Universität Osnabrück
Seminar: Basic methods of probabilistic reasoning - Details
You are not logged in.

General information

Course number 8.3273
Semester SS 2016
Current number of participants 39
Home institute LE Cognitive Science
Courses type Seminar in category Offizielle Lehrveranstaltungen
First date Mon , 04.04.2016 10:00 - 12:00, Room: 66/E01
SWS 2
Sprache Englisch
Contact Hours 2
ECTS points 4

Course location / Course dates

66/E01 Monday: 10:00 - 12:00, weekly (from 04/04/16) (13x)

Comment/Description

Prerequisites: Participants should be well familiar with the most basic concepts of probability theory (probability distributions, independence), logic (propositional, first-order logic) and graph theory (directed, undirected graphs).

Summary: Classical logics often fail to meet the demands of reality since they cannot express uncertainty. In probabilistic reasoning, this shortcoming is overcome by applying methods from probability theory. Probabilistic graphical models belong to the most successful class of methods for probabilistic reasoning and have been applied in different areas like expert systems, medical diagnosis and computer vision. They are also the fundamental building blocks of many formalisms considered in the research area of statistical relational learning. Even though graphical models often perform well computationally, their knowledge representation abilities are quite restricted. Probabilistic logics combine logics and probabilitiy theory and provide higher expressivity. Whereas the first probabilistic logics were designed in a model-theoretic way, some more recent approaches are designed proof-theoretically by extending classical logical programming techniques.