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


12.2. PIF

As KIF is used to facilitate knowledge sharing, PIF (Process Interchange Format) is used to facilitate software process sharing (Lee et al., 1996). The PIF project started in October 1993 for a few groups at MIT, Stanford, the University of Toronto, and Digital Equipment Corporation to share heterogeneous software process descriptions. The goal is to develop an interlingua to automatically support exchange of heterogeneous business process descriptions within and among organizations.

The PIF-CORE provides the basic terminology for describing the simplest entities that can be used to describe the basic elements of any software process. Instead of having to write "ad hoc" translators for each pair of such systems, translators might translate any PIF-CORE description into and out of any target language with the minimum loss of information. Groups of users might extend the PIF-CORE descriptions using Partially Shared Views (PSV) to maximize information sharing among groups. A PSV module is built by specialization of the PIF-CORE or other PSV modules. In this case, translators only translate the additional PSV elements if they know about them. Since PIF is very expressive, translators do not always translate the whole PIF process into the chosen target language. The untranslatable parts are conserved and added when the process is translated back into PIF.

The PIF syntax is based on KIF syntax since: it allows the specification of classes, instances, values, and value restrictions in a structured way; the semantics of KIF reduces the ambiguity in the translation process between PIF and the target languages; it is possible to reuse some works already done in KIF; and Ontolingua translates standard KIF into other knowledge representation languages. More information about PIF requirements, PIF-CORE specification, the PSV mechanism for supporting multiple and partially overlapping class hierarchies, and translators can be found at http://soa.cba.hawaii.edu/pif/.

13. THE MOST WELL-KNOWN ONTOLOGIES

This section gives a brief overview of the most well-known ontologies. Many ontologies, like Ontolingua ontologies at Stanford (http://www-ksl.stanford.edu:5915) and WordNet at Princeton, are freely available over the Internet. A few, like some Cyc ontologies, are, in part, freely available. Others have been developed by companies for their own use and are not available over the Internet. The Ontology Page (http://www.medg.lcs.mit.edu/doyle/top/), also named TOP, identifies worldwide activity aimed at developing formalized ontologies.

This section seeks to summarize the most well-known ontologies, classifying them depending on the area of knowledge included. We will present: knowledge representation ontologies, like the Frame-Ontology in Ontolingua; linguistic ontologies, like GUM and EDR; engineering ontologies, like EngMath and PhysSys; and planning ontologies, like Multis.

13.1. CYC

The Cyc project was started at the Microelectronics and Computer Technology Corporation (MCC) in 1984. Cyc goals (Lenat and Guha, 1990) were to undo the software bottleneck by constructing a foundation of basic common-sense knowledge that would enable a variety of knowledge-intensive products and services. In January 1995, Cycorp (http://www.cyc.com/) improved and commercialized Cyc technology.

Although Cyc goals have remained unchanged over the years, Cyc technology has evolved substantially since then2. Actually, Cyc is a very large, multicontextual knowledge base with inference engines upon which different applications are built. Cyc technology consists of three main parts: the Cyc Knowledge Base (that is, the Cyc ontology), the CycL representation language and inference engines, and the knowledge server utility. The Cyc system is available in Common Lisp and C.


2Cyc's evolution is presented in Enabling agents to work together, in Communications of the ACM, 37(7), 127-142, 1994.

Cyc ontology provides a vast amount of fundamental human knowledge. The ontology consists of a set of terms and assertions that relate those terms. The Cyc ontology is divided into many microtheories. Each microtheory only captures a consistent point of view of a given domain of knowledge. Some areas can handle several different microtheories, representing different perspectives and assumptions, levels of granularity, and distinctions.

CycL is Cyc's knowledge representation language. CycL is a declarative and expressive language, similar to first-order predicate calculus with extensions to handle equality, default reasoning, skolemization, and some second-order features. CycL uses a form of circumscription, includes the unique names assumption, and can make use of the closed-world assumption where appropriate. The Cyc inference engine performs general logical deduction, best-first-search using a set of proprietary heuristics, uses microtheories to optimize inferences in restricting domains, and includes several special-purpose inferencing modules for handling specific classes of inferences.

The knowledge server utility allows multiple people to work together simultaneously on building up the KB; and provides a methodology and several user interfaces to browse, edit, and extend the Cyc KB, put queries to the inference engine, and interact with natural-language and database integration modules.

The applications areas currently available or under development are: natural language processing, integration of heterogeneous databases and data mining, knowledge-enhanced retrieval of captioned information, distributed AI, WWW information retrieval, smart interfaces, and more.

The most representative Cyc application is the Cyc-NL system. Given a query in English, the Cyc-NL system converts it into a CycL expression with free variables. Using CycL inference engines, Cyc gets a CycL answer and answers generating English from CycL statements. So, communication between Cyc and its users might be performed in English instead of by the Application Program Interface used in the 1990s. The Cyc-NL system is used for user-friendly interfaces to database tools and retrieval of information applications.

In data mining and integration of heterogeneous databases applications, Cyc incorporates the semantic-level knowledge. Cyc's database tool handles an NL or CycL query in three phases: the interface phase, which transforms NL queries into CycL expressions; the planner phase, which converts CycL expressions into an intermediate database representation format called CSQL, used to represent high-level logical database queries; and the executor phase, which converts the logical CSQL queries into physical SQL queries. After, the SQL answer is converted into CycL. This application was developed in 1995 to integrate data from tables at pharmaceutical companies.

With regard to knowledge-enhanced searching of captioned information, an image retrieval application and a text retrieval application were developed in 1994 and 1995 to optimize the search in extensive libraries of captioned images and in text documents. As before, once the target images or text documents have been described using Cyc, both systems, integrated with the Cyc-NL system, accept NL queries and make inferences with the image or text descriptions and other knowledge gathered in the Cyc KB.


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.