Skip to content

Latest commit

 

History

History
47 lines (47 loc) · 1.91 KB

2019-10-15-upadhya19a.md

File metadata and controls

47 lines (47 loc) · 1.91 KB
title crossref abstract layout series id month tex_title firstpage lastpage page order cycles bibtex_author author date address publisher container-title volume genre issued pdf extras
Efficient Learning of Restricted Boltzmann Machines Using Covariance Estimates
acml19
Learning RBMs using standard algorithms such as CD(k) involves gradient descent on the negative log-likelihood. One of the terms in the gradient, which involves expectation w.r.t. the model distribution, is intractable and is obtained through an MCMC estimate. In this work we show that the Hessian of the log-likelihood can be written in terms of covariances of hidden and visible units and hence, all elements of the Hessian can also be estimated using the same MCMC samples with small extra computational costs. Since inverting the Hessian may be computationally expensive, we propose an algorithm that uses inverse of the diagonal approximation of the Hessian, instead. This essentially results in parameter-specific adaptive learning rates for the gradient descent process and improves the efficiency of learning RBMs compared to the standard methods. Specifically we show that using the inverse of diagonal approximation of Hessian in the stochastic DC (difference of convex functions) program approach results in very efficient learning of RBMs.
inproceedings
Proceedings of Machine Learning Research
upadhya19a
0
Efficient Learning of Restricted Boltzmann Machines Using Covariance Estimates
836
851
836-851
836
false
Upadhya, Vidyadhar and Sastry, P S
given family
Vidyadhar
Upadhya
given family
P S
Sastry
2019-10-15
PMLR
Proceedings of The Eleventh Asian Conference on Machine Learning
101
inproceedings
date-parts
2019
10
15