Deep Boltzmann Machines

Ruslan Salakhutdinov, Geoffrey Hinton
; Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics, PMLR 5:448-455, 2009.

Abstract

We present a new learning algorithm for Boltzmann machines that contain many layers of hidden variables. Data-dependent expectations are estimated using a variational approximation that tends to focus on a single mode, and data-independent expectations are approximated using persistent Markov chains. The use of two quite different techniques for estimating the two types of expectation that enter into the gradient of the log-likelihood makes it practical to learn Boltzmann machines with multiple hidden layers and millions of parameters. The learning can be made more efficient by using a layer-by-layer “pre-training” phase that allows variational inference to be initialized by a single bottom-up pass. We present results on the MNIST and NORB datasets showing that deep Boltzmann machines learn good generative models and perform well on handwritten digit and visual object recognition tasks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v5-salakhutdinov09a, title = {Deep Boltzmann Machines}, author = {Ruslan Salakhutdinov and Geoffrey Hinton}, booktitle = {Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics}, pages = {448--455}, year = {2009}, editor = {David van Dyk and Max Welling}, volume = {5}, series = {Proceedings of Machine Learning Research}, address = {Hilton Clearwater Beach Resort, Clearwater Beach, Florida USA}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v5/salakhutdinov09a/salakhutdinov09a.pdf}, url = {http://proceedings.mlr.press/v5/salakhutdinov09a.html}, abstract = {We present a new learning algorithm for Boltzmann machines that contain many layers of hidden variables. Data-dependent expectations are estimated using a variational approximation that tends to focus on a single mode, and data-independent expectations are approximated using persistent Markov chains. The use of two quite different techniques for estimating the two types of expectation that enter into the gradient of the log-likelihood makes it practical to learn Boltzmann machines with multiple hidden layers and millions of parameters. The learning can be made more efficient by using a layer-by-layer “pre-training” phase that allows variational inference to be initialized by a single bottom-up pass. We present results on the MNIST and NORB datasets showing that deep Boltzmann machines learn good generative models and perform well on handwritten digit and visual object recognition tasks.} }
Endnote
%0 Conference Paper %T Deep Boltzmann Machines %A Ruslan Salakhutdinov %A Geoffrey Hinton %B Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2009 %E David van Dyk %E Max Welling %F pmlr-v5-salakhutdinov09a %I PMLR %J Proceedings of Machine Learning Research %P 448--455 %U http://proceedings.mlr.press %V 5 %W PMLR %X We present a new learning algorithm for Boltzmann machines that contain many layers of hidden variables. Data-dependent expectations are estimated using a variational approximation that tends to focus on a single mode, and data-independent expectations are approximated using persistent Markov chains. The use of two quite different techniques for estimating the two types of expectation that enter into the gradient of the log-likelihood makes it practical to learn Boltzmann machines with multiple hidden layers and millions of parameters. The learning can be made more efficient by using a layer-by-layer “pre-training” phase that allows variational inference to be initialized by a single bottom-up pass. We present results on the MNIST and NORB datasets showing that deep Boltzmann machines learn good generative models and perform well on handwritten digit and visual object recognition tasks.
RIS
TY - CPAPER TI - Deep Boltzmann Machines AU - Ruslan Salakhutdinov AU - Geoffrey Hinton BT - Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics PY - 2009/04/15 DA - 2009/04/15 ED - David van Dyk ED - Max Welling ID - pmlr-v5-salakhutdinov09a PB - PMLR SP - 448 DP - PMLR EP - 455 L1 - http://proceedings.mlr.press/v5/salakhutdinov09a/salakhutdinov09a.pdf UR - http://proceedings.mlr.press/v5/salakhutdinov09a.html AB - We present a new learning algorithm for Boltzmann machines that contain many layers of hidden variables. Data-dependent expectations are estimated using a variational approximation that tends to focus on a single mode, and data-independent expectations are approximated using persistent Markov chains. The use of two quite different techniques for estimating the two types of expectation that enter into the gradient of the log-likelihood makes it practical to learn Boltzmann machines with multiple hidden layers and millions of parameters. The learning can be made more efficient by using a layer-by-layer “pre-training” phase that allows variational inference to be initialized by a single bottom-up pass. We present results on the MNIST and NORB datasets showing that deep Boltzmann machines learn good generative models and perform well on handwritten digit and visual object recognition tasks. ER -
APA
Salakhutdinov, R. & Hinton, G.. (2009). Deep Boltzmann Machines. Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics, in PMLR 5:448-455

Related Material