University of Cambridge > Talks.cam > Machine Learning @ CUED > Approximation strategies for structure learning in Bayesian networks

Approximation strategies for structure learning in Bayesian networks

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Zoubin Ghahramani.

Note: via Skype!

Structure discovery in Bayesian networks has attracted considerable interest in the recent decades. Attention has mostly been paid to finding a structure that best fits the data under certain criterion. The optimization approach can lead to noisy and partly arbitrary results due to the uncertainty caused by a small amount of data. The so-called full Bayesian approach addresses this shortcoming by learning the posterior distribution of structures. In practice, the posterior distribution is summarized by constructing a representative sample of structures, or by computing marginal posterior probabilities of individual arcs or other substructures. The state-of-the-art sampling algorithms draw orderings of variables along a Markov chain. We have proposed several improvements to these algorithms. In this talk I discuss these improvements.

This talk is part of the Machine Learning @ CUED series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2017 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity