2006 05 09 cts2

Towards Collaborative Environments for Ontology Construction and Sharing: 

Towards Collaborative Environments for Ontology Construction and Sharing Jie Bao, Doina Caragea and Vasant Honavar Artificial Intelligence Research Laboratory Computer Science Department Iowa State University, Ames, IA USA 50011 Email: {baojie, dcaragea, honavar}@cs.iastate.edu May 15, 2006

Outline: 

Outline Motivation Package-based Description Logics: Language Features Package-based Description Logics : Semantics Collaborative Ontology Building Tools

Challenges in Ontology Building: 

Challenges in Ontology Building Collaboration Challenges Integration of local points of view Avoiding inconsistencies and unintended coupling Selective knowledge hiding Partial ontology reuse Scaleability Challenges Editing Storage Reasoning

Local vs Global Semantics: 

Local vs Global Semantics Ontologies represent local views of its producers Biologist: dog species only eats animal Ontology: Dog is Carnivore and all Carnivore only eats Animal Pet owner: pet dog sometimes eats DogFood, which is not animal Ontology: PetDog is Dog and some PetDog eats DogFood; DogFood is CannedFood and not Animal Global semantics could lead to conflicts Localizing knowledge is helpful to reduce such risks

Partial vs All-or-Nothing Reuse: 

Partial vs All-or-Nothing Reuse Lack of modularity: all or nothing Eg: how to import part of the animal ontology? Modular ontologies : more flexible partial reuse Less communication Less memory Less parsing time. Less unwanted junk!

Organizational vs Semantic Structure : 

Organizational vs Semantic Structure Organizational structure: how to arrange terms for better usage and understanding Eg: Computer Science Dictionary and Biology Dictionary Semantic structure: how to relate meanings of terms Eg: ‘Mouse’ is a kind of ‘Animal’ or ‘Mouse’ is part of ‘Computer’

Knowledge Hiding vs Sharing: 

Knowledge Hiding vs Sharing Ontology reflects shared knowledge in general However, the provider may also wish to hide part of it. Privacy, Copyright, Security Partial hiding helps for safer ontology organization Reduce unexpected coupling Separate “details” and “interface”

Ontology Languages Today: 

Ontology Languages Today Description Logics(DL), OWL, OBO (life science ontologies) However, the state of art in ontology languages is reminiscent of the early programming languages Uncontrolled use of global terms Unwanted and uncontrolled interactions between fragments Difficult to reuse: all or nothing

Ontology Languages Needed: 

Ontology Languages Needed Modularity Has localized terminology and semantics Allows partial ontology reuse Utilizes organizational and semantic structure Enables collaborative and scaleable tools Knowledge Hiding Builds safer ontologies Reduces unwanted interactions Hides details (encapsulate semantics)

Outline: 

Outline Motivation Package-based Description Logics: Language Features Package-based Description Logics: Semantics Collaborative Ontology Building Tools

P-DL: 

P-DL P3 protected 1. Whole ontology consists of a set of packages 2. Packages are organized in hierarchies 3. Terms and axioms are defined in packages with scope limitations P 1 P 2 public private P 1 P 2 public private

Package: 

Package A package is an ontology module with clearly defined access interface; Each package is defined with certain ontology language and Import: terms from other packages Interface: terms visible to other packages Each term has a home package 1. Whole ontology consists of a set of packages

Nested Package : 

Nested Package A nested package is part of another package Super package, sub package Form a package hierarchy Could be used to represent the organizational structure Arrange knowledge Enforce hierarchical management of knowledge 2. Packages are organized in hierarchies

Scope Limitation Modifier : 

Scope Limitation Modifier Defines the visible scope of a term or axiom SLM of an ontology term or axiom t is a boolean function V(t,r), where r is a package Package r could access t iff V(t,r) = True. Example SLMs Public (t,r): t is accessible from anywhere Private (t,r): t is only available in the home package Protected(t,r): t is accessible from the home package and its recursive sub packages. 3. Terms has scope limitation

SLM: example(TBC): 

SLM: example(TBC) A schedule ontology Hidden: details of the activity Visible: there is an activity Hidden semantics may still be used in reasoning

Outline: 

Outline Motivation Package-based Description Logics: Language Features Package-based Description Logics : Semantics Collaborative Ontology Building Tools

Local Interpretation: 

Local Interpretation Ontology: Carnivore Animal Interpretation: In any world that conforms to the ontology, for any instance x of Carnivore, x is also an instance of Animal.

Local and Global Interpretations: 

Local and Global Interpretations

Distributed Interpretation: 

Distributed Interpretation Global interpretations may not exist for all packages Distributed interpretations may still exist for selected sets of packages. P1,P2

Outline: 

Outline Motivation Package-based Description Logics : Language Features Package-based Description Logics : Semantics Collaborative Ontology Building

Collaborative Ontology Building: 

Collaborative Ontology Building Ontology modularity facilitates collaborative building Each package can be independently developed Different curators can concurrently edit the ontology on different packages Ontology can be only partially loaded Unwanted interactions are minimized by limiting term and axiom visibility Module access privileges can be controlled by the package hierarchy

The INDUS DAG Editor: 

The INDUS DAG Editor The INDUS DAG Editor

Summary: 

Summary Collaborative ontology building calls for modular ontology representation. Package-based description logics (P-DL) offers an ontology language for modularity and selective knowledge sharing. Efficient collaborative ontology building tools can be realized with P-DL. Ongoing Work Reasoning algorithm Extension to OWL

Slide24: 

Backup

Ontology Languages Today (2): 

Ontology Languages Today (2) Distributed Description Logics (DDL) Allows “bridge rules” between concepts across ontology modules E-Connections Connects DL modules with special types of roles called “links” Limitations Expressivity Semantic Soundness

Interpretation of Importing: 

Interpretation of Importing Domain relations are compositional consistent: r13=r12 O r23 Therefore domain relations are transitively reusable. Domain relation: individual correspondence between local domains Importing establishes one-to-one domain relations between local domains “Copied” individuals are shared between local domains Ensure exact reasoning w.r.t. the integrated ontology

Slide27: 

AnimalI CarnivoreI DogI goofyI fooI DogI PetI PetDogI plutoI eatsI (a) (b) 1 1 1 1 2 2 2 2 2 2 DogFoodI 2 AnimalI 2 Local Interpretation Semantics of foreign terms is not imported One term may have different local interpretations