Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!crdgw1!rodson!bloomer
From: bloomer@rodson.crd.ge.com (John Bloomer)
Newsgroups: comp.ai.neural-nets
Subject: generic gradient decent references needed
Message-ID: <19796@crdgw1.crd.ge.com>
Date: 22 May 91 13:02:24 GMT
Sender: news@crdgw1.crd.ge.com
Reply-To: bloomer@rodson.crd.ge.com (John Bloomer)
Lines: 10

We are using a gradient-decent approach to motion estimation.  PEL-recursive
(Netravali et.al.), Horn and even more recent derivatives (no pun intended)
boil-down to this or at least some functional minimization involving gradients.

Can anyone give me some references on neural approaches to generic
gradient-decent?
 **
John J. Bloomer <jbloomer@crd.ge.com, uunet!rodson.crd.com!bloomer >
KWC317, General Electric Corporate Research and Development, (518)387-6964
PO Box 8, Schenectady NY 12301