What is a GA?
Discuss.
704 thoughts on “Holding tank for general chatter about GAs”
Leave a Reply
You must be logged in to post a comment.
The Skeptical Zone
"I beseech you, in the bowels of Christ, think it possible that you may be mistaken."
What is a GA?
Discuss.
You must be logged in to post a comment.
Well, neither can Avida organisms. That refutes your own argument. 🙂
Seriously, both cases represent Darwinian evolution. There is random variation and differential reproduction. These are the two main ingredients of Darwinian evolution. Both numerical experiments with Avida and in vitro experiments with RNA aptamers have that.
Because I don’t click on every link in every post?
But certainly my bad. I’m glad to have read it now.
Well you can’t reproduce on your own either.
Hi Elizabeth,
I see you accused me of not understanding GAs however my scenario is directly derived from “Evolving Inventions” which uses genetic programming, which is a specialization of genetic algorithms, according to wikipedia (and others).
So perhaps you would care to support your accusation?
Joe G,
You seemed to think that GAs could be “embedded” in the genomes of organisms, and respond to environmental cues, as in epigenetics. It’s a cool idea, but it’s not the principle of GAs, and not the principle of AVIDA (which is relevant to this thread). It’s also not the principle of the GAs in your link! That’s why I think you are confused. But further discussion will have to wait, unless you want to start a thread in the Penguin Colony.
Umm the principle of GAs is to solve a problem. And the option of gene duplications followed by mutation changing function “just happening” is crazy.
Joe G,
There are over 16,000 peer reviewed papers on PubMed documenting exactly that “craziness”.
Really? How was it determined that gene duplication followed by mutation changing function is a blind and undirected process?
Please show the work.
Joe. Do you believe, or have any reason to believe, that mutations resulting in new, beneficial, functions or structures, are always or mostly in some way directed?
A simple yes or no is all that’s needed, and whichever you say, I’m not going to argue it tonight.
It’s important. I’ve written down what I think your answer will be. If my prediction is correct, I get a good shot of Tobermory 15-year-old single malt before beddibyes. If not, it’s tea.
Once again-
I am OK with the following:
http://en.wikipedia.org/wiki/Genetic_programming
and
http://en.wikipedia.org/wiki/Genetic_algorithms
and
http://en.wikipedia.org/wiki/Evolutionary_algorithm
GAs are used to solve problems. In my scenario, in in the SciAm article “Evolving Inventions” tat is exactly what is taking place.
My scenario has a GA directing mutations to solve a problem, for example a new protein or protein complex.
Point mutations, probably not directed-
recombinations, gene duplications, transposons, insertions, deletions-> probably directed.
No, the principle of a GA is to evolve a population of virtual organisms that thrive within a virtual environment that presents them with hazards and opportunities.
Sometimes, the purpose of a GA is to solve a problem, because if you present your problem as the hazards and opportunities of a virtual environment, as the virtual population of organisms evolves to thrive within that environment, it will solve your problem for you.
But a GA is not embedded within the organisms, nor, in a GA, do we normally make the genomes react to their environment by changing their sequences. It’s the other way round – the Darwinian way round. Genomes are mutated/recombined randomly with respect to the environment, and those that are best able to thrive within it are most likely to contribute parts of their genomes to the next generation. So your description was of something that is not, normally, referred to as a GA. The link you gave, however, was.
It “just happens” in the way that hydrogen “just happens” to combine with oxygen to form water if you burn it in air.
That’s not crazy. It’s chemistry.
Well, that’s just not what a GA, is, Joe. In a normal GA, the only “direction” is provided at the level of the fitness function, not at the level of directing the mutations.
As I said, it’s interesting, but the actual opposite of a GA. If we knew what mutations were required to solve a problem, we wouldn’t need the GA!
They’re not undirected. My dentist magically causes all gene duplications, with his midichlorians.
Prove me wrong.
Probably directed? To what end?
Science: You’re doing it backwards.
There is no empirical evidence that suggests that particular mutations are induced by any external entity. The rate of mutation can certainly be influenced, but not the specific changes that occur.
If you have any objective, empirical evidence that suggests otherwise, please present it. Otherwise, your speculation is no different from the idea that tiny, undetectable gremlins are weaving DNA for their own unknowable purposes while carefully making their changes look consistent with known natural processes.
We could step through a GA one line at a time to help Joe get it?
Sure, here’s a fun one.
For extra credit, I’d love to see someone calculate the CSI of the string representing the Steiner Problem solution.
Spoiler Alert: it won’t be Joe.
This one is quite cool, as it does something the GA designer doesn’t expect! It solves the problem (of locomotion) by exploiting a loophole.
Countdown to :”Yes, but the program running it is designed!!”
And this is nice too:
Well, the important point for now is that what is designed is the starting population of self-replicators and the fitness criteria. Not the genomes, as witnessed by the fact that the phenotypes are so varied and surprising.
All computing is evolutionary. It all involves the rule-bound transformation of an input into an output. That’s at least something like what we mean by “evolution” isn’t it? The transformation of something into something else via a process that we can understand, rationalize, and most importantly, not respective of the former, reproduce (“rule-bound”).
There isn’t any evidence that all mutations are via blind and undirected chemical processes.
The only reason to say that all mutations are do to blind and undirected chemical processes is if living organisms are reducible to matter, energy, necessity and chance.
That is why ID is all about origins and my scenario is how a designer could design something that would be “hands free”, ie no intervention required.
Joe G,
You’ve miscomprehended and been helped in this thread already, Joe. I hope to learn more also. Let go of your hate, it’s the path to the dark side! 😉
It still solves the problem it was designed to solve. And how it did it is the whole point of GAs-> the exploit things we miss.
