Below you find a description of slide style you might consider to use, various topics and a description of the intended format of the seminar.

• Proposed is to use prosper latex style (available for fedora as rpm)
• Example slides with CI style (PPPRCI.sty) included (unpack in a directory by: tar -xvf example.tar)
• Example of a script that converts the latex ouput to a PDF file for presentation

List of topics:

A. Philosophical Foundations of Reasoning with Uncertainty:
1. The foundations of statistics reconsidered (Savage, 1961)
2. Why isn't everyone a Bayesian? (Efron, 1986)
3. Judgement under uncertainty: heuristics and biases (Tversky and Kahneman, 1974)
4. Languages and designs for probabilistic judgement (Shafer and Tversky, 1985)
B. Technical Aspects of Probabilistic and Decision-theoretic Models:
1. Structure learning of BNs:
   • Book BAI 6.3, chapter 8;
   • Learning as search paper by Castelo and Kocka
   • Use of genetic algorithms in BN structure learning paper by Larranage et al.
   • Constraint-based learning paper by Cheng et al.
   • Constraint-based learning follow-up paper by Chichering and Meek
2. Learning probabilities and Expectation-Maximization (EM) algorithm:
   • Book: BAI chapter 7;
   • Tom Mitchell, Machine Learning, McGraw-Hill, 1997, Section 6.12 (good examples)
   • Paper by Bilmes
3. Decision networks and influence diagrams:
   • Book: BAI 4.1-4.4
   • Paper by Shachter
4. Dynamic Bayesian networks:
   • Book: BAI 4.5;
   • Paper by Murphy
5. Evaluation (sensitivity analysis and validation methods):
   • Book: BAI chapter 10;
   • Paper by Van der Gaag , Renooij and Coupe
6. Markov logic networks:
   • Paper by Richardson and Domingos
7. Missing data and Bayesian networks learning
   • Paper by Ramoni and Sebastiani
8. Qualitative probabilistic networks (QPNs):
   • Book: BAI, page 250
   • Paper by Wellman
9. Contextual independence:
   • Paper by Boutilier, Friedman, Goldszmidt and Koller
10. Model-based diagnosis using Bayesian networks:
    • Paper by Flesch et al.
C. Cognitive Science, Bayesian Statistics and PGMs :
1. Cognitive science and Bayes
   • Subjective probability paper bu Kahneman and Tversky material
   • Optimal predictions paper by Griffith and Tenenbaum
   • UCI teaching material
   • Paper by by Kersten and Yuille
   • Paper by Griffiths, Kepms and Tenenbaum
   • Fallacies in legal reasoning by Fenton et al
2. Structural statistical modelling:
   • Paper by Tenenbaum, Griffiths and Kemp
   • Paper by Kemp and Tenenbaum
3. Brain image interpretation using Bayesian networks:
   • Paper by Mitchell et al.
   • Paper by Rajapakse and Zhou

• Seminar on Tuesday, preparatory meeting with lecturers on Thursday
• Topics can be prepared and presented in pairs
• Start with 10 minutes where other student think about a problem statement or try to solve an exercise
• Presentation of the topic of about 25-30 minutes
• Discussion (about 5-10 minutes)
• Duration of one presentation: 45 minutes (including discussion)
• Slide style: try to use CI prosper style

Note: you will get a mark for the presentation!