your answer.
2
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Figure 2: Graphical Model for Prob. 2
2.2
[8 pts]
Now, let the distribution of
X
be given by some graphical model instance,
B
= (
G
, P
).
Consider variable
X
i
. What is the minimal subset of the variables,
A
 X  {
X
i
}
, such that
X
i
is independent of the rest of the variables,
X 
A
∪{
X
i
}
, given
A
? Prove that this subset
is necessary and sufficient.
(Hint: Think about the variables that
X
i
cannot possibly be conditionally independent of,
and then think some more)
2.3
Extra Credit [8 pts]
Show how you could efficiently compute the distribution over a variable
X
i
given some as
signment to all the other variables in the network:
P
(
X
i

x
1
, ..., x
i

1
, x
i
+1
, ..., x
n
).
Your procedure should not require the construction of the entire joint distribution
P
(
X
1
, ..., X
n
).
Specify the computational complexity of your procedure.
3
Marginalization [15 pts]
1. Consider the Burglar Alarm network shown in Figure 3. Construct a Bayesian network
over all of the nodes except for Alarm, which is a minimal Imap for the marginal
distribution over those variables defined by the above network.
Be sure to get all
dependencies that remain from the original network.
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '13
 Dr.ZAre
 Probability theory, pts, 3 pts, 4 pts, graphical model

Click to edit the document details