Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!scramble.lm.com!news.math.psu.edu!news3.cac.psu.edu!howland.erols.net!news.sprintlink.net!news-peer.sprintlink.net!news.mathworks.com!tank.news.pipex.net!pipex!usenet2.news.uk.psi.net!uknet!usenet1.news.uk.psi.net!uknet!uknet!bhamcs!bham!sunserver1.aston.ac.uk!usenet
From: branass@aston.ac.uk (Suky Brana)
Subject: Capacity of Committee machines
X-Nntp-Posting-Host: venezuala.aston.ac.uk
Message-ID: <Dz7qwJ.6qA@aston.ac.uk>
Sender: usenet@aston.ac.uk
Reply-To: branass@aston.ac.uk
Organization: Aston University
Date: Sun, 13 Oct 1996 12:05:07 GMT
Lines: 19


I am conducting a numerical study of the capacity of committee machines.
The aim of the project is to study (numerically) the number of arbitrary 
input-output relations such a system can store. 

The commitee machine is a simple two layer neural network with binary-valued input and output.The weights between the hidden and output layer are fixed to +1 and there is a single output node. I will be using the CHIR2 algorithm to adapt the weights between the input and hidden layers.

I will be grateful for information on committee machines e.g factors that determine the capacity of committee machines etc.

references :
Computational Capabilities of Restricted Two-layered Perceptrons (1993)
Learning by CHIR without storing internal representations.

Thank You
Mr Sukhvinder Brana.

branass@aston.ac.uk


