Natural artificial intelligence...

Ye' old general discussion board. Basically, for everything that isn't covered elsewhere. Come here to shoot the breeze, shoot your mouth off, or whatever suits your fancy.
This forum is not for asking programming related questions.

Moderator: General Moderators

alex.barylski
DevNet Evangelist
Posts: 6267
Joined: Tue Dec 21, 2004 5:00 pm
Location: Winnipeg

Natural artificial intelligence...

Post by alex.barylski »

Oxymoron? You bet!

Just finished watching some documentry on dinosuars/evolution, etc...always been interested in that stuff...and because I love what we do as well...I often try and find similarities between them...personally I see tonnes.

Got me thinking...if the genome is anaologous to the machine code of a software program (in concept)...then if given enough time and processing power...I wonder what would happen if you randomly (using real variables pulled from the internet) let a program generate binary files and executed them as Windows binaries or something similar...

With the genuine variability of input from Internet sources (stocks, weather, etc) and say 10000 years...I wonder if the output would ever result in something useful, like a calculator or operating system or perhaps an intelligent, natural (due to osmosis not direct human intervention), but non-organic life form...

If you created a binary executable the same size, as say Windows kernel...and just randomly over a period of 10,000 years tweaked a byte at a time and ran the code...eventually...in theory...you would reproduce the exact results and/or possibly end up with something better...

That is essentially (as my understanding goes) how evolution works and has worked...atomic mutations over billions of years resulted in life as we know it...creepy....it's like living in the matrix...

Cheers :)
User avatar
Benjamin
Site Administrator
Posts: 6935
Joined: Sun May 19, 2002 10:24 pm

Re: Natural artificial intelligence...

Post by Benjamin »

Dr. Thompson peered inside his perfect offspring to gain insight into its methods, but what he found inside was baffling. The plucky chip was utilizing only thirty-seven of its one hundred logic gates, and most of them were arranged in a curious collection of feedback loops. Five individual logic cells were functionally disconnected from the rest–
http://www.damninteresting.com/?p=870
User avatar
Inkyskin
Forum Contributor
Posts: 282
Joined: Mon Nov 19, 2007 10:15 am
Location: UK

Re: Natural artificial intelligence...

Post by Inkyskin »

Along the same lines - In a sense the internet is like a giant brain, with each node (server) like a memory cell. All that's missing is some way to access itself automatically - although thats what DNS does for us. The PC end user would essentially be it's face. Just it talks by displaying a webpage instead of with a mouth. The only thing thats missing is the ability to make all the decisions it wants by itself...

Hang on, isnt that the plot with Skynet in Terminator... All it takes is one stupid scientist or evil genius and were screwed!
malcolmboston
DevNet Resident
Posts: 1826
Joined: Tue Nov 18, 2003 1:09 pm
Location: Middlesbrough, UK

Re: Natural artificial intelligence...

Post by malcolmboston »

Since the matrix ive contemplated that in the future something similar could happen.

Not being taken over by machines, but an actual world created by programming, always has fascinated me.....
User avatar
Chris Corbyn
Breakbeat Nuttzer
Posts: 13098
Joined: Wed Mar 24, 2004 7:57 am
Location: Melbourne, Australia

Re: Natural artificial intelligence...

Post by Chris Corbyn »

Reminds me of i-Robot.
User avatar
Jonah Bron
DevNet Master
Posts: 2764
Joined: Thu Mar 15, 2007 6:28 pm
Location: Redding, California

Re: Natural artificial intelligence...

Post by Jonah Bron »

Sure. You really believe you are an accident? If evoluton is true, then I want to know where the "gases" came from that caused the "Big Bang"?

I don't get it. 8)
User avatar
Benjamin
Site Administrator
Posts: 6935
Joined: Sun May 19, 2002 10:24 pm

Re: Natural artificial intelligence...

Post by Benjamin »

PHPyoungster wrote:Sure. You really believe you are an accident? If evoluton is true, then I want to know where the "gases" came from that caused the "Big Bang"?

