I presume that almost everyone knows that today, January, 12, 1997, is the
actual birthday of HAL, the sentient computer in "2001: A Space Odyssey", the
movie in which he spoke the dying words:
"I am a HAL Nine Thousand computer Production Number 3.
I became operational at the HAL Plant in Urbana, Illinois,
on January 12, 1997."
To celebrate this rather auspicious date (in the true Roman sense of
"auspicious" -- but without the birds, of course, as Jim Wowchuk would be
quick to point out), the University of Illinois at Champagne/Urbana will
later this year celebrate HAL's birth with a "CyberFest." A schedule of
events for the CyberFest is posted at:
http://www.cyberfest.uiuc.edu/scheduleparent.html
The movie "2001" had a profound impact on a great many people of my age.
Indeed, when I first met my wife in 1971, one of the first things that I ever
told her, romantic as it may be, was that I could ever own even a fragment of
a movie, it would the sequence from the opening credits of "2001" through the
appearance of the first obelisk with the ramapithecines to the almost
instananeous (4 Myr) evolution of man and space stations and the docking of
the Pan Am shuttle with the space station, choreographed to the Blue Danube
Waltz. The only words that are spoken during that entire sequence occur at
the very end when Dr. Heywood Floyd says as he steps off the air lock, "See
you next trip."
Not all movies are clearly of equal caliber. "2001" is literature in a way
that very few movies are. But more than that, it's technically accurate. Gene
Siskel, the movie critic for the Chicago Tribune, once sat next to Neil
Armstrong (the other Neil Armstrong, not the one from Robelle) and asked him
what it was really like on the moon. Armstrong asked if he'd seen "2001".
Siskel said, "Sure." "Well," Armstrong said, "that's pretty much what it was
like."
However, where the movie fails is in its unbridled optimism of the
development of artificially intelligent devices. In the near 30 years since
its production, we've progressed no further than we were. All that we're
still building are very fast hand-cranked calculators. Indeed, all we've
really done is begin to discover the depth of the problem and haven't begun
to scratch the surface as far as workable solutions.
This level of "discouragement" is not mine alone. It's fairly rampant among
AI people nowadays -- which means that the fundamental approach adopted by
most people doing this sort of research is undoubtedly wrong. You can get a
sense of this discouragement from a recent USA Today article that's reprinted
at:
http://207.123.208.11:80/life/enter/books/leb617.htm
But the core of the problem really is probably nothing more than the approach
taken. Ultimately, machinery will have to become wholly self-adaptive in a
true evolutionary sense. And there are people out there who think that this
approach offers real promise. One of them is my most recent doctoral student,
David Fogel, who published his dissertation as a 1995 book entitled,
"Evolutionary Computation: Towards a new philosophy of machine intelligence."
A web page describing (at least indirectly) his book can be found at:
http://www.natural-selection.com/misc/evolCompBook.html
Wirt Atmar
|
