Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!newsfeed.internetmci.com!in1.uu.net!redstone.interpath.net!sas!mozart.unx.sas.com!saswss
From: saswss@hotellng.unx.sas.com (Warren Sarle)
Subject: Re: about neural network architecture
Originator: saswss@hotellng.unx.sas.com
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <Dnnron.xG@unx.sas.com>
Date: Sat, 2 Mar 1996 21:00:23 GMT
X-Nntp-Posting-Host: hotellng.unx.sas.com
References: <8BA44EB.01CE0053F1.uuout@almac.co.uk> <yyong-2402961347000001@lvl-mac006.usc.edu> <4gvc7pINNt5t@bhars12c.bnr.co.uk>
Organization: SAS Institute Inc.
Lines: 29


In article <4gvc7pINNt5t@bhars12c.bnr.co.uk>, pgh@bnr.co.uk (Peter Hamer) writes:
|> In article <yyong-2402961347000001@lvl-mac006.usc.edu> yyong@almaak.usc.edu (yong   yuan) writes:
|> >
|> >I recently received the following comment on some of my research network
|> >research:
|> >
|> >"Techniques have been evolved for modeling network architecture that are
|> >driven by the data being considered. These approaches allow for near
|> >optimization of paradigm architecture and the generation of networks that
|> >may perform substantially better than empirical designs."
|> ...
|> I too would love to hear about techniques which meet this criteria better than 
|> those I know of. 
|> 
|> Firstly, I think the poster who mentioned GA was guilty of a non sequitur. 
|> GA are indeed a valuable technique, but to me they represent the automation
|> of the empirical approach, not an alternative to it.

I am thoroughly confused by this question. What is the difference
between "driven by the data" and "empirical"?  I would consider the
former to be a crude definition of the latter.  And GAs use a fitness
function based on the data, so how are GAs not "driven by the data"?

-- 

Warren S. Sarle       SAS Institute Inc.   The opinions expressed here
saswss@unx.sas.com    SAS Campus Drive     are mine and not necessarily
(919) 677-8000        Cary, NC 27513, USA  those of SAS Institute.
