Representing Roles and Purposes James Fan 1, Ken Barker 1, Bruce Porter 1, Peter Clark 2 1 University of Texas at Austin 2 Boeing Company.

Slides:



Advertisements
Similar presentations
Entity Relationship Diagrams
Advertisements

Developing ER-Diagram
What is a domain model? “A domain model captures the most important types of objects in the context of the business. The domain model represents the ‘things’
The EDI Standard for the Belgian Insurance sector Modelling : some basic principles Classes Attributes of classes Associations from/to classes Generalizations.
Ch4: Software Architecture and Design. 1 Object-oriented paradigm  Object-oriented decomposition:  Agents comprised of two parts:  Hidden implementation:
Modeling Main issues: What do we want to build How do we write this down ©2008 John Wiley & Sons Ltd. vliet.
Building Knowledge Bases Compositionally Bruce Porter, Peter Clark Ken Barker, Art Souther, John Thompson James Fan, Dan Tecuci, Peter Yeh Marwan Elrakabawy,
Knowledge Representation. Essential to artificial intelligence are methods of representing knowledge. A number of methods have been developed, including:
7M822 UML Class Diagrams 7 October 2010.
Semantic Nets, Frames, World Representation. Knowledge Representation as a medium for human expression An intelligent system must have KRs that can be.
Albert Gatt LIN 1080 Semantics Lecture 13. In this lecture We take a look at argument structure and thematic roles these are the parts of the sentence.
Page 1 Integrating Multiple Data Sources using a Standardized XML Dictionary Ramon Lawrence Integrating Multiple Data Sources using a Standardized XML.
Chapter 14 (Web): Object-Oriented Data Modeling
Modeling Data The Entity Relationship Model (ER) For Database Design.
McGraw-Hill/Irwin Copyright © 2007 by The McGraw-Hill Companies, Inc. All rights reserved. Chapter 5 Understanding Entity Relationship Diagrams.
Integrating data sources on the World-Wide Web Ramon Lawrence and Ken Barker U. of Manitoba, U. of Calgary
Measurable Properties Name Blood pressure Weight Length Age Shoe size Temperature IQ EQ PERSON A person may have a large number of properties that are.
7M822 UML Class Diagrams advanced concepts 15 September 2008.
Lesson-19 Data Modeling and Analysis
Database Design Concepts Lecture 7 Introduction to E:R Modelling Identifying Entities.
Chapter 14: Object-Oriented Data Modeling
UML Class Diagrams: Basic Concepts. Objects –The purpose of class modeling is to describe objects. –An object is a concept, abstraction or thing that.
Description; compare-contrast; narrative; definition; opinion; cause- effect; classification; process.
Data Modeling Using the Entity-Relationship Model
the Entity-Relationship Model
Irwin/McGraw-Hill Copyright © 2000 The McGraw-Hill Companies. All Rights reserved Whitten Bentley DittmanSYSTEMS ANALYSIS AND DESIGN METHODS5th Edition.
Frames and semantic networks, page 1 CSI 4106, Winter 2005 A brief look at semantic networks A semantic network is an irregular graph that has concepts.
Knowledge Base Content Bruce Porter, Peter Clark Ken Barker, Art Souther, John Thompson James Fan, Dan Tecuci, Peter Yeh Marwan Elrakabawy, Sarah Tierney.
Knowledge Entry as the Graphical Assembly of Components Peter Clark, John Thompson (Boeing) Ken Barker, Bruce Porter (Univ Texas at Austin) Vinay Chaudhri,
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology A Taxonomy of Similarity Mechanisms for Case-Based Reasoning.
Database Design Principles – Lecture 3
For Wednesday Read chapter 13 Homework: –Chapter 10, exercise 5 (part 1 only, don’t redo) Progress for program 2 due.
©Silberschatz, Korth and Sudarshan2.1Database System Concepts Chapter 2: Entity-Relationship Model Entity Sets Relationship Sets Design Issues Mapping.
Definition of a taxonomy “System for naming and organizing things into groups that share similar characteristics” Taxonomy Architectures Applications.
Entity-Relationship Modeling Based on Chapter 12.
CS3773 Software Engineering Lecture 04 UML Class Diagram.
Objectives:1. Classes and Objects 2. Attributes 3. Services 4. Subjects Object-Oriented Analysis – Finding Class-&-Obects.
CS 405G: Introduction to Database Systems Lecture 2 : Database Design I.
CSC241 Object-Oriented Programming (OOP) Lecture No. 3.
Kal Bugrara, Ph.DSoftware Engineering Northeastern University Fundamentals Of Software Engineering Lecture V.
Design Model Lecture p6 T120B pavasario sem.
Semantic Nets, Frames, World Representation CS – W February, 2004.
Disambiguation Read J & M Chapter 17.1 – The Problem Washington Loses Appeal on Steel Duties Sue caught the bass with the new rod. Sue played the.
Object-Oriented Modeling: Static Models. Object-Oriented Modeling Model the system as interacting objects Model the system as interacting objects Match.
Knowledge Entry as the Graphical Assembly of Components Peter Clark, John Thompson (Boeing) Ken Barker, Bruce Porter (Univ Texas at Austin) Vinay Chaudhri,
Domain Classes – Part 1.  Analyze Requirements as per Use Case Model  Domain Model (Conceptual Class Diagram)  Interaction (Sequence) Diagrams  System.
Supertypes and Subtypes
Knowledge Representation. Keywordsquick way for agents to locate potentially useful information Thesaurimore structured approach than keywords, arranging.
Data Modeling Using the Entity-Relationship (ER) Data Model.
Course Review Classes 1-6 & Creating Motivating Learning
Chapter 8 Entity-Relationship Modeling Pearson Education © 2009.
Ontology and the lexicon Nicola Guarino and Christopher A. Welty(2004). An Overview of OntoClean Weber ( 張澄清 ) 2014/04/23 1.
David M. Kroenke and David J. Auer Database Processing Fundamentals, Design, and Implementation Appendix H: The Semantic Object Model.
Database Design, Application Development, and Administration, 6 th Edition Copyright © 2015 by Michael V. Mannino. All rights reserved. Chapter 5 Understanding.
Entity Relationship Diagram ERD. Background Peter Chen developed ERDs in Charles Bachman and James Martin have added some slight refinements to.
DOMAIN CLASSES – PART 1 BTS430 Systems Analysis and Design using UML.
Databases (CS507) CHAPTER 7.
Entity-Relationship Model
Entity-Relationship Model
Visit for more Learning Resources
Object-Oriented Analysis and Design
Entity-Relationship Model
The OO Solution The OO model closely resembles the problem domain
The OO Solution The OO model closely resembles the problem domain
UML Class Diagrams: Basic Concepts
Chapter 5 Advanced Data Modeling
Weak Slot-and-Filler Structures
Entity Relation Model Tingting Zhang.
Cheng (Sp2001): Object-Oriented Development Overview
Presentation transcript:

