home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!bunyip.cc.uq.oz.au!topaz.ucq.edu.au!hewittj
- From: hewittj@topaz.ucq.edu.au
- Newsgroups: comp.ai.neural-nets
- Subject: Boltzmann Machine a la Patrick Simpson.
- Message-ID: <1992Sep10.113218.12280@topaz.ucq.edu.au>
- Date: 10 Sep 92 11:32:18 AET
- Organization: University of Central Queensland, Australia
- Lines: 21
-
-
-
- As part of a 3rd year computing project I have been trying to implement a
- general Boltzmann Machine Neural Network for experimentation.
-
- The algorithm I am using is from the book "Artificial Neural Systems" by
- Patrick Simpson. Simpson's algorithm is the most complete example I have
- found for the B.M (even if it does seem simplistic).
- !! However I can't get it to work - no matter what I try (and believe me,
- I have tried many algorithm variations).
-
- I've read a lot of background material, however nearly all of it is NOT
- very specific on the network architecture and algorithm details.
-
- There must be someone out there who has successfully coded the algorithm
- for the B.M. Could someone p l e a s e suggest some practical readings/
- examples that might be of help ?????
-
- Thanks muchly in advance..
- John Hewit.
-
-