Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!hp4nl!star.cs.vu.nl!pjhamvs
From: pjhamvs@cs.vu.nl (Summeren van Peter)
Newsgroups: comp.ai.neural-nets
Subject: Re: Backprop Training
Message-ID: <7301@star.cs.vu.nl>
Date: 13 Aug 90 20:55:21 GMT
References: <1331@winnie.fit.edu>
Sender: news@cs.vu.nl
Reply-To: pjhamvs@cs.vu.nl (Summeren van Peter)
Organization: VU Dept. of Computer Science, Amsterdam, The Netherlands
Lines: 9

In article <1331@winnie.fit.edu>,
	dfausett@zach.fit.edu ( Donald W. Fausett) writes:
> If the input signal is zero, then the weight
> update is zero => the value of the weight does not change => no learning
> occurs.  When using backpropagation, it is always better to convert binary
> input patterns to bipolar form before training the network.

Except when zero means unknown
Greetings