CSCE 582 {=STAT 582} Lecture Notes
- Errata for [J].
- Notes for Introductory Lectures
.
- Hugin GUI screenshots for introductory
lectures
.
- Transcript of notes of lectures of
January 14, 2016, with the Icy Roads and Wet Lawn
examples (pdf).
- Transcript of notes of lectures of
January 16, 2020, with an example of compositional certainty factor
computation.s (pdf).
- Transcript of notes of lectures of
January 16, 2014, with the
Burglary example and the definitions of d-separation and Markov
blanket (pdf).
- Lauritzen's algorithm for
d-separation with proof of correctness, as discussed in class on 2016-01-21.
- Transcript of notes of lectures of
January 21, 2016, with Lauritzen'a algorithm for d-separation (pdf).
- Transcript of notes of lectures of
January 28, 2020, with the definition of d-separation and
Lauritzen'a algorithm for d-separation (pdf).
- Transcript
of notes of lecture of 2022-02-10 with the statement
of Theorem 2.7 on p.51 of Lauritzen's _Graphical Models_.
- Transcript of notes of lectures of
February 16, 2022, with summary of Section 5.4 [J]
and examples of valuation-based systems.
- Transcript of notes of lecture of
January 23, 2014, with an example of a Dutch Book, used on
2016-01-26 (pdf).
- Transcript of notes of lecture of
January 19, 2012, with the solution of the Icy Roads example
used on 2016-01-26 (pdf), using
table operations on probability distributions.
-
Transcript of notes of lecture of
January 24, 2012, with the solution of the Icy Roads and Wet Lawn examples
used on 2021-02-02 (pdf), using
table operations on probability distributions.
- Slides for Ch.1 [J07]
(pdf).
- Slides for Ch.2 [J07]
(pdf).
- Transcript of notes of lecture of
January 24, 2012, with the solution of the "Wet Grass" and "Athenian Taxis"
examples (pdf), using table operations on probability distributions, used
on 2016-01-28.
- Transcript of notes of lecture of
2014-01-30, with the proof of the chain rule for Bayesian networks.
(This is an informal proof; a formal proof would use induction in the number
of variables in the BN.)
- Tables computed by bucket elimination
for a belief update example
- Tables computed by bucket elimination
for an MPE example
- Transcript of notes of lecture of
2016-02-18, with a sketch of the proof that BNs admit d-separation using
the graphoid axioms
- Transcript of notes
with the description of the stratum method (pdf).
- Slides for Ch.3 [J07]
(pdf).
- Transcript of notes used in the lectures
of 2016-03-17 and 2016-03-22 (pdf).
- Slides for
sections 4-4.6 [J07], on junction tree propagation.
(pdf).
- Paper used in lecture of
April 5, 2016 (local copy, pdf): Finn V. Jensen and Frank Jensen.
"Optimal Junction Trees." Uncertainty in Artificial Intelligence:
Proceedings of the Tenth Conference (UAI-94) (Ramon Lopez de Mantaras and
David Poole, eds.), pp.360-366.
- Transcript of notes of lecture of March
20, 2009 (pdf).
- Transcript of notes of lecture of March
23, 2009 (pdf).
- Transcript of notes of lecture of March
25, 2009 (pdf).
- Transcript of notes of lecture of March
27, 2009 (pdf).
- Transcript of notes of lecture of April
1, 2009 (pdf).
- Slides for Ch.7 [J07]
(pdf).
- Transcript
of notes from the lecture of 2021-03-25 with
suggestions for the extra assignment for graduate students.
- Applying the PC Algorithm to
the Visit to Asia (Chest Clinic) Network with a D-separation oracle and
with perfect conditional independence testing.
- Slides for Ch.9 [J07]
(pdf).
- Slides for Section 6.2 [J07]
on the EM algorithm (pdf).
- Slides for Section 4.8 [J07]
on approximate inference (pdf).
- Notes on how to use
Hugin to construct a Bayesian network implementation of
a classifier.
-
Slides on solving influence diagrams using variable elimination and strong
junction trees by Scott Langevin and Jingsong Wang (fall 2010).