AI-artificial intelligence

Human Anatomy, Physiology, and Medicine. Anything human!

Moderators: honeev, Leonid, amiradm, BioTeam

Post Reply
User avatar
kabuto
Coral
Coral
Posts: 118
Joined: Sun Apr 02, 2006 7:30 am

AI-artificial intelligence

Post by kabuto » Fri Apr 21, 2006 9:16 am

do u think that one day artificial intelligence can surpass our own?
:?: :?:

pdavis68
Death Adder
Death Adder
Posts: 57
Joined: Mon Apr 03, 2006 8:37 pm

Post by pdavis68 » Fri Apr 21, 2006 5:22 pm

As a software engineer, I'll chime in on this one: Absolutely. Not a doubt in my mind.

It's largely an issue of computational power. There have already been developments in software that "evolves" on its own. Basically it uses ideas similar to DNA evolution (crossovers, mutations, etc) to create programs that are more efficient. Some of the programs don't work and die off, other's spawn. This, combined with advances in AI techniques (Neural Nets, Genetic Algorithms, etc) will, I believe, eventually lead to an intelligent system.

I suspect it will work as follows: A basic system will be designed, modeled on the human mind. It will be given "sensory input", methods of communication, and a basic neural structure (virtual, in the program, not a real neural structure). A biologic-style evolution system will cause it to adapt and evolve. Once the system begins, it will adapt and evolve until it starts to show signs of intelligence. The limiting factor will largely be computational power and memory. I suspect it's "doable" now, to some degree, but the computational speed would make it take hundreds of thousands of years to evolve intelligence.

It probably won't happen in my lifetime, but I have little doubt it WILL happen.

User avatar
AstusAleator
King Cobra
King Cobra
Posts: 1039
Joined: Tue Jan 24, 2006 8:51 pm
Location: Oregon, USA

Post by AstusAleator » Fri Apr 21, 2006 9:19 pm

could you please post a link to the program that is evolving ai? This is a particular fascination of mine. In fact i wanted to be the first one to do it, and I'm pissed it's already being done :D. But then again i'm not a software engineer so... go figure.

pdavis68
Death Adder
Death Adder
Posts: 57
Joined: Mon Apr 03, 2006 8:37 pm

Post by pdavis68 » Fri Apr 21, 2006 9:48 pm

Sure, the program is called Tierra and you can get it here:
http://www.nis.atr.jp/~ray/tierra/

There's also Avida which I believe is derived from Tierra:
http://devolab.cse.msu.edu/software/

Understand, this is really just a proof of concept kind of thing and it's not terribly advanced. The computer programs are written in a psuedo-assembly code and it runs on "virtual" cpus.

That said, I haven't really looked at this stuff since 1999, so they may have improved on it quite a bit. Still, it's definitely in its infancy. It's going to take years to apply this to real regular software, let alone neural nets and genetic algorithms.

But this is how these things go. A little step here, a little step there, and eventually you've got something useful.

User avatar
kabuto
Coral
Coral
Posts: 118
Joined: Sun Apr 02, 2006 7:30 am

Post by kabuto » Sun Apr 23, 2006 10:08 am

if they eventually surpass us, what would happen?
is there any chance that they could turn against us (like Terminator, Matrix, etc)? :? :?

pdavis68
Death Adder
Death Adder
Posts: 57
Joined: Mon Apr 03, 2006 8:37 pm

Post by pdavis68 » Sun Apr 23, 2006 1:42 pm

Well, that really depends. Could they turn against us? Sure. But what could they do?

I mean, the first "intelligent" computers are unlikely to be robots (since I suspect the first ones will be rooms full of computers) and they're unlikely to be connected to any manufacturing equipment. If they were connected to the internet, they could probably figure out how to hack into systems and wreak havoc, but there are certain obvious protections you can put in place to prevent that (firewalls, traffic monitoring and restriction, etc).

So, even if they decided we were evil and to do away with us, they wouldn't really have the means. Unlike the terminator movies, there's simply nobody in government stupid enough to hand over a nuclear arsenal to a computer and I don't think that day will ever come.

Also, there's no guarantee that with intelligence will come any sort of emotions. The jury is still out on how much emotions are tied to biochemistry. It could just be a really brilliant machine with no emotions, desires, love, or hate. Things like this are kind of hard for us to grasp because it's outside of our experience. As far as we know, all higher animal forms have some degree of emotions. But it's certainly possible and I would suspect even likely, that an intelligent computer would simply be emotionless.

An interestiing point is that, because their evolution will likely happen quickly, they will probably reach our level of intelligence and pass by it swiftly. Hopefully they would make groundbreaking discoveries for us in a variety of areas. The would, I hope solve problems that we have as yet been unable to solve. Come up with cures for diseases, solve physics problems that we haven't, and so forth.

But no, I don't think they're going to take over the world.

The mention of the matrix is interesting. Years ago I read a philosophical argument that basically stated, if a system like the matrix (really, his argument was more along the lines of the story from the movie The Thirteenth Floor) is possible, where virtual life is created inside of computers, then the probability is exceedingly high that we live in such a world. In other words, if it's even possible at all, then the chances are, we're already living inside one created by someone else.

The argument was very compelling and even described how it would work. I wish I knew where I read it, because I'd love to read it again. The conventional argument is that you'd need a computer that could simulate the entire universe, but his argument was that you didn't need that. All you needed was a computer that could simulate what each individual would experience. If one looked in a telescope, then it would simulate what the telescope would see while that person was looking in it. If you looked in a microscope, it would simulate what was there while you could see it. But the rest of the times, it didn't deal with the microscopic or cosmic level details.

Anyway, it was a really interesting piece. Really made you think.

Post Reply

Who is online

Users browsing this forum: No registered users and 6 guests