lun apr 19 23:45:15 CEST 2010

An  introduction  to neural networks with an implementation using
my Neural++ library

--

I've just released a new version of my paper about  the  analysis
and design of neural networks. A new part has been inserted about
the theoretical background of supervised networks,  referring  to
the  perceptrons,  single-layer and multiple-layer networks, with
Widrow-Hoff or back-propagation training algorithms. A  new  ver-
sion  may soon arrive too, as I'm getting quite interested to the
not supervised networks and I'm reading a lot  about  these  net-
works.  It would be interesting if I could implement some not su-
pervised algorithms in my Neural++ library.  Moreover,  I'm  also
interested  in introducing in my library a robust self-adjustment
of both learning rate  and  inertial  momentum,  so  letting  the
synaptic weights self-adjust in the learning process and converg-
ing to the desired result in a shorter number of steps.

The     link     is     here.


--
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GAT d? a? C++++ U++++ P++++ L+++++ E--- W+++ !w PS+++ PE-- Y++ PGP++ X++
R+ tv-- b+>+++ DI++ G++ e+++ h* r++ z**
------END GEEK CODE BLOCK------