Fascinating. Thanks. As I understand it, information, mutation of information and phenotypic success combine to promote evolution irrespective of mechanism (as long as it's functional.) Would I be right?
you take biological progress as morphological variables, that's a fatal mistake, we are not made in that way, rather, our morphological information is encrypted in recombinable genes, many ppl is decieved by an evolutionary model that is the grandson of the long ago refuted lamarckism.
This only shows the lack of mechanisms for evolution that deals with the DNA and the many-mutations problem.
You either didn’t make any effort whatsoever to look into how the mathematics of Genetic Algorithms are structured, or you did so and are incapable of understanding basic mathematics. These algorithms work at the level of the Genotype to produce the resultant changes in the Phenotype (the morphology) exactly as you say is required.
Just simple clue for the mathematically incapable- they are called GENETIC algorithms because they are genetic in how they work
@HowieInTheUK who you trying to decieve? you can make genetic algorithms with any type of variable, did you really thought that they were coded tru layers of DNA-like encryption? that would be so laughable from you.
you guys still hold on to the old darwinian model, this is no different from the story of Lamarck's giraffes falsified by geneticists 100 years ago
DNA encodes genetic INFORMATION. The MECHANISM of encoding genetic information is irrelevant, it is the INFORMATION itself that counts. Genetic Algorithms encode and handle genetic INFORMATION in the same way as DNA. That is all that matters.
Your nonsensical argument is exactly akin to saying that since two chickens plus two chickens is four chickens ALL ADDITION must be carried out only using chickens
well... what's so special is that these programs can design things, things best adapted to solve the problems at hand. They design things by a blind and simple process, rather than the way WE do, by trying to deduce what design might be best for the situation. The designs of Evolutionary Algorithms are often suprisingly better than the ones that we deduced.
question - why don't you evolve a program to play chess.
We know that it's doable because IBM wrote deep blue that beat kasparov.
Your genetic code would be the primitives of the C programming language in which deep blue is written. (no cheating by programming in logic to the primitives!)
Marc, what you say is EXACTLY what engineers are doing to solve problems that are too complex for them to do with ordinary mathematics. They "breed" solutions and only let the fittest solutions survive to breed with each other again. The solutions also mutate. The selection criteria is what they want the design to do. This optimising technique works a treat. Airplane wings are often designed this way.
also, this has amazing potential for giving the computer the ability to perform creative problem solving. The computer could keep evolving a solution in its virtual world until one that could really work in the real world emerges. Two things would be needed, accurate simulation, and the ability to assess the value of a proto solution.
"evolutionary behaviours are forced on the simulation by the code that controls it"
The form & function of the creatures are determined by genes, the genes can mutate, and the creatures that do better at getting the food get to reproduce. Seems pretty straight forward to me.
..it is even MORE difficult to try to explain how such an "intellegent designer" could have ever come to exist before the universe did... or how he/she/it ever got so bloody intelligent in the first place to be able to design everything.... Only evolution explains how creatures become more complex and more fit for purpose from simple beginnings.....
if you look at the precision which all living things display in their working and even the universe as a whole displays marvelous precision in its working and the tolerances are so fine that it is very difficult to not believe that they were created by and intelligent being!!!!! almost all mutations of living things cause death and the ones that don't definatly do not enhance the workings of the creature they actually make the creature less fit for survival.
in this simulation the goals that are desired are made to influence the progress of the simulation, however there is no proof that the universe tries to favor survival rather than any thraits in organisms that cause death, in fact it is generally seen that the universe favors death more than life!!!! if a life form were to be evolved what is to stop it from being destroyed in the next instant and so not survive!!!!
Well, it certainly might be true that most mutations will be less fit to compete than their parent(s). If so, they are eliminated by the process of natural selection and this unfit mutation will not enter the overall gene pool. The rarer "improved" mutation will survive and multiply explosively.. as evolutionary mathematics shows. As improved mutations reproduce at a greater rate they are more likely to survive in the long term... Q.E.D.
youre not quite right in saying that this shows that evolution will hapen anywhere that reproduction occurs because in this environment evolutionary behaviours are forced on the simulation by the code that controls it, IE the creatures are forced to mutate the goals are forced to be obtained. this would not work in real life but it is a good way of achieving AIs in computers that can accomplish certain tasks!!
Those two grasping limbs send shivers down my spine. This is porn for a biologist.
5teampunker 3 weeks ago
Fascinating. Thanks. As I understand it, information, mutation of information and phenotypic success combine to promote evolution irrespective of mechanism (as long as it's functional.) Would I be right?
benbisley 7 months ago
@benbisley
Absolutely right Ben.... as long as the basic rule that "like begets like" applies. Darwin called it Descent with Modification.
Evolution is a process, an algorithm. As long as the ingredients of the algorithm are there ANYTHING will evolve
HowieInTheUK 7 months ago
you take biological progress as morphological variables, that's a fatal mistake, we are not made in that way, rather, our morphological information is encrypted in recombinable genes, many ppl is decieved by an evolutionary model that is the grandson of the long ago refuted lamarckism.
This only shows the lack of mechanisms for evolution that deals with the DNA and the many-mutations problem.
