In article <6tkhh0$c80$1 at news.tamu.edu>,
wej3715 at scully.tamu.edu (Walter Eric Johnson) wrote:
>david_olmsted at my-dejanews.com wrote:
> :
> : If you want a slightly higher level approach to neural network simulation
> : consider my site at http://www.neurocomputing.org.>> Your web site pushes for the use of some sort of multi-valued
> logic to explaining the way the brain works. What evidence
> is there that your proposal is true? It is not enough to
> come up with a mechanism that says "it could work this way".
> There are often many schemes which appear to explain something
> but which are completely wrong when examined in detail.
>> (I didn't look that closely at your web site. I was supposed
> to be somewhere else an hour ago and so am rather short of
> time at the moment.)
>> Eric Johnson
>Eric,
Quite correct. The first stage of the site(now mostly complete)shows that
asynchronous multivalued logic neural networks can do some behaviorally
relevant information processing and do it better (faster, more robustly) than
any existing neural network. It also shows how multivalued logic operations
can be implemented by neurons. I am hoping that Yale University will soon put
up illustrations of microcircuits so web surfers can compare the those
proposed for multivalued logic operations will real microcircuits. (if not I
will just have to scan them from Gordon Shepherd's most recent book).
The second stage of the site on which I am now working is to provide detailed
reviews of the non-mammalian neuroscience literature in order to provide
macroscopic confirmation of these ideas. The reticular formation has mostly
been reviewed and its neural structure and function confirms the multivalued
logic model assigned to it.
The third stage will be to model the microcicuits of neurons at their ionic
level to provide microscopic confirmation by showing that all
their non-linearities appoximate multivalued logic operations.
So like any new theory only time will prove its correctness.
Sincerely,
David Olmsted
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum