From 15e4871fc224e9e74c93b772b15aea7031f262ab Mon Sep 17 00:00:00 2001 From: jeanpouget-abadie Date: Thu, 5 Nov 2015 09:35:44 -0500 Subject: adding simple bayes function --- notes/extensions.tex | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'notes') diff --git a/notes/extensions.tex b/notes/extensions.tex index dd9933a..cc247ef 100644 --- a/notes/extensions.tex +++ b/notes/extensions.tex @@ -31,7 +31,11 @@ network learning however, we can place more informative priors. We can \item Take into account common graph structures, such as triangles \end{itemize} -We can sample from the posterior by MCMC. +We can sample from the posterior by MCMC. This might not be a very fast solution +however. In the case of the independent cascade model, there is an easier +solution: we can use the EM algorithm to compute the posterior, by using the +parent that \emph{does} infect us (if at all) as the latent variable in the +model. \subsection{Active Learning} -- cgit v1.2.3-70-g09d2