Neural Network 'Near-Death Experiences'
Arthur T. Murray
uj797 at victoria.tc.ca
Mon Jun 24 16:17:18 EST 2002
matt at advancedatatools.com (Matt Lesko) wrote on 24 Jun 2002:
>Hello, I have been searching for information trained neural networks
>that 'relive' their early memories while 'dying' (that is, certain
>connections being severed). This comes from an experiment done by S.
>L. Thaler, and is referenced in Scientific American, May 1993 by
>Phillip Yam ("Daisy, Daisy: Do Computers Have Near-Death
http://www.scn.org/~mentifex/rejuve.html avoids neural-net death.
>I would dearly like to see the original article, or better yet, any
>sort of evidence, for or against, this activity. Putting Thaler's name
>into google gives little help, not only has he apparently passed away
>(although contradicted by another site), most of the links go to the
>rather outlandish Imagination Engines Inc.
>After doing a good bit of research, the only other lead sI have for
>this subject is a paper by Thaler in 'Neural Networks' Vol 8, No 1
>(1995) pgs 55-65 titled: "Virtual Input" Phenomena Within the Death of
>a Simple Pattern Associator. Unfortunately, the college library that I
>access to does not stock that particular journal. A post in
>comp.ai.neural-nets from 1994 mentions "4-2-4 Encoder Death, S.L.
>Thaler in 1993 World Congress on NN, Portland,
>Oregon, July 11-15, p. 180-183." but I can't find where I would be
>able to find this (online or in real-life).
>Anyone have any other information they could provide me with? I
>expected to find at least one bit of follow-up research to Thaler's
>initial work. More research could certainly make a bit of interesting
>paper, or at least something to talk about with your friends when
>really high ;)
>TIA for any help. -- Matt Lesko
More information about the Neur-sci