Representing Roles and Purposes James Fan 1, Ken Barker 1, Bruce Porter 1, Peter Clark 2 1 University of Texas at Austin 2 Boeing Company

Background on RKF Project Text book challenge problem One of the problems is how to deal with variety of types of knowledge, not just partonomy or taxonomy Some of them are obvious: entities and events Some of them are less obvious: entities and roles

A role representation consists of those features of an entity that are due to its participation in some event.

Roles and Purposes Examples: Fred is an employee at IBM. Fred is the gardener's employer. Grasshoppers are the favourite food of many birds. NADH is an important carrier of power in cells. This operator is a short region of regulatory DNA.

Intrinsic and Extrinsic Features Extrinsic features describe an entity relative to other entities and events.[Smith and Medin81], [Barr & Caplan87], for example, salary (of an employee), nutritional value (of a food). Intrinsic features describe an entity in isolation, for example, size (of an object), weight (of an object). Intrinsic features always applicable; extrinsic features may not be applicable.

Intrinsic and Extrinsic Features (Continue) Conclusions: The distinction is important. The representation of a role consists of those extrinsic features of an entity that are due to its participation in some event. This motivates reifying roles.

Foundedness and Semantic Rigidity (due to Guarino) Foundedness: Something is founded if it is defined in terms of relationships to other things. Semantic Rigidity: Something is semantically rigid if its existence is tied to its class; that is if in ceasing to be of kind X, it ceases to be.

Distinguishing Roles and Entities (due to Guarino) Foundedness Semantic Rigidity Roles Entities

Distinguishing Roles and Entities Food is Founded: The properties of food, such as eaten-by and nutritional-value, are extrinsic features of the entity filling the role of food. They relate that entity to others participating in the eating event, such as the eater, and they are applicable only in that context.

