Bounds and estimates for BP convergence on binary undirected graphical models

author: Joris Mooij, Radboud University Nijmegen
published: Feb. 25, 2007,   recorded: January 2005,   views: 4402


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.


Belief Propagation (BP) has become a popular method for inference on graphical models. Accurate approximations for intractable quantities (e.g. single-node marginals) can be obtained within rather modest computation times. However, for large interaction strengths (i.e. potentials that are highly dependent on their arguments) or densely connected graphs, BP can fail to converge. This can be remedied by damping the iteration equations, using sequential update schemes or using entirely different algorithms such as double-loop algorithms, that directly minimize the Bethe free energy. However, the value of this approach is questionable, since there is empirical evidence that failure of convergence of BP often indicates low quality of the Bethe approximation.

See Also:

Download slides icon Download slides: oiml05_joris_becbu_01.pdf (89.2┬áKB)

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: