 |
Science File Information:
|
-
It all starts when the CEO of a computer company decides to axe an artificial
intelligence (AI) project being developed by his company, Eurisko.
Eurisko is a very clever choice as the name of a computer company
inventing an artificial intelligence. First of all, as AI programmer
Brad Wilczek explains to Agent Mulder, it is Greek for "I discover things."
Secondly, Eurisko is the title given to a branch of AI philosophy that suggests
that computers can be programmed to develop their own heuristics, or methods
of solving problems.
For more information, look to
http://www.cs.indiana.edu/hyplan/gasser/C463_4/am.html,
a page from an Indiana University class on AI. It explains the concepts of both Eurisko
and heuristics.
-
Unfortunately for Eurisko's impetuous CEO, the entire office building is
maintained--from the phone lines to the climate control--by a prototype of the
artificial intelligence called the "Central Operating System." The computer does
not take too kindly to someone wanting to shut it down, and it electrocutes the CEO.
Ouch.
To find out how much electricity will kill, and to read a comparison of the resistance
of wet and dry skin, go to
http://www.elmwoodelectric.com/fatalcurrent.html.
-
Scully runs a voiceprint match on the Central Operating System's phone message only to
find that it's a near-exact match for the AI's creator, Brad Wilczek. Voiceprint
analysis is a very real tool of forensic investigators.
The theory behind voiceprint spectrographic analysis is explained at
http://www.aftiinc.com/voice.htm, a page
from Applied Forensic Technologies International,
a company of forensic consultants.
-
Later, Mulder meets with his "Deep Throat" informant to discuss why Brad Wilczek's
computer research would interest the Defense Department. "Have you ever heard of
adaptive networks,
Mr. Mulder?" he asks. Such artificial intelligence, though far from the level of
the COS in this episode, or from HAL 9000
for that matter, is being researched every day at venerable universities around
the nation.
For more information on reinforcement learning, a method used by the University of
Massachusetts' Adaptive Networks Labs, check out
http://envy.cs.umass.edu/anw-home-page.html.
For a broader perspective on AI construction, look to the University of Michigan's
own program on Cognitive and Agent Architectures at
http://ai.eecs.umich.edu/cogarch0/.
|
![[Go back to index page...]](smx5.gif)