Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!swrinde!ucsd!sdcc6!mangani!demers
From: demers@mangani.ucsd.edu (David Demers)
Newsgroups: comp.ai.neural-nets
Subject: Request for data: Boltzmann machines
Message-ID: <14361@sdcc6.ucsd.edu>
Date: 20 Nov 90 00:55:40 GMT
Sender: news@sdcc6.ucsd.edu
Organization: CSE Dept., UC San Diego
Lines: 24
Nntp-Posting-Host: mangani.ucsd.edu



I have a request for data.  I am looking at alternative methods for
finding the global minimum of an energy function in
Boltzmann machines (alternative to simulated annealing).
I would like to know if anyone could help provide either:

1)	a weight matrix and the set of patterns "stored"
for a network with a reasonably large number of units
(minimum 30-ish, max of say 200 or so).  
or
2)	code for Boltzmann learning. I have a Sun4, SGI 
Personal Iris and NeXT at my disposal.

Of course, I could write up the Boltzmann learning algorithm myself, 
but prefer not to if it's been done (and I'm a slow programmer...).  

Thanks for any help!

-- 
Dave DeMers			demers@cs.ucsd.edu
Dept. of Computer Science & Engineering
UCSD
La Jolla, CA 92093-0114		(619) 534-8187 or 534-0118