I don't get it. 8)
Let's not go there.
User avatar
Christopher
Site Administrator
Posts: 13596
Joined: Wed Aug 25, 2004 7:54 pm
Location: New York, NY, US

Re: Natural artificial intelligence...

Post by Christopher »

Hockey wrote:Just finished watching some documentry on dinosuars/evolution, etc...always been interested in that stuff...
Me too... :)
Hockey wrote:Got me thinking...if the genome is anaologous to the machine code of a software program (in concept)...
I think this is where you went off the rails ... ;)
Hockey wrote:That is essentially (as my understanding goes) how evolution works and has worked...atomic mutations over billions of years resulted in life as we know it...creepy....it's like living in the matrix...
That is actually a bit of a misunderstanding of how evolution works. It is in many ways a very un-random process. Given the complex set of conditions in an environment and that each time living things recreate themselves the offspring are slightly different than the parent -- given that -- some percentage of those living things will reproduce at a slightly higher rate than the the rest. Because of that, over a very long time certain combinations will increase in numbers and others will decrease. And it just keeps going. It is in many ways very determinate. Add on top of that largish changes in the environment, also fairly determinate given orbits, solar cycles, etc. and here we are...

That is actually the way genetic algorithms work. But there has to be a clear set of rules to be the environment for them to operate in. The rules are often what you change to improve the results. It is also interesting that when they analyze the results of running genetic algorithms, they are often apparently inefficient and wasteful, and a bit odd sometime bordering on not clearly comprehensible. If you have ever run them you might know what I mean.
(#10850)
User avatar
Benjamin
Site Administrator
Posts: 6935
Joined: Sun May 19, 2002 10:24 pm

Re: Natural artificial intelligence...

Post by Benjamin »

It's funny that Hockey started this post. I had been thinking about writing some code that would "evolve" to perform a certain task. I just need a good starting point. A foot in the door if you will, as to where to start. Maybe someday...
alex.barylski
DevNet Evangelist
Posts: 6267
Joined: Tue Dec 21, 2004 5:00 pm
Location: Winnipeg

Re: Natural artificial intelligence...

Post by alex.barylski »

That is actually a bit of a misunderstanding of how evolution works. It is in many ways a very un-random process
Depends on how you define random I think. ;)

Evolution, is personally...more unpredictable than where I am getting my next lunch...my personal patterns would indicate that I typically eat at Subway...what my first born will look like...I couldn't tell you. :P

Weather is another example, most would say it is unpredictable...especially Pilots...and yet there are certainly patterns to be observed and studied in the circle of meteorologists. In the Newtonian context...everything in theory should be predictable...but Newtonian laws don't apply at the atomic level...

Evolution is IMHO a random event which occurs over millions and billions of years...

I think to say that evolution is *not* random is incorrect. Rather evolution is simply influenced by many variables, some of which are more predictable than others. You can calculate where the Earth will be relative to the Sun, but you can't calculate whether your species is wiped out by another.
astions wrote:It's funny that Hockey started this post. I had been thinking about writing some code that would "evolve" to perform a certain task. I just need a good starting point. A foot in the door if you will, as to where to start. Maybe someday...
I think the answer is in simplicity, after all those are always the best solutions. I think humans trying to copy life would require the understanding of an infinite number of variables, thus making emulation impossible, but simulation perhaps more practical.

Look at games like SimCity or whatever. They simulate life, under a pre-determined number of variables...but I don't think all the variables will ever be understood until we as a species can comprehend infinite as a quantity.

I say, life and evolution is a random event...the more arbitrary you can make the input, the better odds you have of achvieving success. I think life is truely a simple matter of a freak event occuring billions of years ago and from that point on it's been a matter of trial and error. The best to copy that is to do something like I suggest in my original post.

Using man-made languages like Lisp, etc...are excellent for education and experimentation...but not for emulation.
User avatar
Christopher
Site Administrator
Posts: 13596
Joined: Wed Aug 25, 2004 7:54 pm
Location: New York, NY, US

Re: Natural artificial intelligence...

