Variance Approximation in Large-Scale Gaussian Markov Random Fields

author: Dmitry Malioutov, Stochastic Systems Group, Massachusetts Institute of Technology, MIT
published: Aug. 26, 2009,   recorded: June 2009,   views: 4362


Related Open Educational Resources

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.


In this talk we discuss a framework for computing accurate approximate variances in large scale Gaussian Markov Random Fields. We start by motivating the need to compute variances in GMRFs, and discuss related problems in machine learning. Our approach is based on constructing a certain low-rank aliasing matrix which takes advantage of the Markov graph of the model. We first construct such a matrix for models with short-range correlation, and then describe a wavelet-based construction for models with long-range correlation. The approach is based on fast solution of sparse linear systems, and we describe suitable preconditioners. We also describe how the approach can be used for problems with sparse plus low-rank structure, for example in approximate Kalman filtering with large state spaces.

See Also:

Download slides icon Download slides: icml09_malioutov_itva_01.pdf (1.6┬áMB)

Help icon Streaming Video Help

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: