Download - Corporate Semantic Web
11
Corporate Semantic Web
Acacia
http://www.inria.fr/acacia
INRIA Sophia Antipolis
22
Corporate Semantic Web ?
Use Semantic Web approach for Corporate Memory and Corporate Knowledge Management
33ObjectivesObjectives
Objectives
implement and trial a corporate memory management framework based on agents and
ontologies :
CoMMA : Corporate Memory Management with Agents
2 relevant scenarios have been chosen to highlight the problem of information retrieval in the company:
Enhancement of New Employee Insertion in the company,
Performing process that detect, identify and interpret technology movements for matching technology evolutions with market opportunities to disseminate among employees innovative ideas related to Technology Monitoring activities
44ObjectivesObjectives
ObjectivesCorporate knowledge management
aims at facilitating creation, dissemination,
transmission and reuse of knowledge in an
organisation
propose an innovative solution based on integration of technologies:
ontologies or knowledge models
multi-agent architecture of several co-operating agents
meta-information (resource annotation) expressed in RDF format
Machine Learning Techniques for user adaptability
55ObjectivesObjectives
CoMMA Objectives
Author of documents
User ModelsEnterprise Models
RDF ontology
Creator of ontology
Multi-Agent System
Employee
Corporate MemoryCorporate MemoryRDF annotations
(XML) Document
Indexation
KnowledgeKnowledge modelsmodels
.
MLT
Author agent
User agent
Group of interestagent
MLT
MLT
RDF annotations
(XML) Document
66CoMMA ConsortiumCoMMA Consortium
CoMMA ConsortiumEuropean IST project : 2000-2001
3 industrial partners:
Atos Origin (F)
CSTB (Centre Scientifique et Technique du Batiment) (F)
T-Systems Nova (G)
3 academic partners:
INRIA (F)
LIRMM/CNRS (F)
University of Parma (I)
77CoMMA : What is it ? CoMMA : What is it ?
Corporate Memory:Corporate Memory:
An explicit, disembodied and persistent An explicit, disembodied and persistent representation of knowledgerepresentation of knowledge and information in an and information in an organization, in order to facilitate their organization, in order to facilitate their access and access and reusereuse by members of the organization, for their by members of the organization, for their tasks.tasks.
88How ?How ?
How ?How ? Corporate memories are Corporate memories are heterogeneousheterogeneous
and distributed information landscapesand distributed information landscapes Stakeholders are an Stakeholders are an heterogeneous and distributed heterogeneous and distributed
populationpopulation Exploitation of CM involves Exploitation of CM involves heterogeneousheterogeneous
and distributed tasksand distributed tasks
Materialization CMMaterialization CM Exploitation CMExploitation CM
XML:XML: Standard, Structure, Standard, Structure, Extensible, Validate, Extensible, Validate, TransformTransform
RDF: RDF: Annotation, Annotation, Schemas Schemas
Multi-Agent System:Multi-Agent System: Modularity, Distributed, Modularity, Distributed, CollaborationCollaboration
Machine Learning :Machine Learning : Adaptation, EmergenceAdaptation, Emergence
CoCorporate rporate MMemoryemory MManagement through anagement through AAgentsgents
99Overall Schema & OntologyOverall Schema & Ontology
Corporate Memory
Multi-Agents SystemLearning
UserAgent
Learning
Interest GroupAgent
Ontology and Models Agent
UserAgent
Learning
InterconnectionAgent
KnowledgeEngineer
Author and/orannotator of documents
End User
Annotation
Document
Annotation
Document
Annotation
Document
Annotation
Document
Ontology
Models
- Enterprise Model - User's Profiles
Query
1010Example of problem: ambiguityExample of problem: ambiguity
The balance of our pharmaceutical project.The balance of our pharmaceutical project.
Two concepts & one term : ambiguityTwo concepts & one term : ambiguity Ontology :Ontology : object capturing relevant aspects of object capturing relevant aspects of
the meaning of concepts used in our application the meaning of concepts used in our application scenarios scenarios (example)(example)
1111
Use & UsersUse & Users
(1) Scenarios and Data collection(1) Scenarios and Data collection
Building the ontologyBuilding the ontology
Observations
&&
Internal
Interviews Documents
Reuse
External
ExternalExpertise
(Meta-)Dictionaries
Scenarios
(2) From semi-informal to semi-formal(2) From semi-informal to semi-formal
Relation Domain Range View SuperRelation
OtherTerms
Natural LanguageDefinition
Sy Tr Re Pr
Manage OrganizationalEntity;
OrganizationalEntity;
Organization Relation ; Relation denoting that anOrganizational Entity(Domain) is incharge/control of anotherOrganizational Entity(Range)
Tr EO
Created By Document; OrganizationalEntity;Person;
*; Relation ; Relation denoting that aDocument has beencreated by anOrganizational Entity
Us
Attribute Domain Range Type View SuperRelation
Other Terms Natural Language Definition Pr
Designation Thing; literal (string) *; ; ; Identifying word or words bywhich a thing is called andclassified or distinguished fromothers
Us
Family Name Person; literal (string) Person; Designation; Last Name;Surname
The name used to identify themembers of a family
Us
MobileNumber
Person; literal (string) Person; ; ; Mobile phone number Us
Title Document;
literal (string) Document;
Designation; ; Name of a document Us
Class View Superclass
OtherTerms
Natural Language Definition Pr
Thing Top-Level; ; ; Whatever exists animate, inanimate orabstraction.
Us
Event Top-Level;Event;
Thing; ; Thing taking place, happening,occurring; usually recognized asimportant, significant or unusual
Us
Gathering Event; Event; ; Event corresponding to the social act ofa group of Persons assembling in oneplace
Us
(3) RDF(S)(3) RDF(S)
Conceptual VocabularyConceptual Vocabulary
Relations - constraintsRelations - constraintsex: ex: personperson ( (authorauthor) ) documentdocument
Terms & natural language definitionsTerms & natural language definitionsex: 'bike', 'cycle', bicycle' - (ex: 'bike', 'cycle', bicycle' - (bicyclebicycle))
Concepts & links - definitionsConcepts & links - definitionsex: ex: documentdocument reportreport
(4) Navigation and Use(4) Navigation and Use
1212Memory StructureMemory Structure
Corporate Memory
Multi-Agents SystemLearning
UserAgent
Learning
Interest GroupAgent
Ontology and Models Agent
UserAgent
Learning
InterconnectionAgent
KnowledgeEngineer
Author and/orannotator of documents
End User
Annotation
Document
Annotation
Document
Annotation
Document
Annotation
Document
Ontology
Models
- Enterprise Model - User's Profiles
Query
1313Illustration of the cycleIllustration of the cycle
Organizational Entity (X) : Organizational Entity (X) : The entity X is or is a The entity X is or is a sub-part of an sub-part of an organization. organization.
Person (X): Person (X): The entity X is The entity X is living being pertaining to living being pertaining to the human race.the human race.
Include (Organizational Include (Organizational Entity: X, Organizational Entity: X, Organizational Entity / Person Y) :Entity / Person Y) :the organizational entity X the organizational entity X includes Y as one of its includes Y as one of its members.members.
Manage (Person: X, Manage (Person: X, Organizational Entity: Y) :Organizational Entity: Y) : The person X watches and The person X watches and directs the organizational directs the organizational entity Yentity Y
b - Ontologyb - Ontology
Person(Person(RoseRose))Person(Person(FabienFabien))Person(Person(OlivierOlivier))Person(Person(AlainAlain))
Organizational Organizational Entity(Entity(INRIAINRIA))
Organizational Organizational Entity(Entity(AcaciaAcacia))
Include(Include(INRIA, AcaciaINRIA, Acacia))
Manage(Manage(Rose, AcaciaRose, Acacia))
Include(Include(Acacia, RoseAcacia, Rose))Include(Include(Acacia, FabienAcacia, Fabien))Include(Include(Acacia, OlivierAcacia, Olivier))Include(Include(Acacia, AlainAcacia, Alain))
c - Situation &c - Situation &AnnotationsAnnotations
a - Realitya - Reality
1414Model-based Annotated MemoryModel-based Annotated Memory
Corporate Semantic WebCorporate Semantic WebRDF & RDFS : XML framework for Web RDF & RDFS : XML framework for Web
resources descriptions resources descriptions Use it for Intranets Use it for Intranets OOntology in RDFSntology in RDFS Description of the Description of the SSituation in RDF:ituation in RDF:
User ProfilesUser ProfilesOrganization modelOrganization model
AAnnotations in RDF describing nnotations in RDF describing DDocumentsocuments
OO SS
AADD
OO SS
AADD
Annotated ArchivesAnnotated Archives
ModelModelOO SS
AADD
++
++
++
++MemoryMemory
1515End-UsersEnd-Users
Corporate Memory
Multi-Agents SystemLearning
UserAgent
Learning
Interest GroupAgent
Ontology and Models Agent
UserAgent
Learning
InterconnectionAgent
KnowledgeEngineer
Author and/orannotator of documents
End User
Annotation
Document
Annotation
Document
Annotation
Document
Annotation
Document
Ontology
Models
- Enterprise Model - User's Profiles
Query
1616Interfacing UsersInterfacing Users
User InterfacesUser InterfacesAnnotating documentsAnnotating documentsQuerying the memoryQuerying the memoryHide complexity (ontology, agents,...)Hide complexity (ontology, agents,...)Present the resultsPresent the results
Push technologyPush technology Improve information flowingImprove information flowingProactive diffusion of annotationsProactive diffusion of annotationsCommunities of interestCommunities of interest
1717Profiles & LearningProfiles & Learning
Organizational modelOrganizational model Users' Profiles:Users' Profiles:
Administrative Information (link to Org. model)Administrative Information (link to Org. model)Explicit preferencesExplicit preferencesFavorite queries / annotationsFavorite queries / annotationsCharacteristics derived from past useCharacteristics derived from past use
Learning techniques:Learning techniques:Represent, learn and compare current use profiles Represent, learn and compare current use profiles to improve future use.to improve future use.Learning during a login sessionLearning during a login sessionRanking resultsRanking results
1818Multi-agent ArchitectureMulti-agent Architecture
Corporate Memory
Multi-Agents SystemLearning
UserAgent
Learning
Interest GroupAgent
Ontology and Models Agent
UserAgent
Learning
InterconnectionAgent
KnowledgeEngineer
Author and/orannotator of documents
End User
Annotation
Document
Annotation
Document
Annotation
Document
Annotation
Document
Ontology
Models
- Enterprise Model - User's Profiles
Query
1919Principal interest of MAS in CoMMAPrincipal interest of MAS in CoMMA
One functional architecture leading to several One functional architecture leading to several possible configurations in order to adapt to the possible configurations in order to adapt to the broad range of environments that can be found in broad range of environments that can be found in a companya companyArchitecture:Architecture: Agent kinds and their relationship Agent kinds and their relationship
Fixed at design timeFixed at design timeConfiguration:Configuration: Exact topography of a given MAS Exact topography of a given MAS
Fixed at deployment timeFixed at deployment time Flexible distribution :Flexible distribution :
Locally adapt to resources and usersLocally adapt to resources and usersGlobal capitalization through cooperationGlobal capitalization through cooperation
Integration of different technologiesIntegration of different technologies
2020Societies, Roles and InteractionsSocieties, Roles and Interactions
Ontology and Model SocietyOntology and Model Society
Ontologist AgentsOntologist Agents
Annotations SocietyAnnotations Society
MediatorsMediators
ArchivistsArchivists
Users' societyUsers' society
Profile Profile ManagersManagers
Profiles Profiles ArchivistsArchivists
InterfaceInterfaceControllersControllers
Interconnection SocietyInterconnection Society
FederatedFederatedMatchmakersMatchmakers
2121ConclusionConclusion
Corporate Memory
Multi-Agents SystemLearning
UserAgent
Learning
Interest GroupAgent
Ontology and Models Agent
UserAgent
Learning
InterconnectionAgent
KnowledgeEngineer
Author and/orannotator of documents
End User
Annotation
Document
Annotation
Document
Annotation
Document
Annotation
Document
Ontology
Models
- Enterprise Model - User's Profiles
Query
Done
Done
2222
TECHNOLOGY MONITORING
ANNOTATION
PUSH
Engineer (internal / informal
sources)
Archivist (external sources)
Docs+
An
notation
s
User
Area referent
CoordinationStrategic orientation
RETRIEVAL Query
Index card ,Synthesis,
The Technology Monitoring scenarioThe Technology Monitoring scenario
The diffusion of innovative ideas among employeesThe diffusion of innovative ideas among employees
Authors
2323
The actors of the Technology Monitoring scenario :The actors of the Technology Monitoring scenario :
Archivist Archivist in charge of feeding the systemin charge of feeding the system -> Author -> Author
Engineer and ResearcherEngineer and Researcherwatching his expertise Areawatching his expertise Area -> User -> Userfeeding the system with new informationfeeding the system with new information -> Author -> Authorin charge of identifyingin charge of identifying correspondents and coordinating correspondents and coordinating
thematic groupsthematic groups -> Area referent-> Area referent
The actorsThe actors
2424Examples of Supported tasksExamples of Supported tasks
For the For the AuthorsAuthors:: Indexing information by annotating companies, Indexing information by annotating companies,
people, documents...people, documents...
