Scaling up Natural Gradient by Sparsely Factorizing the Inverse Fisher Matrix

author: Roger Grosse, Department of Computer Science, University of Toronto
published: Dec. 5, 2015,   recorded: October 2015,   views: 30
Categories

See Also:

Download slides icon Download slides: icml2015_grosse_fisher_matrix_01.pdf (6.4 MB)


Help icon Streaming Video Help

Related content

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Lecture popularity: You need to login to cast your vote.
  Bibliography

Description

Second-order optimization methods, such as natural gradient, are difficult to apply to high-dimensional problems, because they require approximately solving large linear systems. We present FActorized Natural Gradient (FANG), an approximation to natural gradient descent where the Fisher matrix is approximated with a Gaussian graphical model whose precision matrix can be computed efficiently. We analyze the Fisher matrix for a small RBM and derive an extremely sparse graphical model which is a good match to the covariance of the sufficient statistics. Our experiments indicate that FANG allows RBMs to be trained more efficiently compared with stochastic gradient descent. Additionally, our analysis yields insight into the surprisingly good performance of the “centering trick” for training RBMs.

Link this page

Would you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !

Write your own review or comment:

make sure you have javascript enabled or clear this field: