Computer Ontology – Final Project Presentation

Rajesh Karunamurthy

Khalid Hassan Md.Mahmudur Rahman

Ali Kiani

Domain and Scope

Application domain : Personal Computer, specifically Hardware

Scope: Helping a person in buying a Computer and/or helping a Salesperson in a computer hardware shop to assist a Customer in buying a Computer

AssumptionsAssumptions : If user is ~( “Smart”) enough to choose a PC for his application he takes the help of a salesperson

Limitation of the Ontology

I don’t care whether my DMA chip is from INTEL or NEC …until it does its job

The project will not deal with the details of the manufacturing materials required for making the components, which make the Computer system

Scenario…C the center Dot..

Development Process

1. Research on Domain…become a Domain Expert2. Designing the Classes and its Hierarchy3. Defining the Slots and Facets of the slots4. Creation of Ontology using editor( OilEd)5. Verifying the Ontology ( Racer and FACT)6. GOTO STEP 2(or)3(or)4(or)5 and Refine …and refine

Research

Ontology development is moving from AI Labs to Desktop of Domain Experts (Knowledge System Lab of Stanford University)

Thanks…to – Designer of Largest Library in the World

SO……

Ontology Developer -> ~ ( Need not be a Domain expert)

Class Designing

Initial Design - One and Only one Hierarchy….Is it good…?…..NO Refinement….Ended in TWO main Hierarchies - Computer

component and Computer…with 290 classes… Many Small Hierarchies to support the main hierarchies …But Why?

… Types can not be designed using the existing editors

Class Hierarchy …Top Down Design (1/2)

Class Hierarchy …Bottom Up Design (2/2)

Slots and Facets (1/3)

From AI LiteratureSlots - Properties of

the concepts (Classes)

- Describes various features and attributes of the concept

Facets - Restrictions on slots

From Our Design Slots are defined only for

the main hierarchies Level of Abstraction

restricts the number of slots….still 30 slots

Facets - Better c this example - The IBMC laptop/IBMC desktop can have only IBMC OS and IBMC processor

Slots and Facets (2/3)

Slots and Facets (3/3)

Creation of Ontology

Creation … done by discussion Entering in the OilEd…..OilEd is simple and easy… Moreover it has

Activity log..others don’t support

Ontology Verification (1/2)

Verification- Are we Building the Ontology Right..?

Tools used: FACT, Racer and Rice(RACER interactive client environment) …..

In FACT When an Ontology is right..? If the activity log in OilEd gives no errors for subsumptions, satisfiablility, classes and instances …U GOT IT RIGHT..!!!

IS Our Ontology Right….click on the cat

Ontology Verification (2/2)

Last Words…..From [1]

1) There is no one correct way to model a domain— there are always viable alternatives. The best solution almost always depends on the application that you have in mind …SO Our Design is RIGHT

2)Ontology development is necessarily an iterative process…Yeah …Experience talks

3)Concepts in the ontology should be close to objects (physical or logical) and relationships in the domain of interest….. True ..THIS is reflected from our class hierarchy

Problems Faced …

Current Editors doesn’t support many functionalities….still this area is

OilEd .. doesn't support spaces in the names…No types like string or boolean or float or int is supported…SLOW…some times had problems even in saving the files…

Fact does not support AboxRacer is better….

Future Developments

Lots of things can be done…Developments depends on Goal

This ontology can be modified to support chip level details to the professionals seeking specific support…

The computer hardware can be isolated and can be extended for the companies developing hardware…

QUESTIONS…?