For the For the Area referentsArea referents:: Identifying resources, skills about given business Identifying resources, skills about given business
domainsdomains
For the For the UsersUsers: : Being automatically informed about relevant Being automatically informed about relevant
information according to their profile (push mode)information according to their profile (push mode)Querying the system (pull mode)Querying the system (pull mode)
2525
Search
Presentation
Question
FAQ
Help
Relation
Be Evaluated
tutor
Corporate Memory
Human resource
Updating profile
Newcomer
NEI Scenario: the “insertion of new employees” in the company NEI Scenario: the “insertion of new employees” in the company concerns the new employees who need to handle a lot of new concerns the new employees who need to handle a lot of new
information about their enterprise in a very short time, to be information about their enterprise in a very short time, to be rapidly efficientrapidly efficient
2626The actorsThe actors
The NE who just arrived in his new companyThe NE who just arrived in his new companynot familiar with the environmentnot familiar with the environmentneeding answers to many standard questionsneeding answers to many standard questions
The tutorThe tutorperson responsible to support NEs during the first person responsible to support NEs during the first
weeksweekswith CoMMA responsible to fill the annotation basewith CoMMA responsible to fill the annotation base
2727CoMMA solutionCoMMA solution
5 major components:5 major components: An ontology (O’CoMMA)An ontology (O’CoMMA) A multi-agent system,A multi-agent system, A Semantic search engine (CORESE),A Semantic search engine (CORESE), A machine learning algorithmA machine learning algorithm A GUIA GUI
The CoThe CoMMAMMA technical solution for the technical solution for the implementation of a Corporate memory.implementation of a Corporate memory.
The CoMMA Solution
2828CoMMA solutionCoMMA solution
splitting resources / system:splitting resources / system:
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
2929CoMMA solutionCoMMA solution
splitting resources / system:splitting resources / system: the document resourcesthe document resources
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
3030CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
splitting resources / system:splitting resources / system: the document resourcesthe document resources the configuration resourcesthe configuration resources
3131CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
splitting resources / system:splitting resources / system: the document resourcesthe document resources the configuration resourcesthe configuration resources
• OntologyOntology
3232CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Ontology O’CoMMAOntology O’CoMMA Dedicated to corporate memory,Dedicated to corporate memory, Represented in RDFS,Represented in RDFS,
3333CoMMA solutionCoMMA solution
rdfs:Class for concepts of the rdfs:Class for concepts of the ontology, ontology, Possibility to use class inheritancePossibility to use class inheritance
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Ontology
Engineer
Employee
rdfs:subClassOf
3434CoMMA solutionCoMMA solution
rdf:Property for relations of the ontology,rdf:Property for relations of the ontology,specialization of properties :specialization of properties :
director subPropertyOf managerdirector subPropertyOf manager
director director manager manager
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Ontology
isInterestedByEngineer
Employee
rdfs:subClassOf
Topic
3535CoMMA solutionCoMMA solution
rdfs:label for synonyms and multi- rdfs:label for synonyms and multi- language of the ontology,language of the ontology,
Use of stylesheet to filter terminology Use of stylesheet to filter terminology and multi-language.and multi-language.
