Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!newsfeed.internetmci.com!news.mathworks.com!uunet!in1.uu.net!bcstec!nntp
From: Mike Fouche <mikef@hsvaic.hv.boeing.com>
Subject: Re: Can you teach NN without knowing the transfer function?
X-Nntp-Posting-Host: e677744.hv.boeing.com
Content-Type: text/plain; charset=us-ascii
Message-ID: <DnLo9o.Jx1@bcstec.ca.boeing.com>
Sender: nntp@bcstec.ca.boeing.com (NNTP News Access)
Content-Transfer-Encoding: 7bit
Organization: Missiles & Space Division
References: <4gp4ik$md7@blaze.cs.jhu.edu>
Mime-Version: 1.0
Date: Fri, 1 Mar 1996 17:51:24 GMT
X-Mailer: Mozilla 1.2 (Windows; U; 16bit)
Lines: 8

You should be able to using the Finite Differencing method to numerically calculate 
the gradient of the net without knowing the transfer function.  One of our learning 
codes uses Finite Differencing in addition to the analytical calculation of the 
gradient.  It serves as a reality check in some cases.

Mike Fouche
Boeing Missiles & Space Division

