Course details

8.3273

Basic methods of probabilistic reasoning

 S
WS 2019 Dr. rer. nat. Nico Potyka OFFLINE
2h/wk
4 ECTS 		B.Sc modules:
CS-BWP-AI - Artificial Intelligence
CS-BWP-NI - Neuroinformatics
KOGW-WPM-KI - Artificial Intelligence
KOGW-WPM-NI - Neuroinformatics
M.Sc modules:
CC-MWP-AI - Artificial Intelligence
CC-MWP-NI - Neuroinformatics
CS-MWP-AI - Artificial Intelligence
CS-MWP-NI - Neuroinformatics
 Mon: 14-16

Classical reasoning methods are too weak for many practical applications because they cannot deal with uncertainty. Various qualitative (e.g. non-monotonic reasoning) and quantitative (e.g. fuzzy reasoning, possibilistic reasoning) extensions have been proposed to overcome these shortcomings. In this course, we will discuss probabilistic methods that build up on probability theory. For instance, probabilistic logics extend classical logics by replacing the truth values 0 and 1 with the probability interval [0,1]. Reasoning results in classical logic are well founded by the laws of probability theory. However, in order to perform probabilistic reasoning in reasonable time, we often have to make independency assumptions. Probabilistic graphical models represent the independency structure graphically and offer various reasoning and learning algorithms that can exploit this structure. While Probabilistic graphical models can deal with large problems, their logical expressiveness is rather weak. Frameworks from the field of statistical relational AI like Markov Logic and ProbLog try to offer a better tradeoff between expressiveness and efficiency. The first part of the course will consist of lectures, where some of the main formalisms for probabilistic reasoning are introduced. This includes Bayesian and Markov networks, which are the main families of probabilistic graphical models. In particular, we will talk about the basic reasoning (variable elimination, belief propagation) and learning (parameter/structure) methods. In the second part of the course, participants will present advanced Topics and applications. Prerequisites: Basic knowledge in computer science and math, in particular probability Theory.