The course covers selected topics of artificial intelligence (AI) such as state space search, logic programming, natural language processing or machine learning. The emphasis is put to those AI topics which are either closely bound to a declarative programming language Prolog or which can be well presented within this context. There are explained principles of Prolog - a language designed for AI problem solving. Characteristic programming methodogy is introduced using specific solutions of typical AI problems (state-space search, simple expert system, etc.). Special attention is given to AI methods developed for communication with a computer in (written) natural language and to methods of common sense reasoning. Inductive logic programming will be introduced as a new perspective method extending significantly application possibilies of machine learning.
Machine learning - overview.
Selected symbolic methods of inductive learning - trees, rules.
Programming tools for AI and their requested properties. Principles of declarative programming languages.
Principles of Prolog. Logic programming and proof by resolution.
Implementations of state space search.
Review of Prolog solutions for characteristic AI problems. Typical applications.
New directions in logic programming. Constraint logic programming (CLP).
Utilisation of natural language communication within AI systems. Phases of natural languege processing.
Role of syntax and semantics of a sentence. DCG grammars. Design of a module for natural language communication.
Principles of ILP systems, review of properties of used training examples, practical applications of ILP.
Background knowledge and its role in inductive logic programming (ILP).
Computational learning theory.
Common sense reasoning and its partial automation. Naive physics. Qualitative simulation.
Russell, S., Norvig, P.: Artificial Intelligence, A Modern Approach. Prentice Hall Series in AI, Englewood Cliffs, New Jersey 1995.