What is leaning in neural networks ?

pinchas tandeitnik pini1 at bgu.ac.il
Wed Jun 9 13:35:28 EST 1993


I'll be happy to receive your comments on the following questions:

What is leaning in neural networks ?

For the follow chart :

  |--------------|                                  |--------------|
  |  presynaptic |        |---------|               | postsynaptic |
  |              |------->| synapse |-------------->|              |
  |  cell        |        |---------|               | cell         |
  |--------------|    g(t)=A*t*exp(-t/tau)          |--------------|
   action potential                                  action potential
   at frequency f1                                   at frequency f2

 - Does a general learning means that the synaptic weight (A) changes ?
 - Does leaning mean also changing the synaptic time constant ?
 - Does f2 is a function of f1 and the synaptic parameters (A , tau) ?
 - What is the relation (magnitude) between f1 and f2 to the
   synaptic time constant ?
 - To me as  Electrical and Computer Engineering it seems the most
   natural way to model synaptic connection as a convolution between
   the presynaptic voltage and the synaptic conductance in order to
   receive the postsynaptic current, but the usual model is
   i_synapse=g(t)*[v(t)-v_synapse] ?

 Any comments and references will be helpful.

	         Sincerely yours

                                            Pinchas Tandeitnik.

   Department of Electrical and Computer Engineering,
   Ben-Gurion University of the Negev,
   P.O.Box 653, Beer-Sheva 84105, ISRAEL.
   email: pini at bguee.bgu.ac.il

More information about the Neur-sci mailing list