Social Choice Theory and Artificial Intelligence by ecg16852


									                     Social Choice Theory and Artificial Intelligence

                                            PD Dr. Pascal Hitzler
                                                Dr. Guilin Qi
                                             Seminar, SS 2008
                                      AIFB, Universität Karlsruhe (TH)

 Institute AIFB Karlsruhe, Germany
                                      General Expectations

 You gets one recent research paper
 You are expected to find closely related literature by yourself
 You select the material from the paper and the related
  literature, which you want to present

 Presentations are 60 minutes
 Presentations must be in English language
 A manuscript (10 pages) must be handed in at the end of the
 Grades are combined: 2/3 presentation, 1/3 manuscript
 You are also expected to come to all the meetings

 Institute AIFB Karlsruhe, Germany
                                          Organisational Issues

 We will present the offered topics now
 We will distribute the topics among you (now or within the
  next week)
 We will fix a schedule for the meetings
        • Can we move the meeting time/weekday away from Tuesdays 17:30?
        • Shall we make a block seminar?

 Supervision will be done by Dr. Qi
 You are expected to contact him at least once, namely when
  you have an outline/structure of your talk
 Make contact any time when you have questions

 Institute AIFB Karlsruhe, Germany
1.  PageRank
2.  survey, user preferences (elicitation)
3.  qualitative decision theory (desire)
4.  Conditional Preference networks (qualitative)
5.  preferences over sets (tradeoffs-enhanced CP-nets)
    [builds on 4]
6. belief merging – relationships between different merging
    operators (needs some logic)
7. belief merging – strategy-proofness (also needs some logic)
8. belief revision, merging – and voting
9. aggregating conflicting beliefs (some logic)
10. voting rules as maximum likelihood estimators (no logic)
11. judgement aggregation (Arrow's theorem) (little bit of logic)
12. judgement aggregation under constraints
 Institute AIFB Karlsruhe, Germany

To top