13 Ontologies and Knowledge-Based Systems

How do you go about representing knowledge about a world so it is easy to acquire, debug, maintain, communicate, share, and reason with? This chapter explores how to specify the meaning of symbols in intelligent agents, how to use the meaning for knowledge-based debugging and explanation, and, finally, how an agent can represent its own reasoning and how this may be used to build knowledge-based systems. As Smith points out in the quote above, the problems of ontology are central for building intelligent computational agents.

• 13.1 Knowledge Sharing
• 13.2 Flexible Representations
  • 13.2.1 Choosing Individuals and Relations
  • 13.2.2 Graphical Representations
  • 13.2.3 Primitive Versus Derived Relations
• 13.3 Ontologies and Knowledge Sharing
  • 13.3.1 Description Logic
  • 13.3.2 Top-Level Ontologies
• 13.4 Querying Users and Other Knowledge Sources
  • 13.4.1 Functional Relations
  • 13.4.2 More General Questions
• 13.5 Implementing Knowledge-Based Systems
  • 13.5.1 Base Languages and Metalanguages
  • 13.5.2 A Vanilla Meta-interpreter
  • 13.5.3 Expanding the Base Language
  • 13.5.4 Depth-Bounded Search
  • 13.5.5 Meta-interpreter to Build Proof Trees
  • 13.5.6 Ask the User Meta-interpreter
  • 13.5.7 Delaying Goals
• 13.6 Review
• 13.7 References and Further Reading
• 13.8 Exercises