MR-Received: by mta STKMTS; Relayed; Mon, 21 Oct 1996 14:11:14 +0000
MR-Received: by mta VALMTS; Relayed; Mon, 21 Oct 1996 15:01:09 +0000
Alternate-recipient: prohibited
Date: Mon, 21 Oct 1996 12:44:03 +0000 (GMT)
From: "Tomas Andersson @GOO" <ANDERSSON.TOMAS@A1STKAI1.STKMTS.SOO.mts.dec.com>
Subject: Best topology for multiplication?
Message-ID: <B817ZWNQN4QK6*/R=SOO/R=STKMTS/R=A1STKAI1/U=ANDERSSON.TOMAS/@MHS>
MIME-version: 1.0
Content-type: TEXT/PLAIN; CHARSET=US-ASCII
Content-transfer-encoding: 7BIT
Posting-date: Mon, 21 Oct 1996 14:11:00 +0000 (GMT)
Importance: normal
Priority: normal
Sensitivity: Company-Confidential
UA-content-id: B817ZWNQN4QK6
X400-MTS-identifier: [;41114112016991/8550487@STKMTS]
A1-type: MAIL
Hop-count: 2
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!news.duq.edu!newsgate.duke.edu!news.mathworks.com!howland.erols.net!EU.net!usenet2.news.uk.psi.net!uknet!usenet1.news.uk.psi.net!uknet!dispatch.news.demon.net!demon!mail2news.demon.co.uk
Newsgroups: comp.ai.neural-nets
X-NNTP-Posting-User: ANDERSSON.TOMAS@a1stkai1.stkmts.mts.dec.com
X-Mail2News-Path: news.demon.net!server21.digital.fr!mail.vbo.dec.com!valmts.vbe.dec.com!mts.dec.com
Lines: 20


 I'm evaluating the performance of a neural network OLE server with
 respect to a number of "elementary" problems, and I need to know
 if there is a simple answer to the following simple question:

 What is the most suitable topology for a multi-layered perceptron
 that is trying to learn how to calculate the function

	F = A * B

 where A and B are real numbers ranging from -1 to +1?

 I'm primarily interested in an obtaining an accurate solution, but
 learning speed is also a relevant parameter.

 Thanks for your time,

 - T. A. -
 

