Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!yeshua.marcam.com!usc!math.ohio-state.edu!howland.reston.ans.net!EU.net!sun4nl!news.nic.surfnet.nl!tudelft.nl!galaxy.twi.tudelft.nl!parmul
From: parmul@kgs.twi.tudelft.nl (PM project_Bernard)
Subject: Re: NN/GA Hybrids
Message-ID: <parmul.782727665@galaxy.twi.tudelft.nl>
Lines:       22
Sender: news@news.tudelft.nl (UseNet News System)
Nntp-Posting-Host: galaxy.twi.tudelft.nl
Organization: Delft University of Technology
References: <387eht$7qa@newsbf01.news.aol.com>
Date: Fri, 21 Oct 1994 08:21:05 GMT

emangman@aol.com (EmanGman) writes:

>I am a graduate student at George Mason University in Fairfax Virginia.  I
>am using a NN/GA hybrid for pattern classification in imagery.  

>My search of the literature has shown that GA algorithms have been used to
>develop the weights and/or structure of neural nets.  I have not been
>able, however,  to locate any other  NN/GA hybrids.  Is anybody aware of
>any NN/GA hybrids (other than using a GA to develop the weights and/or
>structure of a NN)?
Yes, sorta...
We used a GA to develop the weights and the parameters for the activation 
functions, such as the steepness for a sigmoid, and the radius for a 
RBFN, simultaneously. When doing this however, one has to avoid epistasis, 
for example by taking a suitable representation of the NN.

Bernard

-- 
========================================================================
Bernard Kroes                   | e-mail: parmul@kgs.twi.tudelft.nl
========================================================================
