In article <ysizpwc9e1g9.fsf at sapo.usc.edu> brannon at sapo.usc.edu (Terrence M. Brannon) writes:
"Brains, Machines, and Mathematics" by Michael A. Arbib shows that any
digital computer is equivalent to the simplest of neurons, the
McCulloch-Pitts neuron because
1) A digital computer can be created using of AND, OR, and NOT gates
2) the above three gates can be simulated by giving McCulloch-Pitts
neurons varying integration characteristics.
That's not actually a proof that a computer is equivalent to a neuron:
only that a neuron is equivalent to a computer. Did he also show that
*everything* the neuron does can be maped to AND, OR, and NOT gates?
-jen
--
---------------------------------------------------------------------
Jen Mankoff jmankoff at cc.gatech.edu 226E
http://www.cs.oberlin.edu/students/jmankoff/home.html
---------------------------------------------------------------------
