point of these asinine fitness functions? Well, Mung is engaged in abject definition lawyering, hoping that he can find some scheme that technically qualifies as “cumulative selection” but nevertheless fails to locate the target in a reasonable time. In this case he is trying to find a fitness function that satisfies the criterion “rewards proximity to the target” while still causing the program to fail.
What he is too dim to realize is that his entire project is predicated on a simple logic error. No one has claimed that cumulative selection always succeeds. So even if schemes using those idiotic fitness functions actually did qualify as “cumulative selection”, their failure to find the target wouldn’t show what Mung wants it to show.
If were smarter, he would have realized that you don’t need to concoct Mungish fitness functions in order to cause cumulative selection to fail. You can do it via much simpler interventions, like setting the mutation rate to an extremely small value.
The problem (for Mung) is stark and obvious: the fact that cumulative selection can fail in some cases does not mean that it fails in all, and it doesn’t mean that Weasel fails to demonstrate the power of cumulative selection.*
(the asterisk is a note that Joe F. pointed out the same thing) |
This kills the creature/population. /s
Rumraket introduces a relevant and related concept when he responds to a question from colewd (smoothness of landscape)
| Quote | | Obviously if you’re selecting towards the same phrase with a smooth hill to climb, the chance of convergence is one hundred percent. Interestingly, this is why convergence happens in nature. You’ll note how both fish and dolphins have that typical hydrodynamic torpedo-like shape with a pointy nose and long slim body. That’s an example of convergence and there’s a perfectly good, sensible explanation for it through natural selection. |
Mung perks up at the sight of something to fling a turd at and responds.
| Quote | Is that all that’s required to demonstrate the power of cumulative selection? A simple hill-climbing algorithm?
Wouldn’t it be amazing if nature consisted of smooth hills with targets at the top and a function to direct the genomes of living organisms to the tops of those hills? |
petrushka adds ...
| Quote | | What would be really amazing is if you could demonstrate that people like Larry Moran or Jerry Coyne or even Dawkins have ever suggested that nature has smooth, easy to climb hills. |
Mung flings again.
| Quote | We have the Dawkins Weasel program which is designed to produce the result it does. It would probably fail to find the target phrase in far more scenarios than those in which it successfully finds the target phrase. Assume for the sake of argument that cumulative selection is still in operation in those scenarios. How would we know?
Or should we just declare that it is so and go home. |
One wonders, did Mung just (fail to) relate and discover the reason for the extinction of 99+% of species that ever lived on the planet because they could not reach or maintain their position on the moving target of surviving in the environment? Who knew that the power of cumulative of selection or the combination of mutation + selection as related to a genomic pool has limits and is a cruel and harsh mistress?
I suppose if one is too intently focused on proving the map just doesn't work one might not realize it's connection, it's limits, and it's usefulness in understanding the real world.
(Apologies if I'm a bit off here or not explaining myself well. I have issues with that as it is and I'm a bit overtired. I'm just fascinated by watching people try as had as they can to go "nuh-uh" or intentionally try very hard to not understand. Go Tard.)
Edit: Yay! An edit button thingy! If it wasn't due to reaching 100 posts, my thanks to whoever flipped the bit.
--------------
"So I'm a pretty unusual guy and it's not stupidity that has gotten me where I am. It's brilliance."
"My brain is one of the very few independent thinking brains that you've ever met. And that's a thing of wonder to you and since you don't understand it you criticize it."
~Dave Hawkins~
|
