Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!hellgate.utah.edu!caen!uflorida!gatech!mcnc!rti!bdrc!curry
From: curry@bdrc.bd.com (Joan Curry)
Newsgroups: comp.ai.neural-nets
Subject: references for conjugate gradient and cascade-correlation
Message-ID: <2695@bdrc.bdrc.bd.com>
Date: 18 Feb 91 20:44:15 GMT
Organization: Becton Dickinson Rsrch Cntr, RTP, NC
Lines: 23

Apologies if this has been submitted recently; our newsfeed has been
_very_ erratic lately. (And since I can't count on seeing answers in
the news, I'd appreciate getting any responses by email - I'll post
a summary if anyone requests it.)

Now ...

I am looking for references to reading material on:

 (1) improvements to the error minimization method of standard
     backpropagation - the one I've heard the most favorable
     comments about is conjugate gradient methods (which I don't
     really know what are), but I'm also interested in references
     to anything else that net-net-hackers think is good.

 (2) cascade-correlation method for building up a network.

Many thanks,
Joan
-- 
Joan Curry		Becton Dickinson Research Center
curry@bdrc.bd.com	21 Davis Drive, RTP, NC 27709
...!mcnc!bdrc!curry	(919) 990-2110