[ CogSci Summaries home | UP | email ]

D. Lenat and R. Guha, Building Large Knowledge Based Systems: Representation and Inference in the Cyc Project. Addison-Wesley Publishing, 1990.

Author of the summary: J. William Murdock, 1997, murdock@cc.gatech.edu

Cite this paper for:

Keywords: Knowledge, Ontology

Systems: CYC

Summary: Chapter 1: Observes that current expert systems are
excessively brittle and incapable of effectively handling novel
situations.  Argues that general knowledge is needed to support more
flexible reasoning.  Discusses the importance of deep knowledge for
effective analogy.  Observes the problems in sharing of knowledge
between expert systems which use even subtly different ontologies.
Introduces the CYC project.  States that the set of primitives being
defined is converging.  Argues against natural language understanding
and machine learning as shortcuts to building knowledge bases.
States, however, an expectation that with sufficient knowledge,
natural language understanding will be possible to allow the system to
extend itself (by 1994).

Chapter 2: Presents the goals of the CYC project.  Discusses the user
interface for data entry.  Presents a language, CyCL, which uses a
frame-based syntax with a more powerful (but less tractable)
predicate-calculus constraint language for describing conceptually
challenging information.  Discusses implementation details.  Presents
some very general inferencing rules.  Discusses the constraint language
in some detail.

Summary author's notes:

Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies ( jim@jimdavies.org )
Last modified: Tue Mar 9 18:03:20 EST 1999