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Ontology
Engineer
Employee
rdfs:subClassOfIngénieur
cadre technique
rdfs:label
3636CoMMA solutionCoMMA solution
rdfs:comment for natural language rdfs:comment for natural language definitiondefinition the link between definition and concept the link between definition and concept
is kept is kept ontology “tracontology “trackkability”ability”
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Ontology
Engineer
Employee
rdfs:subClassOf
Professionnel quitravaille dans unecertaine branche d
ingenierie ou l’on utilise laconnaissance
scientifique pourresoudre des
problemespratiques
rdfs:comment
Professional whoworks in some
branch ofengineering using
scientificknowledge to solvepractical problems.
rdfs:comment
3737RDFS Example : ClassRDFS Example : Class
<rdfs:Class rdf:ID="Document"><rdfs:Class rdf:ID="Document"> <rdfs:subClassOf rdf:resource="#Entity"/><rdfs:subClassOf rdf:resource="#Entity"/>
<rdfs:subClassOf rdf:resource="#EntityConcerningATopic"/><rdfs:subClassOf rdf:resource="#EntityConcerningATopic"/>
<rdfs:subClassOf rdf:resource="#NumberableEntity"/><rdfs:subClassOf rdf:resource="#NumberableEntity"/>
<rdfs:comment xml:lang="en">Entity including elements serving as <rdfs:comment xml:lang="en">Entity including elements serving as a representation of thinking.a representation of thinking.
</rdfs:comment></rdfs:comment>
<rdfs:comment xml:lang="fr">Entite comprenant des elements de <rdfs:comment xml:lang="fr">Entite comprenant des elements de representation de la pensee.representation de la pensee.
</rdfs:comment></rdfs:comment>
<rdfs:label xml:lang="en">document</rdfs:label><rdfs:label xml:lang="en">document</rdfs:label>
<rdfs:label xml:lang="fr">document</rdfs:label><rdfs:label xml:lang="fr">document</rdfs:label>
</rdfs:Class></rdfs:Class>
3838RDFS Example : PropertyRDFS Example : Property
<rdf:Property rdf:ID="Title"><rdf:Property rdf:ID="Title"><rdfs:subPropertyOf rdf:resource="#Designation"/><rdfs:subPropertyOf rdf:resource="#Designation"/>
<rdfs:range rdf:resource="&rdfs;Literal"/><rdfs:range rdf:resource="&rdfs;Literal"/>
<rdfs:domain rdf:resource="#Document"/><rdfs:domain rdf:resource="#Document"/>
<rdfs:comment xml:lang="en">Designation of a document.<rdfs:comment xml:lang="en">Designation of a document.
</rdfs:comment></rdfs:comment>
<rdfs:comment xml:lang="fr">Designation du document.<rdfs:comment xml:lang="fr">Designation du document.
</rdfs:comment></rdfs:comment>
<rdfs:label xml:lang="en">title</rdfs:label><rdfs:label xml:lang="en">title</rdfs:label>
<rdfs:label xml:lang="fr">titre</rdfs:label><rdfs:label xml:lang="fr">titre</rdfs:label>
</rdf:Property></rdf:Property>
3939CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
splitting resources / system:splitting resources / system: the document resourcesthe document resources the configuration resourcesthe configuration resources
• Ontology, Enterprise modelOntology, Enterprise model
4040Enterprise ModelEnterprise Model
<c:LegalCorporation rdf:about="http://www.inria.fr/"/><c:LegalCorporation rdf:about="http://www.inria.fr/"/>
<c:NationalOrganizationGroup rdf:about="http://www.inria.fr/"><c:NationalOrganizationGroup rdf:about="http://www.inria.fr/">
