BARBELITH underground
 

Subcultural engagement for the 21st Century...
Barbelith is a new kind of community (find out more)...
You can login or register.


Quantum Computers and AI

 
 
Mystery Gypt
00:16 / 06.02.02
so i read what i can here and there in Wired about quantum computing developments, the main thrust of which i suppose is an incredible increase in processing power because they can do simultaneous outcomes, right?

so is there work being done that would make quantum computing the key to creating real AIs? is there a necessary connection between the two? or are they totally different animals?
 
 
kid coagulant
12:55 / 06.02.02
I'm reading Ray Kurzweil's 'Age of Spiritual Machines' right now and I'm on the chapter where he's talking about the possibility of quantum computers and quantum mechanics in general (he says that quantum computers will be to digital computers what a hydrogen bomb is to a firecracker). It's whizzing right over my head, I'm afraid. Sounds like quantum copmuters and AI will could used together/will be one and the same. It's fascinating stuff (especially some of the encryption stuff and the processing speeds and putting computers into coffee cups). Here are some links from the book: http://qso.lanl.gov/qc/ http://www.research.ibm.com/quantuminfo/
 
 
kid coagulant
15:59 / 06.02.02
Kurzweil's website is also a pretty good resource: http://www.kurzweilai.net/
Lots of cool shit here.
 
 
captain piss
16:56 / 06.02.02
Still to look at the stuff in your links, invix, but- for a starter:
From what I know- quantum computing would supply the necessary leap in processing power that, some believe, might be essential for machines to become self-aware. However, the central issue in the feasibility of AI is really algorithms rather than hardware.

But putting aside the question of software and 'how' to approach the task of AI- there's the question of how much processing power or computer hardware is needed to approximate a human brain- I think its still a hotly debated question. I’ve just been reading an essay talking about a conference held in 1992 (called ‘How We Will Build a Machine that Thinks’ )* in which a minority believed the largest 1992 computers were within three orders of magnitude of the power of the human brain (the best computers in 2002 are roughly six times better than the best of 1992).

But, basically, I don’t think computing power is the biggest problem. The guys designing silicon chips today are having a hard enough job doing sensible things with the complexity we already have.
For the next few years, the challenge – in the grand project which will, depending on what you believe, lead to something like AI - seems to be that hardware complexity is running far ahead of our ability to do sensible things with it. So it’s possible to use the space on a silicon chip to do beefier signal processing (much faster and more complex graphics or sound rendering, for instance), but the effort of designing more complex and ‘intelligent’ systems on a chip is currently floundering- there basically just isn’t design software sophisticated enough to let the techy guys exploit the complexity that’s possible with a silicon chip.
Which is kind of a roundabout way of saying that the problem of computing power- which quantum computing might well solve- is not the main one facing those trying to make AI happen.

* the essay is at http://www.ugcs.caltech.edu/~phoenix/vinge/vinge-sing.html

[ 06-02-2002: Message edited by: Meme Buggerer ]

[ 06-02-2002: Message edited by: Meme Buggerer ]
 
 
kid coagulant
17:33 / 06.02.02
The thing I can't figure out is why are we basing artificial intelligence on the human brain? I mean, we don't even understand the human brain yet, why can't we come up w/ something else for it? Having a hard time getting my head around this.
Here's more AI stuff, from slashdot.com:
'Arguing A.I.: The Battle for Twenty-first Century Science' http://slashdot.org/books/02/02/04/2248211.shtml
 
 
The Sinister Haiku Bureau
18:31 / 06.02.02
I've metioned this/these guy(s) before, but it's worth mentioning again. http://www.sysopmind.com/beyond.html This guy claims the figure 10^17 floating point operations a second ins the approx computing power of the human mind. He also outlines how to build an AI, and bring about the singularity, using a language he's designing called Flare, along with all sorts of other stuff. It's pretty fascinating stuff.
And yeah, I've only read one of Kurzweil's essays, but he's hella cool.
 
 
Mystery Gypt
20:42 / 06.02.02
quote:Originally posted by invix:
why are we basing artificial intelligence on the human brain?


because that is the only thing most scientists can agree is "intelligent."

[ 06-02-2002: Message edited by: Mystery Gypt ]
 
 
Magic Mutley
18:40 / 07.02.02
I'd go further - I'd say we've created the word 'intelligence' to refer to processing information in a human-like manner (Hmmmm...I've always wondered why we got to be the most intelligent species on the planet )

A quantum computer would process information very differently to a normal computer, and the algorithms would be different - the ability to handle huge numbers of calculations in parallel is closer to how the human brain is structured. Dunno, but it's certainly going to be interesting when they get one going...
 
  
Add Your Reply