Artificial Intelligence
Artificial Intelligence
14 - 16 x 2 hours lesson +
14 x 2 hours exercises
Katharina Morik
The lecture is part of the post-graduate programme. It is obligatory for all
students of computer science. Therefore, it is held once a year. A script
of 223 pages offers the basic material for the lecture. The exercises are
done in Prolog.
Lessons and script cover the following topics:
- What is AI?
the topics, the approaches; Turing test, chinese room
- Problem solving as search
depth-, breadth- first search, hill-climbing, A*
and/or trees, minimax, alpha-/beta-pruning
- Production systems
interpreter strategies, RETE, expert systems
- Knowledge acquisition
KADS, sloppy modeling
- Problem solving by theorem proving
propositional logic: syntax, semantics, inference
predicate logic: syntax, semantics, interpretation, Herbrand theorem,
substitution, unification, resolution
- Knowledge Representation
semantic networks, frames
description logics: syntax, semantics, hybrid inference
- Natural language systems
architecture of a NLS
syntax: DCG, chart-parsing, Cocke-Younger-Kasami parser
lexicon: word form clauses
syntactic features and unification
how to develop a grammar for an application
semantics: reference, ambiguities, semantic representation languages
- Machine learning
learning as search: version space, ID3
conceptual clustering (star method, UNIMEM)
deductive learning
ILP: theta-subsumption, generalized theta-subsumption
learnability: identification in the limit (MIS), PAC-learning