Weighted Spectral Learning and the Efficiency Sharpening Algorithm

author: Michael Thon, Jacobs University Bremen
published: Oct. 6, 2014,   recorded: December 2013,   views: 1591

See Also:

Download slides icon Download slides: nipsworkshops2013_thon_spectral_learning_01.pdf (574.3┬áKB)

Help icon Streaming Video Help

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.


Predictive State Representations (PSRs), Observable Operator Models (OOMs) and Stochastic Multiplicity Automata (SMA) are basic types of models that have a common underlying algebraic structure. We briefly review this general structure and derive a generic learning framework into which the spectral learning algorithms fall. We introduce weights into the learning algorithm that reflect the accuracy of the estimates in the Hankel matrix, and show that this indeed improves the quality of the model estimation. Finally, we show that there is a close relationship between these weighted spectral learning algorithms and the efficiency sharpening algorithm for learning OOMs. The latter is an iterative algorithm that estimated the underlying subspace and weights from a previous model estimate, which has some computational advantages.

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: