Top Banner
 Creating integrated domain, task and competency model Luciano Serafini FBK-irst, Trento, Italy Joint WORK with the partner of the APOSDLE EU PROJECT
33

Creating integrated domain, task and competency model

Nov 01, 2014

Download

Education

telss09

Luciano Serafini
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Creating integrated domain, task and competency model

   

Creating integrated domain, task and competency model

Luciano SerafiniFBK­irst, Trento, Italy

Joint WORK with the partner of the APOSDLE EU PROJECT

Page 2: Creating integrated domain, task and competency model

   

Overview● Semantic technologies for organizations● Examples of conceptual models● The APOSDLE meta­model● Basic facts about ontologies● Proposal for a modeling activity

Page 3: Creating integrated domain, task and competency model

   

In complex organizations

● Too many applications are being built with proprietary structures that are non­interoperable

● Many are busy mapping across islands using at best databases, and XML, but at worst documents and spreadsheets

● Semantic technology is a key enabler for realizing the renewed vision for integrating systems in complex organizaiton

Page 4: Creating integrated domain, task and competency model

   

Semantic technologies enables...

● System interoperability● Model­based systems engineering● Organizational memory● Knowledge management and reuse● Learning in work space

Page 5: Creating integrated domain, task and competency model

   

Basic reference architecture

M1

M3

M2

M4

Application

Application

Application

Application

Page 6: Creating integrated domain, task and competency model

   

Short introduction to APOSDLE learning platform

Page 7: Creating integrated domain, task and competency model

Third Annual Review Meeting, Graz

19 May, 2009 / 7

Integrated support for 

learner, knowledgeable person and worker

learning activities within work and learning processes

within computational work environment

utilizing organizational memory

 Self­Directed (SD) Work­Integrated Learning (WIL) 

APOSDLE Key Distinctions:Learning Perspective

Page 8: Creating integrated domain, task and competency model

Third Annual Review Meeting, Graz

19 May, 2009 / 8

APOSDLE Key Distinctions:Technological Perspective

● Hybrid Approach: Coarse grained semantic models complemented with soft computing approaches– Automatic discovery of work task/topic based on user interactions

– Automatic maintenance of user profiles based on user interactions 

– Automatic identification of similarities based on text, multi­media data and semantic analysis

– Automatic identification of prerequisite relations based on semantic analysis

Page 9: Creating integrated domain, task and competency model

Third Annual Review Meeting, Graz

Use Cases?

RE Process

UC ...

AlanUC Actor !

Sara

Database File­Server

LMS CMS ...

BackendSystems

SemanticStructures

IntegratedKnowledgeStructure

Domain­Model (Ontology) Process Model

Competency Performance 

Structure

Work Context

Skills

User Profiles Associative Network

LearningGoal Model

Process ModelDomain  Model(Ontology)

Working tools Learning tools

Collaboration tools 

OrganizationalIT-Infrastructure

APOSDLE Platform

APOSDLE Tools

Users

APOSDLE P3

Page 10: Creating integrated domain, task and competency model

Integrated Modeling of Domain, Tasks and Learning 

Goals

3rd Review Meeting, Graz

Page 11: Creating integrated domain, task and competency model

   

Short introduction on ontology engineering

Page 12: Creating integrated domain, task and competency model

   

The goal of conceptual modeling

● To construct a conceptualization of a domain that describes the aspects of a domain which are relevant to a certain (set of) application.

● What is a Conceptualization? It is a formal representation of a domain in terms of a set of Concepts and a set of Relations between concepts. 

Page 13: Creating integrated domain, task and competency model

   

Example of ConceptualizationConceptual Graphs

Page 14: Creating integrated domain, task and competency model

   

Example of ConceptualizatioTopic maps

Page 15: Creating integrated domain, task and competency model

   

Example of conceptualizationSemantic networks

Page 16: Creating integrated domain, task and competency model

   

Example of conceptualizationRDF­gaphs

Page 17: Creating integrated domain, task and competency model

   

Example of conceptualizationTaxonomic classification

Page 18: Creating integrated domain, task and competency model

   

Example of conceptualizationPartinomy

Page 19: Creating integrated domain, task and competency model

   

Example of conceptualizationWeb Ontology

Page 20: Creating integrated domain, task and competency model

   

Formal ontology

● A formal ontology is a special type of conceptualization based on logic

● ( + )( + ) Advantages of logic:

– It is an UNAMBIGUOUS language

– It is MACHINE UNDERSTANDABLE– It is possible to implement AUTOMATIC 

REASONING ALGORITHMS● ( – )( – ) Drawbacks of logic: it is NOT INTUITIVE for humans. 

