RE: [asa] NASA - Climate Simulation Computer Becomes More Powerful

Long ago, Bernie Alder, who did very pioneering work in computer simulations of hard disks and spheres, referred to computers as fast idiots. I have yet to hear a more accurate description than that.
Moorad

From: asa-owner@lists.calvin.edu [mailto:asa-owner@lists.calvin.edu] On Behalf Of Rich Blinne
Sent: Thursday, September 10, 2009 9:03 PM
To: Randy Isaac
Cc: asa@calvin.edu
Subject: Re: [asa] NASA - Climate Simulation Computer Becomes More Powerful

On Thu, Sep 10, 2009 at 5:58 PM, Randy Isaac <randyisaac@comcast.net<mailto:randyisaac@comcast.net>> wrote:

The complexity of Deep Blue was greater than its creators in the sense of being able to calculate an enormous number of possible scenarios and calculate all their weighting values. No human can do that. But humans have a much better intuitive sense of figuring it out without such brute force calculations. That's a complexity no computer can touch.

I've witnessed several cycles of "great disappointments" concerning artificial intelligence, particularly AI that had little or no constraints. People tried to give a program some axioms and see if it could reconstruct Euclid to no avail. On the other hand if an inference engine is more tightly constrained like some medical diagnosis programs more progress was made (this is not unlike the Big Blue example above). Everything was in the extrinsic weighting values. The inference engine, genetic algorithm, or game tree search itself is pretty stupid and contained little information. Likewise in Dembsky's latest IEEE paper active information comes extrinsically from the environment and not intrinsically from the genetic code. Thus, no matter how "powerful" your computer is you cannot achieve true intelligence from it other than the faux one like ELIZA playing the Rogerian psychotherapist. Once the author of ELIZA, Joseph Weizenbaum, saw that he left the field in disgust. I do find hooking PARRY up to ELIZA amusing, though. :-)

Rich Blinne
Member ASA

To unsubscribe, send a message to majordomo@calvin.edu with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Fri Sep 11 09:57:37 2009