Friday, September 14, 2007

Turing was there genetic algorithms and a biological approach to solving the AI

Turing was there

  • 29 April 2000
  • From New Scientist Print Edition. Subscribe and get 4 free issues.
  • Craig Webster, Auckland, New

I enjoyed your interview with Stephen Grand, creator of norns—some of the most sophisticated artificial life around (1 April, p 42). However, I found his bald statement that artificial intelligence has failed because Alan Turing's predictions about it have not come to pass a little glib.

Grand proposes the use of genetic algorithms and a biological approach to solving the AI problem—the very same things which Turing himself proposed way back in the 1940s. Turing discussed "genetical" searches and unusual neural network architectures, and carried out some of the first work in a field which has since become known as morphogenesis.

For a small example of functional Turing nets (to our knowledge the world's first) see http://home.clear.net.nz/pages/cw/unorganized.htm.

The development of the programmable digital computer based on Turing's designs so eclipsed these ideas that they have been more or less forgotten. If Turing hadn't died at the age of 42, or if these ideas had been developed, the world might have been quite a different place today. We would probably have more than just Macs and PCs on our desks.

From issue 2236 of New Scientist magazine, 29 April 2000, page 50

No comments: