Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!satisfied.elf.com!news.mathworks.com!news.kei.com!wang!news
From: peter@swamp.indigo.co.il (Peter Gordon)
Subject: TRAINREC
Organization: Indigo Ltd
Date: Thu, 26 Jan 1995 08:04:03 GMT
Message-ID: <PETER.95Jan26100404@swamp.indigo.co.il>
Sender: news@wang.com
Lines: 33


I have obtained and read the article 

TRAINREC: A System for Training Feedforward and Simple
          Recurrent Networks Efficiently and Correctly.

          ftp://thalamus.wustl.edu/
          ( I am not sure of the full address) 

          by Kalman and Kwansky


1. Is there any software available based on their theory 
   which I could play with.

2. Is there a cookbook paper that describes simply all the
   steps needed.

3. They talk about an SVD (singular value decomposition) and say that 

       A = UGV  (page 12) 
 
   They do not (apparently) give the method for calculating U, G and V.
   Could someone show me how to do this?


I am going to have a lot more questions with this :-)

Thanks 

Peter