Post by Christopher »

Hockey wrote:Depends on how you define random I think. ;)
I don't think we get to pick and choose the definitions of random and prediction and probability -- they have specific scientific meanings distinct from their common, non-scientific usage.
Hockey wrote:Evolution, is personally...more unpredictable than where I am getting my next lunch...my personal patterns would indicate that I typically eat at Subway...what my first born will look like...I couldn't tell you. :P
Give a biologist your genome and they could make a fairly accurate prediction, but if they didn't know you they would have to guess at where you would have lunch. Prediction has to do with the knowledge and information available.
Hockey wrote:Weather is another example, most would say it is unpredictable...especially Pilots...and yet there are certainly patterns to be observed and studied in the circle of meteorologists. In the Newtonian context...everything in theory should be predictable...but Newtonian laws don't apply at the atomic level...
Again, weather is increasingly more accurately predicted as the knowledge and data improves. The current state of weather prediction allows for greatly improved flight planning than years ago.

As far as Newton's equations, they certainly can be applied at the atomic level -- they just are not as accurate as other equations. Let's not mistake the thing itself for the math that tries to model it.
Hockey wrote:Evolution is IMHO a random event which occurs over millions and billions of years...
Evolution is not "a random event." Billions and trillions of events occur over millions of years. Difficult to predict due to the complexity of the system, but not impossible to comprehend. Archaeologists make predictions that are proven out in the fossil record. Biologists can do repeatable experiments on rapidly reproducing species the produce predictable evolutionary results.
Hockey wrote:I think to say that evolution is *not* random is incorrect. Rather evolution is simply influenced by many variables, some of which are more predictable than others.
Again, I think you are mistaking a process with random or difficult to predict events in it, with a random event. On one level your daily walk to Subway is completely random. You never step in the same places or follow the same path. On the atomic level it is a complex system that is difficult to predict. But on a more pertinent level, your walk to Subway every day is utterly predictable.
Hockey wrote:You can calculate where the Earth will be relative to the Sun, but you can't calculate whether your species is wiped out by another.
Biologists actually keep lists of the species that are being wiped out by other species, and how long they have, and what can be done to alter that course. I think people are not aware of the advanced state of biology today. Certainly modern medicine would not be possible without the very detailed understanding of evolution that biologist currently have.
(#10850)
alex.barylski
DevNet Evangelist
Posts: 6267
Joined: Tue Dec 21, 2004 5:00 pm
Location: Winnipeg

Re: Natural artificial intelligence...

Post by alex.barylski »

Evolution is not "a random event." Billions and trillions of events occur over millions of years. Difficult to predict due to the complexity of the system,
Trillions and billions of events...exactly...essentially you cannot say for sure...so we must assume infinite like the stars in the sky. I think infinite has more in common with random than you care to accept. ;)

--- Back on topic ---

Evolution is partially influenced by the chaos theory or the butterfly effect or whatever...completely random choices we make influence our very own evolution. Had I not taken that right step in my high school hallway and ended up in the bathroom, I may have been shot and my existance would have ended - no more evolution on a very small scale.

It's clear to me we simply see different sides of the coin. Natural selection, is what I think you keep refering to, as it is the non-random, deliberate result of generations of evolutionary progress. Only the strong survive.

What I believe, is that (unless you believe in a higher power or divinity) the only way for change to occur in simple life forms is for something to randomly change - what triggers that event is a infinite number of variables (weather, food sources, prey, etc).

Both the requirements of randomness for adaptation to occur and the literally billions of variables which factor into the equation, IMHO make evolution a very random event.

When a species is confronted with adversity, it it given two choices: Adapt or Die. Thousands or maybe millions of species have gone extinct since the start of time. If evolution was so predictable, why didn't dinosaurs figure out how to adapt? The next step should be obvious - but because evolution is random and those adaptive changes didn't occur quick enough they died!