IloveYOUviruses 10 months ago
@IloveYOUviruse
Totally incorrect!
You either didn’t make any effort whatsoever to look into how the mathematics of Genetic Algorithms are structured, or you did so and are incapable of understanding basic mathematics. These algorithms work at the level of the Genotype to produce the resultant changes in the Phenotype (the morphology) exactly as you say is required.
Just simple clue for the mathematically incapable- they are called GENETIC algorithms because they are genetic in how they work
HowieInTheUK 10 months ago
@HowieInTheUK who you trying to decieve? you can make genetic algorithms with any type of variable, did you really thought that they were coded tru layers of DNA-like encryption? that would be so laughable from you.
you guys still hold on to the old darwinian model, this is no different from the story of Lamarck's giraffes falsified by geneticists 100 years ago
IloveYOUviruses 10 months ago
@IloveYOUviruses
Nonsense!
DNA encodes genetic INFORMATION. The MECHANISM of encoding genetic information is irrelevant, it is the INFORMATION itself that counts. Genetic Algorithms encode and handle genetic INFORMATION in the same way as DNA. That is all that matters.
Your nonsensical argument is exactly akin to saying that since two chickens plus two chickens is four chickens ALL ADDITION must be carried out only using chickens
HowieInTheUK 10 months ago
ok...the evolutionary processes are pre-programmed, to select from a wide variety of changes...so whats so special about this program?
ciphro 3 years ago
well... what's so special is that these programs can design things, things best adapted to solve the problems at hand. They design things by a blind and simple process, rather than the way WE do, by trying to deduce what design might be best for the situation. The designs of Evolutionary Algorithms are often suprisingly better than the ones that we deduced.
HowieInTheUK 3 years ago
question - why don't you evolve a program to play chess.
We know that it's doable because IBM wrote deep blue that beat kasparov.
Your genetic code would be the primitives of the C programming language in which deep blue is written. (no cheating by programming in logic to the primitives!)
unkie888 4 years ago
you would play your population against deep blue or some other chess program as the measure of fitness.
I suspect it won't be doable (without the cheating I have seen in many of the virtual evo. sims)
unkie888 4 years ago
If you put the following text into Google search you will get to downloadable video of Sims work (with commentary).......
sims_evolved_virtual_creatures_1994
Good luck
HowieInTheUK 4 years ago
Is this downloadeble from somewhere? PLease post me a link if you can on my email or tomy messages. Thanks
KronosDeret 4 years ago
Marc, what you say is EXACTLY what engineers are doing to solve problems that are too complex for them to do with ordinary mathematics. They "breed" solutions and only let the fittest solutions survive to breed with each other again. The solutions also mutate. The selection criteria is what they want the design to do. This optimising technique works a treat. Airplane wings are often designed this way.
HowieInTheUK 4 years ago
also, this has amazing potential for giving the computer the ability to perform creative problem solving. The computer could keep evolving a solution in its virtual world until one that could really work in the real world emerges. Two things would be needed, accurate simulation, and the ability to assess the value of a proto solution.
marcdesm 4 years ago
What does pmdvy mean by
"evolutionary behaviours are forced on the simulation by the code that controls it"
The form & function of the creatures are determined by genes, the genes can mutate, and the creatures that do better at getting the food get to reproduce. Seems pretty straight forward to me.
marcdesm 4 years ago
..it is even MORE difficult to try to explain how such an "intellegent designer" could have ever come to exist before the universe did... or how he/she/it ever got so bloody intelligent in the first place to be able to design everything.... Only evolution explains how creatures become more complex and more fit for purpose from simple beginnings.....
HowieInTheUK 4 years ago
Comment removed
robotroadkill 2 years ago
Comment removed
robotroadkill 2 years ago
if you look at the precision which all living things display in their working and even the universe as a whole displays marvelous precision in its working and the tolerances are so fine that it is very difficult to not believe that they were created by and intelligent being!!!!! almost all mutations of living things cause death and the ones that don't definatly do not enhance the workings of the creature they actually make the creature less fit for survival.
pmdvy 4 years ago
in this simulation the goals that are desired are made to influence the progress of the simulation, however there is no proof that the universe tries to favor survival rather than any thraits in organisms that cause death, in fact it is generally seen that the universe favors death more than life!!!! if a life form were to be evolved what is to stop it from being destroyed in the next instant and so not survive!!!!
pmdvy 4 years ago
Well, it certainly might be true that most mutations will be less fit to compete than their parent(s). If so, they are eliminated by the process of natural selection and this unfit mutation will not enter the overall gene pool. The rarer "improved" mutation will survive and multiply explosively.. as evolutionary mathematics shows. As improved mutations reproduce at a greater rate they are more likely to survive in the long term... Q.E.D.
HowieInTheUK 4 years ago
youre not quite right in saying that this shows that evolution will hapen anywhere that reproduction occurs because in this environment evolutionary behaviours are forced on the simulation by the code that controls it, IE the creatures are forced to mutate the goals are forced to be obtained. this would not work in real life but it is a good way of achieving AIs in computers that can accomplish certain tasks!!
pmdvy 4 years ago
If you keep your promise not to get too mathematical I'm up for it!
andwar99 4 years ago