<c:Designation>Institut National de Recherche en Informatique <c:Designation>Institut National de Recherche en Informatique et Automatique</c:Designation>et Automatique</c:Designation>
<c:HasForActivity><c:Research/></c:HasForActivity><c:HasForActivity><c:Research/></c:HasForActivity>
<c:IsInterestedBy><c:ComputerScienceTopic/><c:IsInterestedBy><c:ComputerScienceTopic/></c:IsInterestedBy></c:IsInterestedBy>
<c:IsInterestedBy><c:MathematicsTopic/></c:IsInterestedBy><c:IsInterestedBy><c:MathematicsTopic/></c:IsInterestedBy>
……
4141
<c:LocalOrganizationGroup rdf:about="http://www-sop.inria.fr/"><c:LocalOrganizationGroup rdf:about="http://www-sop.inria.fr/">
<c:Designation>UR Sophia Antipolis de l'INRIA: Institut <c:Designation>UR Sophia Antipolis de l'INRIA: Institut National de Recherche en Informatique et National de Recherche en Informatique et Automatique</c:Designation>Automatique</c:Designation>
<c:HasForActivity><c:Research/></c:HasForActivity><c:HasForActivity><c:Research/></c:HasForActivity>
<c:IsInterestedBy><c:ComputerScienceTopic/><c:IsInterestedBy><c:ComputerScienceTopic/></c:IsInterestedBy></c:IsInterestedBy>
<c:Include><c:Include><c:ProjectGroup <c:ProjectGroup rdf:about="http://www.inria.fr/recherche/equipes/acacia.en.html"rdf:about="http://www.inria.fr/recherche/equipes/acacia.en.html"/></c:Include>/></c:Include>
<c:Include><c:Include><c:ProjectGroup rdf:about="http://www-sop.inria.fr/tropics/"/><c:ProjectGroup rdf:about="http://www-sop.inria.fr/tropics/"/></c:Include></c:Include>
<c:Include><c:Include><c:ProjectGroup rdf:about="http://www-sop.inria.fr/cafe/"/><c:ProjectGroup rdf:about="http://www-sop.inria.fr/cafe/"/></c:Include></c:Include>
4242CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
splitting resources / system:splitting resources / system: the document resourcesthe document resources the configuration resourcesthe configuration resources
• Ontology, Enterprise model, User profilesOntology, Enterprise model, User profiles
4343User Profile ExampleUser Profile Example
<c:IndividualProfile rdf:about="#"><c:IndividualProfile rdf:about="#"><c:CreationDate>an 2000</c:CreationDate><c:CreationDate>an 2000</c:CreationDate>
<c:Title>Employee profile of Olivier Corby</c:Title><c:Title>Employee profile of Olivier Corby</c:Title>
</c:IndividualProfile></c:IndividualProfile>
<c:Employee <c:Employee rdf:ID = "http://www-sop.inria.fr/acacia/personnel/corby/"> rdf:ID = "http://www-sop.inria.fr/acacia/personnel/corby/">
<c:FamilyName>Corby</c:FamilyName><c:FamilyName>Corby</c:FamilyName>
<c:FirstName>Olivier</c:FirstName><c:FirstName>Olivier</c:FirstName>
<c:HasForOntologicalEntrancePoint><c:KnowledgeModelingTo<c:HasForOntologicalEntrancePoint><c:KnowledgeModelingTopic/></c:HasForOntologicalEntrancePoint>pic/></c:HasForOntologicalEntrancePoint><c:HasForOntologicalEntrancePoint><c:ObjectProgrammingTo<c:HasForOntologicalEntrancePoint><c:ObjectProgrammingTopic/></c:HasForOntologicalEntrancePoint>pic/></c:HasForOntologicalEntrancePoint>
4444CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
splitting resources / system:splitting resources / system: the document resourcesthe document resources the configuration resourcesthe configuration resources the multi agent system frameworkthe multi agent system framework
4545CoMMA solutionCoMMA solution
Learning
User Agent
Gui: building an annotation.Gui: building an annotation.
