home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!cam-eng!ajr
- From: ajr@eng.cam.ac.uk (Tony Robinson)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: search for literature on recurrent networks
- Keywords: recurrent, networks
- Message-ID: <1992Sep5.100106.26838@eng.cam.ac.uk>
- Date: 5 Sep 92 10:01:06 GMT
- References: <1992Sep4.155607.4536@a.cs.okstate.edu>
- Sender: ajr@eng.cam.ac.uk (Tony Robinson)
- Organization: Cambridge University Engineering Department, UK
- Lines: 19
- Nntp-Posting-Host: dsl.eng.cam.ac.uk
-
- In comp.ai.neural-nets siva@a.cs.okstate.edu (KAVUTURU SIVA RAMA) writes:
- >
- > I am searching for the literature on recurrent networks. If you
- > have any information regarding recurrent networks please send me.
- >
-
- A recent message to comp.ai.neural-nets gave details of how to access the
- anonymous ftp site svr-ftp.eng.cam.ac.uk. In directory "reports" you will
- find seven reports on recurrent nets. This includes my PhD thesis "Dynamic
- Error Propagation Networks" (how I wish I'd used the word "recurrent") and a
- recent turorial which has two pages on recurrent nets, giving about six
- different forms of supervised net and three unsupervised nets.
-
- Much of what a standard back-prop net does can be done with standard
- statistical techniques, often better and faster (e.g. weighted nearest
- neighbours, Gaussian mixtures, etc). But there seems to be nothing to match
- the processing power of a recurrent net classifier.
-
- Tony [Robinson]
-