– Difficulties to read logic

– Difficulties to formalize concepts in logic

Page 21: Creating integrated domain, task and competency model

   

Try yourself:

Elephants are gray  mammal which have a trunck

Elephant = Mammal ⊓ ∃ bodyPart.Trunk ⊓ ∀color.Gray

Elephants are heavy mammals, except for Dumbo elephants that are light

Elephant = Mammal ⊓

(∀weight.heavy ⊔ (Dumbo ⊓ ∀weight.Light)a

Page 22: Creating integrated domain, task and competency model

   

A four slide Introduction to ontologies

(1)The language of ontologies

(2)It's meaning 

(3)Expressing general knowledge (Tbox)

(4)Expressing specific knowledge (Abox)

Page 23: Creating integrated domain, task and competency model

   

Ontologies are formal theories based on a formal language:

● Basic components of the language of ontologies are 

● CONCEPTS (aka CLASSES, TYPES)– ANIMAL, FELINE, CAT, TAIL, ...

● RELATIONS (aka ROLES, ATTRIBUTES)– LOVES, IS_FRIEND_OF, LIVES_IN

● INDIVIDUALS (aka OBJECTS, CONSTANTS)– Garfield, John, Italy, France, ...

Page 24: Creating integrated domain, task and competency model

   

Formal languages has an unambiguous interpretation

● Every concept A is interpreted in a set.– CAT = (Fido, Garfield, Felix, cat1, cat2, ...}– TAIL = {tail­of­fido, tail­of­garfield, … }

● The elements of a concept are called Instances of the concepts– Fido, Garfield, ... are instances of the concept CAT, 

● Every relation R is interpreted in a set of pairsof instances– LOVES = {<john,mary> <paolo,elena>, 

<luciano,cecilia> … }

Page 25: Creating integrated domain, task and competency model

   

Axioms are statements in the formal language which holds on the 

domain we want to describe (1)

● A Subclass of B means that all the instances of A are also instances of B – CAT SubClass of ANIMAL means that each cat is 

also an animal

● R Subrelation of S = all the pairst in R are also contained in S

– IS_FRIEND_OF SubRelation KNOWS  means that it's not possible for two individuals to be friends without knowing eachother

Page 26: Creating integrated domain, task and competency model

   

● o ofType C (also written as C(o)) means that the object o is contained in the set of instances of C

– Italy ofType COUNTRY, means that COUNTRY = {.... italy … }

● o R o' (also written as R(o,o)) means that the object o is in relation R with the object o', I .e. that R = {… <o,o'> …} 

– Trentino is_pert_of Italy, means that trentino region is a part of the italian territory.

Axioms are statements in the formal language which holds on the 

domain we want to describe (1)

Page 27: Creating integrated domain, task and competency model

   

Ontology engineering

● Ontology engineering is the “art” of constructing useful, correct, compact and computationally sustainable conceptualizations in the form of formal ontologies. 

● Usually those who retain knowledge about a certain domain (domain experts) are not experts in logic and are not interested in becoming expert. 

● Usually experts in logic (knowledge engineers) have superficial and commonsense knowledge about a certain domain. 

● In ontology engineering domain experts and knowledge engineers need to collaborate to build useful and correct ontology based conceptualizations.

Page 28: Creating integrated domain, task and competency model

Collaborative Modelling

3rd Review Meeting, Graz

Page 29: Creating integrated domain, task and competency model

Collaborative Modeling…June 3, 2009

3rd Review Meeting, Graz

Page 30: Creating integrated domain, task and competency model

…with dedicated supporting tools 

Page 31: Creating integrated domain, task and competency model

   

Two collaborative tools for ontology  engineering

● Moki = Modelling WiKi  is a collaborative tool that provides support for enabling domain experts, who do not necessarily have knowledge engineering skills, to model business domains and simple processes directly.

● Collaborative Protege is an extension of the existing Protege system that supports collaborative ontology editing as well as annotation of both ontology components and ontology changes.

Page 32: Creating integrated domain, task and competency model

   

Proposal for a modeling experience● We constitute n modelling groups G(1) .... G(n)

● Mon­Tue

– G1..G(n/2) models moki producing models M1.. M(n/2)

– G(n/2+1)...Gn model with collaborative protege and produce models M(n/2+1) … M(n)

● Wed­Thu

– G1..G(n/2) revse the models M(n/2+1)...M(n) in collaborative protege 

– G(n/2+1)...Gn revise the models M(1)...M(n/2) in moki● Fri discussion on the experience and evaluation of the results

Page 33: Creating integrated domain, task and competency model

   

Domain model and task model for... Technology enhanced learning

● The resulting model should allow to represent 

– Classification of results, methodologies, scientific articles, and tools in the area of TAL

– Construction of a semantic social network in which people and organizations and activities are connected by common/complementary interests

● The resulting model could be used for

– Searching for results, paper, people, projects, possible collaborations

– Learning about TEL

– Keyword selection for semantic tagging 

– ...