This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 10708 Graphical Models: Homework 2 Due October 17th, beginning of class October 3, 2005 1 [15 pts] Clique Tree Imaps In order to formalize the relationship between clique trees and Bayesian Networks, in this question you will prove that if P factorizes according to a Bayesian Network, then any clique tree T for this BN is an Imap for P . Specifically; In a clique tree, consider a separator S ij between two cliques C i and C j . Let X be any set of variables in the C i side of the tree, and Y be any set of variables in the C j side of the tree. Prove that P  = ( X Y  S ij ). 2 [15 pts] Clique Tree Factorization 2.1 [7 pts] Prove that the clique beliefs ( C i ) = P ( C i ) and edge beliefs ij ( S ij ) = P ( S ij ) form a fixed point for the belief propagation algorithm for a clique tree, i . e . , if we start BP with these beliefs, no messages will change them. 2.2 [8 pts] Using the independencies you showed in Question 1, prove that in a clique tree for a BN we can represent the joint distribution by: P ( X ) = producttext i P ( C i ) producttext ij P ( S ij ) . You should not prove by corollary from the correctness of BP in clique trees. (Hint: combine the chain rule of probabilities with the definition of conditional probabilities.) 1 3 [6 pts] Triangulation 3.1 [3 pts] A C F G E B D Figure 1: Moralized Graph for Prob. 2.1 Consider the graph in Figure 1.Consider the graph in Figure 1....
View Full
Document
 Fall '07
 CarlosGustin

Click to edit the document details