Distinguishing Roles and Entities [Guarino92] Food lacks Semantic Rigidity: For example, a grasshopper is food when eaten by a bird, but when it is no longer considered food, it is still a grasshopper.

Distinguishing Roles and Entities Person is not Founded: The properties of a person, such as age and sex, are intrinsic features. They are defined independently of other entities and events. Person has Semantic Rigidity: when a person ceases to be a person, she ceases to be.

Finding Roles Experiments: Extracted "-ee", "-er", "-or", "-ist" words from wordlist, sampled, and reviewed them using the foundedness and semantic rigidity criteria. 6% of nouns are roles. Reviewed the 3,000 most frequently used nouns in BNC. 6% of nouns are roles.

Examples friend, minister, staff, student, teacher, manager, wife, worker, director, leader, doctor, player, husband, customer etc..

Representational Requirements for Roles 1Role instances are created and destroyed dynamically, Fred’s job was eliminated. 2Roles can be transferred between entities, John got Bill's job. 3Multiple roles can be played by one entity, Fred is both an employer and an employee. 4Entities of unrelated types can play the same role, Fred works for IBM and Lisa.

Ways to Represent Roles[Steimann00] 1Simple label 2Inheritance 3Adjunct instance

I. Simple Label A role is just a label denoting a participant in an event; it is not reified. Example: Pro: simple. Con: Extrinsic features of the entities that are playing roles cannot be held by roles because roles are not reified. IBM: size: 50,000 agent-of: _Employ1 Fred: height: 5’10 object-of: _Employ1 employer-of: employee-of:

II. Inheritance Roles are reified subtypes of entities. Example. Person age: sex: height: … Employer NumOfEmployees: … subsumes

II. Inheritance Advantages: Roles are reified. An entity can play multiple roles through multiple inheritance. Roles are dynamically created/destroyed if a dynamic classification system is available. Disadvantages: Paradox[Steimann00] Fred EmployeeEmployer instance

Paradox Due to Inheritance But employers are not both person AND organizations! But Employer is not sibling of Person or Organization!

An Attempt to Salvage Inheritance But, not every person is an employer!

Lessons From the First Two Approaches Reifying roles allows extrinsic features to be associated with them. Classes of roles and entities should not be related by subsumption. Only instances of roles and entities are related.

III. Adjunct Instance Adjunct: something joined or added to another thing but not essentially a part of it. [Merriam-Webster] Instances of role types joined with instances of entity types. Example: Fred age:35 sex:male  _Employer2 NumOfEmployees:1  plays _Employ3TheGardener agent object event entity role

FOL of the Previous Example Person(Fred)  Employer(_Employer2)  Event(_Employ3)  Person(TheGardener)  agent-of(Fred, _Employ3)  plays(Fred, _Employer2)  object-of(TheGardener, _Employ3)

The Approach We Are Using 1.Based on adjunct instance approach. 2.Role hierarchy is separated from entity hierarchy. Event Role State Action Agent Instrument Be-Touching Move Operator Vehicle Entity Tangible Intangible Object Information

A Representation for Roles a Event an Entity a Role relation plays agent _Employ2 IBM _Employer3 plays _Employ4 Bill _Employer4 agent plays

Representing Purposes Using Roles Hammers are used to hit things. The purposes of hammers are: not an event not an entity but a way that an entity participates an event The shoe was used as a hammer.

A Representation for Purposes an Entity a Role an Event purpose in-event _Hammer1 a Hammer-Role a Hit purpose in-event _MyShoe a Hammer-Role plays Hammers are used to hit things. The shoe was used as a hammer

Duplication of Hierarchies

Non-reified Roles For all x such that isa(x, Hammer)  exists y, z such that isa(y, Instrument)  isa(z, Hammering)  purpose(x, y)  in-event(y, z) Exists p, h such that isa(myShoe, Shoe)  isa(h, Hammer)  plays(myShoe, p)  purpose(h, p)

Conclusion Roles are different from entities: Entities are things that are. Roles are things that are but only in the context of things that happen. A representation of roles needs to meet 4 requirements. Role instances are created and destroyed dynamically. Roles can be transferred between entities. Multiple roles can be played by one entity. Entities of unrelated types can play the same role.