Discrete restricted Boltzmann machines

Guido Montúfar, Jason Ryder Morton

Research output: Contribution to conferencePaper

5 Citations (Scopus)

Abstract

We describe discrete restricted Boltzmann machines: probabilistic graphical models with bipartite interactions between visible and hidden discrete variables. Examples are binary restricted Boltzmann machines and discrete naïve Bayes models. We detail the inference functions and distributed representations arising in these models in terms of configurations of projected products of simplices and normal fans of products of simplices. We bound the number of hidden variables, depending on the cardinalities of their state spaces, for which these models can approximate any probability distribution on their visible states to any given accuracy. In addition, we use algebraic methods and coding theory to compute their dimension.

Original languageEnglish (US)
StatePublished - Jan 1 2013
Event1st International Conference on Learning Representations, ICLR 2013 - Scottsdale, United States
Duration: May 2 2013May 4 2013

Conference

Conference1st International Conference on Learning Representations, ICLR 2013
CountryUnited States
CityScottsdale
Period5/2/135/4/13

Fingerprint

fan
Probability distributions
Fans
coding
Ludwig Boltzmann
interaction
Visible
Interaction
Hidden Variables
Distributed Representation
Inference
Cardinality
Nave
Bayes Model
Algebra

All Science Journal Classification (ASJC) codes

  • Education
  • Computer Science Applications
  • Linguistics and Language
  • Language and Linguistics

Cite this

Montúfar, G., & Morton, J. R. (2013). Discrete restricted Boltzmann machines. Paper presented at 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, United States.
Montúfar, Guido ; Morton, Jason Ryder. / Discrete restricted Boltzmann machines. Paper presented at 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, United States.
@conference{a591541acee84a58a8d5a49a1d2dcec7,
title = "Discrete restricted Boltzmann machines",
abstract = "We describe discrete restricted Boltzmann machines: probabilistic graphical models with bipartite interactions between visible and hidden discrete variables. Examples are binary restricted Boltzmann machines and discrete na{\"i}ve Bayes models. We detail the inference functions and distributed representations arising in these models in terms of configurations of projected products of simplices and normal fans of products of simplices. We bound the number of hidden variables, depending on the cardinalities of their state spaces, for which these models can approximate any probability distribution on their visible states to any given accuracy. In addition, we use algebraic methods and coding theory to compute their dimension.",
author = "Guido Mont{\'u}far and Morton, {Jason Ryder}",
year = "2013",
month = "1",
day = "1",
language = "English (US)",
note = "1st International Conference on Learning Representations, ICLR 2013 ; Conference date: 02-05-2013 Through 04-05-2013",

}

Montúfar, G & Morton, JR 2013, 'Discrete restricted Boltzmann machines' Paper presented at 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, United States, 5/2/13 - 5/4/13, .

Discrete restricted Boltzmann machines. / Montúfar, Guido; Morton, Jason Ryder.

2013. Paper presented at 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, United States.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Discrete restricted Boltzmann machines

AU - Montúfar, Guido

AU - Morton, Jason Ryder

PY - 2013/1/1

Y1 - 2013/1/1

N2 - We describe discrete restricted Boltzmann machines: probabilistic graphical models with bipartite interactions between visible and hidden discrete variables. Examples are binary restricted Boltzmann machines and discrete naïve Bayes models. We detail the inference functions and distributed representations arising in these models in terms of configurations of projected products of simplices and normal fans of products of simplices. We bound the number of hidden variables, depending on the cardinalities of their state spaces, for which these models can approximate any probability distribution on their visible states to any given accuracy. In addition, we use algebraic methods and coding theory to compute their dimension.

AB - We describe discrete restricted Boltzmann machines: probabilistic graphical models with bipartite interactions between visible and hidden discrete variables. Examples are binary restricted Boltzmann machines and discrete naïve Bayes models. We detail the inference functions and distributed representations arising in these models in terms of configurations of projected products of simplices and normal fans of products of simplices. We bound the number of hidden variables, depending on the cardinalities of their state spaces, for which these models can approximate any probability distribution on their visible states to any given accuracy. In addition, we use algebraic methods and coding theory to compute their dimension.

UR - http://www.scopus.com/inward/record.url?scp=84925368276&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84925368276&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:84925368276

ER -

Montúfar G, Morton JR. Discrete restricted Boltzmann machines. 2013. Paper presented at 1st International Conference on Learning Representations, ICLR 2013, Scottsdale, United States.