CISC 884: Knowledge Representation

Catalog Description:
Representation of knowledge in computers, common-sense knowledge, implementation methods: semantic nets, structured objects, logic, and production systems; specialized problems: representation of beliefs, temporal events and naive-physics concepts.

Current Texts:
Readings in Knowledge Representation
Ronald J. Brachman and Hector J. Levesque, eds.
Morgan Kaufmann Publishers, Inc., 1985

Goals:
To learn the major issues currently being researched in the area of knowledge representation. To become familiar with the major approaches and to attempt to apply them to the representation of knowledge of the commonsense world.

Content:

• Problems in knowledge representation:
  • theoretical
  • epistemological
  • implementational
• Role of logic in knowledge representation:
  • formal logic
  • common sense
• Representation of commonsense concepts:
  • naive physics
  • time
  • knowing and believing
• Major approaches to representing knowledge:
  • semantic networks - what's in a link?
  • frames - more fuzzy ideas
  • procedural representations - contrast with declarative representations
• Major unsolved problems:
  • default reasoning
  • handling exceptions
  • analogical reasoning

Required Background: Introductory course in artificial intelligence (CISC 681).

Helpful Background: A knowledge of first-order predicate logic.