Now I know why you guys don’t understand what i am saying…
Leaving aside the problematic word “reducible” (epi- and emergent phenomena make reductionism untenable) for the nonce, please provide any objective, empirical evidence you have for anything other than physics and chemistry being used by living organisms. Where, exactly, is this élan vital you seem to be asserting?
Perhaps that is what you think- however it is strange that you can’t support it.
Where? In the organisms- just as the information- in the organisms-
You have no idea what IDists have been saying, do you?
Umm, we are talking about GAs- you do understand GAs are written to solve problems- or maybe you don’t.
They would be directed to the end of the GA running at the time.
In the same way as bananas were designed for our enjoyment, I’m sure.
Elizabeth- a GA can be written to solve any problem- YOU don’t get to tell people what their GAs must be like.
2- You miss the point- ID says there is MORE than just physics and chemsitry at work- there is also information
3- Yeah making water means we can make a living organism by mixing chemicals together
4- The internal GA/ GP is how the organisms evolve by design/ are designed to evolve.
5- So just demonstrate your claims- that living organisms are reducible to physics and chemistry and ID would be falsified and my internal GA/ GP will be neatly done away with
I eagerly await the arrival of your paper.
There is a way to demonstrate you understand how to write a program, including a GA program. It is the way all of us who have written complex programs, such as a GA, do it.
All you have to do to show us how a GA program operates is to lay out the algorithm or algorithms by which the program works. Do it the way those of us who have done this sort of thing do it in our work; write out a step-by-step algorithm in words, numbering the steps as you go.
Can you do that for us?
Elizabeth- obviously you are not following along- you don’t need to know the mutations, just what you need- what the protein needed is. The GA solves the problem of the missing protein by directing its manufacture via, guess what, mutations to the DNA to produce the required mRNA- whatever those mutations are.
Mike, can you address my arguement or not- can you do that for me, for once?
You aren’t making an argument; you are demonstrating to us that you don’t know what a GA is.
Bottom line Mike, I am sick of your false accusations. Especially when it is obvious that you have no idea what I am claiming.
I’m sorry, I can’t parse this. Could you please explain exactly what you are claiming is present in living organisms in addition to physical and chemical processes?
I’m trying to understand what you are saying. Please help me by stating your claims clearly and unambiguously.
I’m sorry, I can’t parse this.Could you please explain exactly what you are claiming is present in living organisms in addition to physical and chemical processes?
I’m trying to understand what you are saying.Please help me by stating your claims clearly and unambiguously.
ID claims that living organisms are not reducible to matter, energy, necessity and chance- ie that there is more to be a living organism that physics and chemistry.
ID claims that there is, for one, also information. Just as information runs a computer, information runs organisms.
It’s safe to say that most of us are having trouble understanding what you are saying, Joe. I suspect that this is not because we are stupid but because your writing is careless. It doesn’t take much observation to see that it takes you less than a minute to read a comment an write a response. You don’t even take time to digest what your opponent is saying, let alone read some source he or she cited (and ponder it).
If you don’t make an effort at learning, you won’t learn.
What do you mean by “information runs a computer”? Computers run according to known laws of physics.
What do you mean by “information runs a computer”?Computers run according to known laws of physics.
That’s a laugh- I would love to see any modern computer run without a program, ie information.
Go tell that to computer programmers….
Computer programs are compiled into machine language instructions and then typically loaded into RAM (or pre-loaded into ROM), after which they are accessed by a CPU that follows those instructions by raising and lowering voltages on the integrated circuits within the CPU and external components. All of that is based purely on known laws of physics. There is no free-floating “information” independent of the physical components.
We can certainly choose to model the patterns of the state of the RAM chips using information theory, but the computer as a system is operating based on its configuration and well understood electronic and mechanical principles.
So, getting back to the original question, please provide any objective, empirical evidence you have for anything other than physics and chemistry being used by living organisms (or computers, for that matter). Where, exactly, is this élan vital you seem to be asserting?
Amazing. Joe G still can’t grasp the simple fact that a GA can’t run INSIDE the organism that is evolving. The GA needs to be EXTERNAL where it can affect the selection pressures on the organism and thus impact the direction their evolution takes.
fantastic:
[embedded by Lizzie]
Joe, walk us throught how your “cell internal GAs” handle speciation events.
For example, the genetic evidence shows the last common ancestor of the felidae and canidae lived around 52 million years ago. The GAs in that population must have had all the ‘targets’ stored for all our modern felids and canids. Then, when that population split into the two different lineages, somehow the GAs knew to only send the ‘targets’ for modern cats into one lineage, and only send the targets’ for modern canids into the other. The felidae lineage divided, and divided, and divided again until we got lions, tigers, panthers, pumas, saber toothed cats, domestic moggies, etc. The canidae lineage divided, and divided, and divided again into foxes, and dingos, and wolves, and hyenas, etc.
Each step of the way, the GAs somehow knew to send only the ‘targets’ of that particular lineage down its specific line.
How did they manage to do that Joe?
Don’t forget that ID isn’t anti-evolution, right?
Joe, presumably you aren’t saying that computers disobey the laws of physics?
That something other than electromotive force pushes the electrons around?
This is the problem, it seems to me with regarding “reductionism” as “reducing” everything to “mere” matter and energy. A computer is clearly made up of matter and energy and nothing else. People don’t tend to regard computers as having an immortal soul (although perhaps you do).
Yet despite being made of nothing but matter and energy, they can do really interesting things – perform interesting functions. They do this by virtue of being part of a system.
Why can’t the same be true of living things?