OWLprojectace

DesigningCoreDesigningCoreOntologiesOntologiesAnsgarAnsgarScherpScherpCarstenCarstenSaathoffSaathoffThomasFranzThomasFranzSteffenSteffenStaabStaabNr.
5/2011Nr.
5/2011ArbeitsberichteausdemArbeitsberichteausdemFachbereichInformatikFachbereichInformatikDieArbeitsberichteausdemFachbereichInformatikdienenderDarstellungvorlufigerErgebnisse,dieinderRegelnochfürsptereVerffentlichungenüberarbeitetwerden.
DieAutorensinddeshalbfürkritischeHinweisedankbar.
AlleRechtevorbehalten,insbesonderediederbersetzung,desNachdruckes,desVortrags,derEntnahmevonAbbildungenundTabellen–auchbeinurauszugsweiserVerwertung.
The"ArbeitsberichteausdemFachbereichInformatik"comprisepreliminaryresultswhichwillusuallyberevisedforsubsequentpublication.
Criticalcommentsareappreciatedbytheauthors.
Allrightsreserved.
Nopartofthisreportmaybereproducedbyanymeansortranslated.
ArbeitsberichtedesFachbereichsInformatikISSN(Print):1864-0346ISSN(Online):1864-0850Herausgeber/Editedby:DerDekan:Prof.
Dr.
GrimmDieProfessorendesFachbereichs:Prof.
Dr.
Bátori,Prof.
Dr.
Burkhardt,Prof.
Dr.
Diller,Prof.
Dr.
Ebert,Prof.
Dr.
Furbach,Prof.
Dr.
Grimm,Prof.
Dr.
Hampe,Prof.
Dr.
Harbusch,jProf.
Dr.
Kilian,Prof.
Dr.
vonKorflesch,Prof.
Dr.
Lmmel,Prof.
Dr.
Lautenbach,Prof.
Dr.
Müller,Prof.
Dr.
Oppermann,Prof.
Dr.
Paulus,Prof.
Dr.
Priese,Prof.
Dr.
Rosendahl,Prof.
Dr.
Schubert,Prof.
Dr.
Staab,Prof.
Dr.
Steigner,Prof.
Dr.
Sure,Prof.
Dr.
Troitzsch,Prof.
Dr.
Walsh,Prof.
Dr.
Wimmer,Prof.
Dr.
ZbelKontaktdatenderVerfasserAnsgarScherp,CarstenSaathoff,ThomasFranz,SteffenStaabInstitutWeSTFachbereichInformatikUniversittKoblenz-LandauUniversittsstrae1D-56070KoblenzEMail:scherp@uni-koblenz.
de,saathoff@uni-koblenz.
de,franz@uni-koblenz.
de,staab@uni-koblenz.
deDesigningCoreOntologiesAnsgarScherp,CarstenSaathoff,ThomasFranz,andSteffenStaabWeST,UniversityofKoblenz-Landau,GermanyE-mail:{scherp|saathoff|franz|staab}@uni-koblenz.
deFebruary22,2011AbstractOneofthekeyfactorsthathindersintegrationofdistributed,heterogeneousinformationsystemsisthelackofaformalbasisformodelingthecomplex,struc-turedknowledgethatistobeexchanged.
Toalleviatethissituation,wepresentanapproachbasedoncoreontologies.
Coreontologiesarecharacterizedbyahighde-greeofaxiomatizationandformalprecision.
Thisisachievedbybasingonafoun-dationalontology.
Inaddition,coreontologiesshouldfollowapattern-orientedde-signapproach.
Bythis,theyaremodularandextensible.
Coreontologiesallowforreusingthestructuredknowledgetheydeneaswellasintegratingexistingdomainknowledge.
Thestructuredknowledgeofthecoreontologiesisclearlyseparatedfromthedomain-specicknowledge.
Suchcoreontologiesallowforbothformallyconceptualizetheirparticulareldsandtobeexiblycombinedtocovertheneedsofconcrete,complexapplicationdomains.
Overthelastyears,wehavedevelopedthreeindependentcoreontologiesforeventsandobjects,multimediaannotations,andpersonalinformationmanagement.
Inthispaper,wepresentthesimultaneoususeandintegrationofourcoreontologiesattheexampleofacomplex,distributedsocio-technicalsystemofemergencyresponse.
Wedescribeourdesignapproachforcoreontologiesanddiscussthelessonslearnedindesigningthem.
Finally,weelaborateonthebeautyaspectsofourcoreontologies.
1IntroductionDomainsthatrequiretheexchangeofahighamountofcomplex,structuredknowledgesuchasmedicalsystems,distributedmediamanagement,andemergencyresponsehaveahighpressureforsystemsintegrationinordertofacilitateefcientcommunicationandinformationexchange.
Forexample,inemergencyresponsedifferententitiessuchasanemergencyhotline,policedepartment,redepartment,andemergencycontrolcenterareinvolved.
