Principles of Knowledge Representation

Lectures: 12 h

Recommended for:
Foundational course for ECSEL.

Goals:
Knowledge representation is concerned with the systematic and formal description of real-world phenomena, with an emphasis on discrete-level descriptions of objects and processes with a complex structure. It relies on discrete mathematics (in the sense of elementary set theory, graphs, etc) and on formal logic as its conceptual tools.
The research area of knowledge representation has its roots in artificial intelligence research, but there is ample opportunity for interactions with several other fields, ranging from model-building in control engineering, via databases, to systems for human-machine interaction and natural language systems.
In addition, there is a core of common concepts which underlie the representational issues both in programming languages, databases, and knowledge systems.
The present course will present both the basic concepts of knowledge representation per se, and its connections to these neighboring areas.

Prerequisites:
MSc in a non-computer-science area and some programming experience.

Organization:
The course will include lectures, problem solving sessions, and computational exercises.

Contents:
The course consists of the following parts.

1. 1. Discrete structures. A brief recapitulation of the kinds of systems that one wishes to
   caracterize in formal knowledge representations, with examples.
   1. 2. First-order Predicate Logic. Syntax of logic formulae; their semantics; the concept of
      semantic entailment; inference systems.
   2. 3. Description of processes in first-order logic. Representations of time, persistence, indirect
      effects, delayed causation, etc.

Literature:
To be specified.

Examiner:
Erik Sandewall.

Credit:
3 credit points (for graduate students not having studied this material before).

Page responsible: Director of Graduate Studies