Hi Ted and All:

>I really don't understand what is being discussed here.  From my point of
view (as
>a professional anatomist) I understand "his neural network system" as being
>essentially the same as his nervous system, i.e. his brain, spinal cord and
>nerves.  I don't think that is what is being meant here, so Allen, could
you clue
>me in on the real topic.

A neural net computer program just simulates the dendritic connections at
the end of the neurons.  There are several "layers" of neurons, and there
are connections possible from each neuron in one layer to each neuron in
another layer.  Then values are assigned to the neurons (frequencies, etc.
in this case) and the program is allowed to run through many trials trying
to predict a certain result (swarming in this case).  The computer "learns"
on its own through trial and error which pathways come closest to predicting
the desired result.

You can download some simple neural networks to playwith on your own
computer for free from shareware or Simtel.  Neural networks have been used
very successfully for predicting some odd things like the stock market and
generating computer speech.  Where the factors are so complex that all the
variables are difficult or impossible to put into equations to guide the
computer then neural nets are often very useful.

There is a professor in California named Terry Sejnowski I believe who is
doing work modelling bee brain/behaviour using neural nets.  I got a copy of
one of his papers because I have a continuing fascination with the size and
capability of the bee brain, but I would have to say that it was one of the
most incomprehensible (to me) things I have ever read.

Regards, Stan