Evolution, I choose to believe, is a series of random changes, just like how I explain in my first post. A little genetic tweak when two different members of that species procreate and bam...you have a derivative of the parents. If those changes weren't random...why would we have mutations which cause diseases, like cancer? If it's so predictable, why would life choose bad genes?

For that reason and others, I believe that those tweaks in genes are random, at least under the context of which genes are inherited from mother and father and what positive or negative mutations occur during development in the womb. It's to my understanding, that your genes can actually mutate or adapt during your lifetime, which might be influenced by environment and other factors...in these instances...adaptation is perhaps a deliberate act.

However in saying that, I don't believe the intelligence exists at the celluar level...at least to the point required to make smart decisions on how to mutate. When an organism which is water born is suddently faced with shortage of water, the cells do not intelligently decide to change gene "A" because they know that is the gene responsible for them being water bound. Rather, the next generation will receive some form of celluar memory from it's ancestors which indicate that a change is in order. The cell mutates and begins the stage towards sure death or adaptation. Perhaps the order in which genes are changed is known (first try turning on gene A, then gene B, C, etc) but the results are clearly unknown. The idea is over hundreds of generations, the cells change random genes with each iteration. Those changes will dictate whether their kind continue or seize to exist. Again, I feel it is this random-ness which is responsible for mutations which cause disease, like cancer. It is because the cells have no way in knowing the direct results of genetic mutations which occur over the generations that result in failure or success.

I personally feel evolution is better described as a "trial and error" process...rather than "predetermine and execute" the latter of which should result in perfection...otherwise...it's trial and error. If you cannot nail an answer all the time, like you can in mathematics, is it because at least one variable is not understood...so you are going into the unknown...and I am willing to bet most would agree that the unknown is more analogous to random, than it is to ascertainable. ;)

p.s-If it's one thing I appreciate about your replies...it is your Devils Advocate perspective. I may not always agree with you...of course that doesn't mean I think your wrong either...but being challenged to see things from another angle is always interesting and sometimes enlightening.

Cheers :)
User avatar
Benjamin
Site Administrator
Posts: 6935
Joined: Sun May 19, 2002 10:24 pm

Re: Natural artificial intelligence...

Post by Benjamin »

Hockey wrote:I don't believe the intelligence exists at the celluar level...
Intelligence IS the cellular level. That's you.

An analogy?

Transistors. Out of the many in a CPU, which one runs the PC?

And yes, that's exactly what I mean.
alex.barylski
DevNet Evangelist
Posts: 6267
Joined: Tue Dec 21, 2004 5:00 pm
Location: Winnipeg

Re: Natural artificial intelligence...

Post by alex.barylski »

astions wrote:
Hockey wrote:I don't believe the intelligence exists at the celluar level...
Intelligence IS the cellular level. That's you.

An analogy?

Transistors. Out of the many in a CPU, which one runs the PC?

And yes, that's exactly what I mean.
Yes and no...

Your intelligence is dictated by a neural network of billions of cells...not just one. As a simple cell, I personally do not believe it is capable of making such high level "intelligent" choices. Like, gene #35464 should be turned on because we need to adapt to drier than normal land.

It's more like, something is wrong/different here...lets change something and see what happens (cross fingers and wait and see).

Just like a transistor is little more than a simple atomic unit of computational power...it's in combination with millions of others that gives a computer it's amazing abilities.

If you found yourself surrounded by water and thought as hard as you could...you would not likely develop gills and suddently breathe under water (trust me - I've tried). But if you had children in water and they did the same and so on...over a couple thousand years...your genes might change enough through random intervention and possibly some deliberate changes you may very well develop gills and flippers and fins, etc...
User avatar
Benjamin
Site Administrator
Posts: 6935
Joined: Sun May 19, 2002 10:24 pm

Re: Natural artificial intelligence...

Post by Benjamin »

Intelligence and Evolution are two completely different things. Your meshing them together. Things are not near as complicated as they seem sometimes.
Post Reply