Theseentitiesneedtoexchangeamongothersknowledgeaboutwhathappenedduringanincident,taskscommunicatedbetweentheentities,andmediainformationthatdocumentstheincident.
Duetothelackofappropriatelyintegratedsystemsatthedifferentemergencyresponseentities,thecomplex,structuredknowl-edgeiscurrentlyexchangedvianaturallanguageonthephone.
Thisisveryerror-proneandinefcient.
Rather,thedifferent,heterogeneoussystemsusedbytheemergencyre-sponseentitiesshouldbeintegratedtoprovideamoreefcientandeffectiveexchangeoftheknowledge.
Oneofthekeyfactorsthathindersintegrationofthesesystemsisthelackofaformalbasisformodelingthecomplex,structuredknowledgethatistobeexchanged.
Sofar,thisproblemhasnotbeensolvedduetothelackofnetworkedon-tologiesthatprovideaexiblemeanstomodelthecomplexstructureoftheknowledgeexchangedandatthesametimeprovideaformalsemanticstothatstructure.
Inthispaper,weproposeanapproachbasedoncoreontologiestoalleviatethissituation.
Anontologyallowsforformallyrepresentingtherelevantconceptsandre-lationsofaconsidereddomaininamachinereadableformat(Oberleetal.
,2009b;Oberle,2006).
Coreontologiesprovideaprecisedenitionofstructuralknowledgeinaspeciceldthatspansacrossdifferentapplicationdomains,e.
g.
,softwareservices,personalinformationmanagement,knowledgeorganization,multimediaannotations,andothers(Oberle,2006).
Theycombineanumberofspecicpropertiesthathavebeenderivedfromreportedexperiencesindesigningcoreontologies(Oberleetal.
,2007,2006;Oberle,2006)andthedevelopmentofourowncoreontologies(Arndtetal.
,2009;Scherpetal.
,2009a;Staabetal.
,2008;Arndtetal.
,2007;Franzetal.
,2007).
Thesepropertiesareaxiomatizationandformalprecision,modularity,extensi-bility,reuseability,andseparationofconcerns.
AxiomatizationandFormalPrecision.
Ahighdegreeofaxiomatizationandformalprecisionisprovidedbycoreontologies.
Bythis,acommonunderstandinginaparticulareldisestablishedinordertoensureinteroperabilitythroughmachineaccessiblesemantics.
Systemscanreasonabouttherepresentedknowledgeandcarryoutsemanticchecksonitsvalidity.
Theaxiomatizationandformalpreci-sionisachievedbybasingonafoundationalontology.
Modularity.
Coreontologiesshouldfollowapattern-orienteddesignapproach.
Bythis,theyaremodularwithintheeldforwhichtheyaredesigned.
Similartode-signpatternsinsoftwareengineering(Gamma,2007),ontologydesignpatternsprovideamodelingsolutiontoarecurrentontologydesignproblem(GangemiandPresutti,2009).
Thecoreontologyisacompositionofsuchontologyde-signpatternswithappropriatedependenciesbetweenthepatterns(GangemiandPresutti,2009).
Thisenablesapathwayforextensibilityandreuseability.
Extensibility.
Beingmodular,acoreontologyallowsforaddingnewandupdatingorremovingmodules,i.
e.
,ontologydesignpatternsitdenes.
Bythis,thecoreon-tologyisabletoreectsystemevolution(cf.
adaptabilityin(Vrandeˇcic,2009)).
Itisextensibletowardsnewdevelopmentsandfunctionalrequirementsthatarise.
Reuseability.
Differentsystemsarebuiltfordifferentpurposesandusersindifferentdomains.
Beingmodular,acoreontologysupportsreuseofitsmodules,i.
e.
,theontologydesignpatternsdespiteofthedifferentfocianddomainstheconcretesystemshave.
Atthesametimeacoreontologystillguaranteesformalprecisionoftheoverallknowledgeitrepresents.
Inadditiontothereuseofthedomain-independent,structuredknowledgedenedbythecoreontologiesthemselves,alsoexistingdomainknowledgecanbereused.
Coreontologiesareabletoin-corporateexistingdomainontologiesandmakeuseofthatdomainknowledgeratherthanrequiringtoremodelit.
SeparationofConcerns.
Thestructuralknowledgedenedinacoreontologyisclearlyseparatedfromthedomain-specicknowledge.
Thisallowscoreontolo-giestobeappliedinarbitraryapplicationdomains.
Domain-specicknowledgesuchasadomainontologyonemergencyresponseorsportscanbeintegratedandreusedwithoutaffectingthecoreontologyitself.
Inthispaper,wewillshowthatcombingthesepropertiesinacoreontologycanleadtoelegantsolutionsandinteroperabilityincomplexapplicationdomains.
