Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


13.2. EXAMPLE OF A KNOWLEDGE REPRESENTATION ONTOLOGY: THE FRAME-ONTOLOGY

This ontology (Gruber, 1993a) plays an important role since:

  1. It is an ontology that captures the representation primitives most commonly used in frame-based representation languages. However, it does not seek to completely capture the semantics of existing knowledge representation languages. Examples of terms included in this ontology are: "Subclass-of," "Instance-of," "Minimum-Slot-Cardinality," etc.
  2. It is a knowledge representation ontology. It gives a set of second-order relations (relations that can take other relations as arguments) that allows ontologies to be codified using frame-based conventions. All Ontolingua ontologies use definitions from this ontology.
  3. It is the basis upon which Ontolingua translators are built, enabling people that use different knowledge representation languages to share ontologies.

The ontology (http://www-ksl.stanford.edu:5915) has been specified using KIF 3.0, and the entire ontology can be translated into pure KIF without information loss. The basic ontological commitments are that relations are tuples, functions are special cases of a relation in which the last term is unique for the previous terms, and classes are unary relations.

13.3. LINGUISTIC ONTOLOGIES

The Generalized Upper Model (GUM) is a general task and domain-independent linguistic ontology, developed by Bateman and colleagues at GMD/IPSI, Germany (Bateman, Magnini, and Fabris, 1995). To make it portable across different languages, the GUM ontology only includes the main linguistic concepts and how they are organized across languages, and omits details that differentiate languages. This philosophy allowed the use of GUM to create ontologies about specific languages, like Italian, German, and English, by entering the semantic distinctions of each language. Further extensions include languages closely related with the above (French and Spanish) and totally different languages (Chinese and Japanese).

The GUM ontology uses two hierarchies to model the domain. A conceptual taxonomy for concepts and a relational taxonomy for relations between concepts. The ontology provides detailed information about different kind of relations, processes, objects, etc. The ontology has been implemented in LOOM. A full description of both hierarchies can be found at http://www.darmstadt.gmd.de/publish/komet/gen-um/newUM.html. Actually, GUM is being used in different natural language processing applications (Penman, Komet, TechDoc, Alfresco, GIST, etc.), which demonstrates the high level of portability of this ontology.

The EDR Electronic Dictionary (Yokoi, 1995) is a very large KB of world knowledge, built upon the generalized electronic dictionary. The generalized electronic dictionary is an integrated collection of data and knowledge about the language to be used for natural language processing. It structures linguistic information along three axes: the description unit that includes superficial, conceptual, and deep knowledge; the descriptive unit that includes words, sentences, text, and documents; and the type of language unit, like Japanese, French, and English.

WordNet is a general-purpose concept lexicon that was developed at Princeton University. It can be used both as an online dictionary or thesaurus for reference purposes, and as a taxonomic lexical database.

13.4. ENGINEERING ONTOLOGIES

EngMath (Gruber and Olsen, 1994) is an Ontolingua ontology (http://www-ksl.stanford.edu:5915), developed for mathematical modeling in engineering. The ontology includes conceptual foundations for scalar, vector, and tensor quantities, physical dimensions, units of measure, functions of quantities, and dimensionless quantities. This ontology provides:

  1. Engineering models and domain theories represented in machine and human notation.
  2. A formal specification of a shared conceptualization and vocabulary for software agents in engineering domains. The EngMath ontology enables unambiguous communication between software agents in the SHADE project.
  3. A set of definitions to be reused for other engineering ontologies, for example, in the PhysSys ontology.

PhysSys ontology authors (Borst et al., 1996) propose incremental construction of ontologies in engineering domains by isolating types of knowledge in different ontologies, which are then used to build new ontologies by means of mapping ontologies. Mapping ontologies define interrelationships between ontologies.

PhysSys is a mapping ontology for modeling, simulating, and designing physical systems. The authors present three conceptual viewpoints of a physical system: system layout, physical processes underlying behavior, and descriptive mathematical relations. Three engineering ontologies formalize each of these viewpoints: a component ontology, a process ontology, and the EngMath (Gruber and Olsen, 1994) ontology.

There are two kinds of interrelationships between these three ontologies in PhysSys: interrelationships between the component and process ontologies, and interrelationships between the process ontology and the EngMath ontology. So, PhysSys proposes a new form of building and reusing ontologies.

13.5. PLANNING ONTOLOGIES

Multis (Mizoguchi, Vanwelkenhuysen, and Ikeda, 1995) is a task ontology consisting of generic nouns, generic verbs, generic adjectives, and other task-specific concepts. This ontology is used for scheduling tasks and developing a task analysis interview system that enables domain experts to build executable task models. It has been implemented in Ontolingua.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.