TU Wien:Knowledge-based Systems VU (Egly, Eiter, Tompits)/Oral Exam 2021-06-30

Aus VoWi
Zur Navigation springen Zur Suche springen

The questions were asked by Prof. Eiter and Prof. Tompits.

Eiter[Bearbeiten | Quelltext bearbeiten]

Intro Slides (unexpected for me)[Bearbeiten | Quelltext bearbeiten]

What is the Architecture of a KBS? (Additionally to the architecture on the slide, he wanted to hear about a "learning" component - not in the slides?)

What questions does a KBS usually answer? (Why-Questions and How-Questions - not in the slides?)

RBS[Bearbeiten | Quelltext bearbeiten]

What parts does a rule-based system consist of?

How does a rule-based system work? What is the cycle called (RAC)?

What is RETE?

How does a NOT-Node in RETE work?

DL[Bearbeiten | Quelltext bearbeiten]

What are Description Logics, explain the Vocabulary

We heard about Concept Constructors that also include roles, which are those and explain them?

What can you say about the complexity of KB-satisfiability? Why would you prefer tableaux over type elimination?

Is there another way to solve KB-satisfiability? (FOL2) Why don't we just use FOL2 Satisfiability - what complexity does FOL2 have? (NExp-Complete)

Tompits[Bearbeiten | Quelltext bearbeiten]

Logic Programs[Bearbeiten | Quelltext bearbeiten]

What syntactical kinds of programs did we learn about?

How is a rule defined syntactically? How does this definition change w.r.t normal, extended, disjunctive programs?

How do we interpret a classical program semantically? What is a herbrand model (including Herbrand basis, Herbrand universe and what they are, important to notice that we restrict the usage of functions so that the Herbrand universe is finite)? How is the fixed-point semantics defined? How to we calculate the least fixed-point in a set-theoretical sense? (by the union of all F↑k)

How are the semantics for normal logic programs defined?

What do you have to change about the semantics when considering extended disjunctive logic programs? (No least herbrand model, but minimal)


There were no questions about TMS and probabilistic reasoning.