4646
Learning
User Agent
CoMMA solutionCoMMA solution
Machine Learning technique:Machine Learning technique:
use feedbacks to learn document relevancyuse feedbacks to learn document relevancy
feedback from one user can be generalized feedback from one user can be generalized to users having the same fields of interest,to users having the same fields of interest,
is designed for both pull mode and push is designed for both pull mode and push modemode
4747CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
DocumentAgent(s)
Multi-agent system:Multi-agent system:document sub societydocument sub society
4848
DocumentAgent(s)
CoMMA solutionCoMMA solution
Multi-agent system:Multi-agent system:document sub societydocument sub society
CORESE a semantic search engineCORESE a semantic search engine relies on RDF(S) and conceptual relies on RDF(S) and conceptual
graph theory,graph theory, use of the inheritance graph of RDFS use of the inheritance graph of RDFS
(specialization and generalization),(specialization and generalization), Inference mechanismsInference mechanisms manage the annotation distributionmanage the annotation distribution Java API wrapped into an agentJava API wrapped into an agent
4949RDF AnnotationRDF Annotation
<c:ResearchReport <c:ResearchReport rdf:about='http://www.inria.fr/rapports/sophia/RR-3819.html'>rdf:about='http://www.inria.fr/rapports/sophia/RR-3819.html'>
<c:CreatedBy><c:CreatedBy>
<c:Person rdf:about='http://www.inria.fr/nada.matta'><c:Person rdf:about='http://www.inria.fr/nada.matta'>
<c:FamilyName>Matta</c:FamilyName><c:FamilyName>Matta</c:FamilyName>
<c:FirstName>Nada</c:FirstName><c:FirstName>Nada</c:FirstName>
</c:Person> </c:Person>
</c:CreatedBy></c:CreatedBy>
<c:CreatedBy><c:CreatedBy>
<c:Person rdf:about='http://www.inria.fr/olivier.corby'><c:Person rdf:about='http://www.inria.fr/olivier.corby'>
<c:FamilyName>Corby</c:FamilyName><c:FamilyName>Corby</c:FamilyName>
<c:FirstName>Olivier</c:FirstName><c:FirstName>Olivier</c:FirstName>
</c:Person> </c:Person>
</c:CreatedBy></c:CreatedBy>
5050RDF AnnotationRDF Annotation
<c:CreatedBy><c:CreatedBy>
<c:ProjectGroup rdf:about= <c:ProjectGroup rdf:about= 'http://www.inria.fr/recherche/equipes/acacia.en.html'>'http://www.inria.fr/recherche/equipes/acacia.en.html'>
<c:Designation>Acacia</c:Designation><c:Designation>Acacia</c:Designation>
<c:hasCreated rdf:resource=<c:hasCreated rdf:resource='http://www.inria.fr/rapports/sophia/RR-3819.html'/>'http://www.inria.fr/rapports/sophia/RR-3819.html'/>
</c:ProjectGroup> </c:ProjectGroup>
</c:CreatedBy></c:CreatedBy>
<c:CreationDate>11-1999</c:CreationDate><c:CreationDate>11-1999</c:CreationDate>
<c:Title> Méthodes de capitalisation de mémoire de projet <c:Title> Méthodes de capitalisation de mémoire de projet
</c:Title></c:Title>
5151CoMMA solutionCoMMA solution
Final user
ArchitectureArchitecture
Multi Agent system
Document authors and annotators.
Knowledge manager
Corporate Memory
User Agent
Learning
Annotation
Document
Ontology
Model Enterprise model
Annotation
Document
Annotation
Document
Annotation
Document
Request
Ontology and Model Agent
Learning
User Agent
Connecting Agent
Userprofile
Userprofile
Connecting Agent Document
Agent(s)
Learning
User Agent
Connecting Agent
Multi-agent system:Multi-agent system: Interconnecting sub societyInterconnecting sub society
5252CoMMA solutionCoMMA solution
Connecting Agent
Multi-agent system:Multi-agent system: Interconnecting sub societyInterconnecting sub society
A Distributed annotations management algorithmRelies on:• metrics that evaluate the semantic similiarity of annotations• complex protocols between « connecting agents » and « document agents » to rebuild the splitted annotation.
