creating integrated domain, task and competency model
DESCRIPTION
Luciano SerafiniTRANSCRIPT
Creating integrated domain, task and competency model
Luciano SerafiniFBKirst, Trento, Italy
Joint WORK with the partner of the APOSDLE EU PROJECT
Overview● Semantic technologies for organizations● Examples of conceptual models● The APOSDLE metamodel● Basic facts about ontologies● Proposal for a modeling activity
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
Semantic technologies enables...
● System interoperability● Modelbased systems engineering● Organizational memory● Knowledge management and reuse● Learning in work space
Basic reference architecture
M1
M3
M2
M4
Application
Application
Application
Application
Short introduction to APOSDLE learning platform
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
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
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
Integrated Modeling of Domain, Tasks and Learning
Goals
3rd Review Meeting, Graz
Short introduction on ontology engineering
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.
Example of ConceptualizationConceptual Graphs
Example of ConceptualizatioTopic maps
Example of conceptualizationSemantic networks
Example of conceptualizationRDFgaphs
Example of conceptualizationTaxonomic classification
Example of conceptualizationPartinomy
Example of conceptualizationWeb Ontology
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
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
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)
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, ...
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> … }
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
● 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)
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.
Collaborative Modelling
3rd Review Meeting, Graz
Collaborative Modeling…June 3, 2009
3rd Review Meeting, Graz
…with dedicated supporting tools
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.
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
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
– ...