Approximate inference in Boltzmann machines
Inference in Boltzmann machines is NP-hard in general. As a result approximations are often necessary. We discuss first order mean field and second order Onsager truncations of the Plefka expansion of the Gibbs free energy. The Bethe free energy is introduced and rewritten as a Gibbs free energy. Fr...
Hlavní autoři: | Welling, M, Teh, Y |
---|---|
Médium: | Journal article |
Jazyk: | English |
Vydáno: |
2003
|
Podobné jednotky
-
Linear response for approximate inference
Autor: Welling, M, a další
Vydáno: (2004) -
Linear response algorithms for approximate inference in graphical models.
Autor: Welling, M, a další
Vydáno: (2004) -
Approximate inference by Markov chains on union spaces
Autor: Welling, M, a další
Vydáno: (2004) -
Rate-coded restricted Boltzmann machines for face recognition
Autor: Teh, Y, a další
Vydáno: (2001) -
Collapsed variational inference for HDP
Autor: Teh, Y, a další
Vydáno: (2009)