5353CoMMA solutionCoMMA solution
PR
EL
IMIN
AR
AN
AL
YSIS
SP
EC
IFIC
AT
ION
ANALYSIS OF INTERVIEWS
EXPLICITITATION OF ACTORS, ENTITIES,
RELATIONSHIPS, FUNCTIONALITIES
IMPLICIT WORLD OF THE COMPANY EXPLICIT WORLD OF THE COMPANY
COLLECTION OF DOCUMENT ABOUT THE COMPANY
ANALYSIS OF ORGANIZATION
INFORMAL MODEL OF THE
ENTERPRISE
FORMAL USER
MODEL
ONTOLOGIES ENTERPRISE MODEL
USER PROFILE
COMPLEMENTAR INTERVIEWS
OBJECT & AGENT WORLD
DOCUMENT TEMPLATES
XML, RDF,…
INTERVIEWS
REQUIREMENTS
FORMAL MODEL OF ENTERPRISE
Agent sub-societies
DE
SIGN
(UM
L)
Technical, Organisational, user
Data collection
DEPLOYMENT
Agent roles
Agent interactions
The CoMMA Methodology
5454Other project resultsOther project results
Other project results
• O ’CoMMA ontology • Extension of RDF(S) language for representing knowledge• CORESE + new inference mechanisms• Techniques of categorization of RDF-annotated documents• Multi-agent architecture for IR• RDF-based JADE ontology & content language • Management of distribution of annotations and of queries.• Machine learning techniques
5555
Specific Layer
High layer
Middle Layer
Ontology O ’CoMMAOntology O ’CoMMA
Aspects Aspects EnterpriseEnterprise
Aspects Aspects DocumentDocument
Aspects Aspects UserUser
Aspects Aspects DomainDomain
Method: Data collection, Terminological Phase , Method: Data collection, Terminological Phase , Structuration, Validation, Formalization in RDFSStructuration, Validation, Formalization in RDFS
Result: Result: 420 concepts, 50 relations, 630 terms, 420 concepts, 50 relations, 630 terms, 12 levels of depth12 levels of depth
5656
RDF Schema
RDF Annotations
RDF Query
CORESE search engineCORESE search engine
CGSupport
CGFact Base
Query Graph
Resultsin CG
Results in RDF
CORESE
Translation Projection Translation
5757Relation propertiesRelation properties
On
tolo
gie
Country Company Person
employs subdivisionOfactivityInanimate
Entity
nationality
domain
range
RD
FS
subClassOfrangedomain typesubPropertyOf
Property
Resource
Class
Literal
An
no
tatio n
RD
F www.T-Nova.de www.DeutscheTelekom.desubdivisionOf activity
Telecom
5858Relation propertiesRelation properties
RDF(S)
rdfs:Class
rdf:Property
domain, range
Resource
Property
RDF Annotation
CG
Concept Type
Relation Type
Signature
Concept
Relation
Conceptual Graph
RDF & CG
5959Relation propertiesRelation properties
Relation properties
Transitivity, Symmetry, Reflexivity, Inverse for RDF properties:
• Annotations are augmented with new knowledge deduced from these properties
• Transitivity, symmetry and inverse are computed once and added to annotations
• Reflexivity is computed on the fly according to queries
6060Inference Rules for RDFInference Rules for RDF
Inference Rules for RDF• Augment the ontology with rules that enable to deduce and add new knowledge to annotations
IF a team participates to a consortium AND a person is a member of the team
THEN the person participates to the consortium
IF [Person:?p]-(member)-[Team:?t]-(participates)-[Consortium:?c]THEN [Person:?p]-(participates)-[Consortium:?c]
• Forward chaining inference engine
6161Inference Rules for RDFInference Rules for RDF
RDF Rule Syntax<cos:rule><cos:if><c:Person rdf:about=‘?p’>
<c:member><c:Team rdf:about=‘?t’>
<c:participate><c:Consortium rdf:about=‘?c’/>
</c:participate></c:Team
</c:member></c:Person</cos:if><cos:then><c:Person rdf:about=‘?p’>
<c:participate><c:Consortium rdf:about=‘?c’/>
<c:participate></c:Person</cos:then></cos:rule>
6262ConclusionConclusion
Conclusion
The CoMMA system is implemented : A Corporate Semantic Webhttp://www.si.fr.atosorigin.com/sophia/comma
tested at T Nova Systems (Deutsche Telekom) and CSTB
testbed for Corporate Semantic Web technologies :
XML, Agents, Ontology, Semantic metadata, Learning
6363ConclusionConclusion
Conclusion (2)
Corese semantic engine : RDF(S) and Conceptual Graphs
tested at Renault on a design project memory
tested with the Gene Ontology