I need an algorithm for Boltzmann machine. Also need to know what is
an Boltzmann machine, and how is it useful, and how it works ?

---

Request for Question Clarification by hedgie-ga on 27 Apr 2005 08:15 PDT

Hmmm -- nirajaq-ga 
  If you do not know what it is - how do you know you need one?

And what is wrong with all the links you get when you type the term
into a seracj engine, definitions like?

"..The moral of this analogy for artificial nets is to employ
simulated annealing: to start the net off at a relatively high
temperature, as it cycles - at which you will find that coalitions of
units can be easily formed and broken, and gradually cool the net
down..."
http://www.mdx.ac.uk/www/psychology/cog/psy3250/Boltz/boltz.htm
 
What technical level, high school, colege, ... do you program, what is yur aim .. 
the more you tell us, more likely you get a useful answer.

Hedgie

---

Clarification of Question by nirajaq-ga on 27 Apr 2005 11:44 PDT

well, i study at a university, hence need in that detail

basically, BM is an problem solving algorithm but i need to know how it works

most importantly i need the algorithm, once i get the algorithm, ill
be fine with rest of the parts. thanks

OK - here are few links to click on, read and to click on references:

General intro to the field:

See paragraph which starts:
Boltzmann machine The Quick Facts about: Boltzmann machine
Quick Summary not found for this subjectBoltzmann machine can be
thought of as a noisy Hopfield network. Invented by Geoff Hinton and
Quick Facts about: Terry Sejnowski
Quick Summary not found for this subjectTerry Sejnowski (1985), the
Boltzmann machine was important because it was one of the first neural
networks in which learning of latent variables (hidden units) was
demonstrated....


http://www.absoluteastronomy.com/encyclopedia/n/ne/neural_network.htm

Then read: (SEARCH TERMS tellu you what researcher entered to the engine
to get suggested links. It is often useful to get additional references by 
entering those terms into Google or other SE).

SEARCH TERMS: Boltzmann machine algorithm
SEARCH TERMS: minimalization nonlinear

e.g.
A polynomial time algorithm for Boltzmann Machine learning

Boltzmann Machines (BMs) [1], are networks of binary neurons with a
stochastic neuron dynamics, known as Glauber dynamics. Assuming
symmetric connections between neurons, the probability distribution
over neuron states  will become stationary and will be given by the
Boltzmann-Gibbs distribution .


http://www.mbfys.kun.nl/~bert/cambridge/



... and learning rules in general Boltzmann machines, and towards the design of
 special-case algorithms. An attractive special case are networks, ...
http://www.doc.ic.ac.uk/~srueger/www-pub/dbmfd-imacs97.pdf 



SEARCH TERMS (category) :Simulated annealing

See:

The original Metropolis scheme was that an initial state of a
thermodynamic system was chosen at energy E and temperature T, holding
T constant the initial configuration is perturbed and the change in
energy dE is computed. If the change in energy is negative the new
configuration is accepted. If the change in energy is positive it is
accepted with a probability given by the Boltzmann factor exp -(dE/T).
This processes is then repeated sufficient times to give good sampling
statistics for the current temperature, and then the temperature is
decremented and the entire process repeated until a frozen state is
achieved at T=0.

http://www.cs.sandia.gov/opt/survey/sa.html



demo:
http://www.taygeta.com/annealing/demo1.html

book:
Simulated Annealing and Boltzmann Machines: A Stochastic Approach to
Combinatorial Optimization and Neural Computing
 Emile Aarts, Jan Korst

http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0471921467.html


Potentially useful library
http://attrasoft.com/imagefinder60/polyapplet/

more references:
http://citeseer.ist.psu.edu/context/3330/0

://www.google.com/search?hl=en&ie=UTF-8&q=Simulated+annealing,+Boltzmann&spell=1

Free Unsolicitated Advice: do not be afraid to use the Search Engine

enjoy

Hedgie