Duetothecharacteristicsoftheirdesignsuchcoreontologiescanbeexiblycombinedtocovertheneedsofconcrete,complexapplicationdomains.
Thus,fromourperspectivetheyaretobeconsideredbeautifulontologies.
Overthelastyears,wehavedevelopedthreeofthosebeautifulcoreontologies.
Thesecoreontologieshavebeenusedoveralongtime,areverystablewithrespecttotheirdesign,andthusprovideasustainablesolutionforensuringinteroperabilityincomplexsocio-technicalsystemssuchasemergencyresponse.
ThesecoreontologiesaretheEvent-Model-F,COMM,andX-COSIMO.
ThecoreontologyEvent-Model-Fisdesignedformodelingeventsandobjects(Scherpetal.
,2009a).
Itallowsforrep-resentinghumanexperienceandparticipationinrealworldoccurrencesandprovidescomprehensivesupportformodelingtimeandspace,objectsandpersons,aswellasmereological,causal,andcorrelativeeventrelationshipsandeventinterpretations.
TheCoreOntologyonMultimedia(COMM)(Arndtetal.
,2009;Staabetal.
,2008;Arndtetal.
,2007)isdesignedfordescribingarbitrarydigitalmediadata.
Itallowsfor(se-mantic)annotationsofmediadataandtheirdecompositions.
Finally,theCross-ContextSemanticInformationManagementOntology(X-COSIMO)isdesignedforsemanticinformationmanagementandcommunication(Franzetal.
,2007).
Itsupportsmod-elingthecommunicationtakingplacebetweendifferentpersonsandsystemsandtheinformationassociatedwithitsuchastaskdescriptions.
Ourthreecoreontologies,i.
e.
,theEvent-Model-F,COMM,andX-COSIMOarebasedontheDescriptiveOntologyforLinguisticandCognitiveEngineering(DOLCE)(Gangemietal.
,2002;Masoloetal.
,2003).
Asfoundationalontology,DOLCEaimsatmodelingtheverybasicandgeneralconceptsandrelations(BorgoandMasolo,2009;Oberle,2006)thatmakeupourworld,e.
g.
,objects,events,partici-pation,andparthood.
Foundationalontologiesaregenericacrossmanyelds(Oberle,2006).
Theyhavealargescopeandarehighlyreusableindifferentmodelingsce-narios(BorgoandMasolo,2009).
Bytheirnature,foundationalontologiesaremuchbroaderthancoreontologiessuchasourEvent-Model-F,COMM,andX-COSIMO.
Coreontologiesprovidearenementtofoundationalontologiesbyaddingdetailedconceptsandrelationsintheirspeciceld.
DOLCEalreadyhasprovedtobeagoodmodelingbasisforcoreontologiessuchas(Scherpetal.
,2009a;Arndtetal.
,2007;Franzetal.
,2007;Oberleetal.
,2006,2007).
TheEvent-Model-F,COMM,andX-COSIMOhavebeencarefullyalignedwiththefoundationalontologyDOLCE+DnSUltralight1(DUL),alightweightversionofDOLCE.
Bythisalignment,ourcoreontologiescanbeexiblycombinedtocovertheneedsofcomplexapplicationdomains.
Ourcoreontologiesfollowapattern-orientedontologydesignapproach,i.
e.
,theydeneasetofontologydesignpatternstargetedforthespeciceldtheymodel.
ThesepatternsarebasedontheverygenericpatternsDULprovidessuchastheDescriptionsandSituations(DnS)patternandtheInforma-tionObject(IO)pattern(BorgoandMasolo,2009).
Byusingafoundationalontologyandfollowingapattern-orienteddesignapproach,thecoreontologiespossessasolid,semanticallyprecisebasis.
Atthesametimethesecoreontologiesbecomemodu-larandextensiblewithrespecttotheiruseinconcreteapplicationsandtochangesinfunctionalrequirements.
ByapplyingtheDnSpattern,ourcoreontologiesallowforaclearseparationofthestructuredknowledgecapturedbythecoreontologyandthedomainknowledgeprovidedbyadomainontology.
Thus,theyallowforintegratingandreusingexistingdomainontologies.
Theremainderofthepaperisorganizedasfollows:Inthenextsection,wemotivatetheneedforcoreontologiestomodelcomplex,structuredknowledgebypresentingascenarioofacomplex,socio-technicalsysteminthedomainofemergencyresponse.
InSection3,wedemonstratethesimultaneoususeandsmoothinterplayofourthreecoreontologiesEvent-Model-F,COMM,andX-COSIMOintheemergencyresponsescenario.
Itdemonstratestheuseofthethreecoreontologiestomodelthecomplex,structuredknowledgethatneedstobeexchangedbetweenthedifferentsystemsin-volved.
ThepropertiesofcoreontologiesandourdesignapproachfordevelopingsuchcoreontologiesarepresentedindetailinSection4.
TheconcretedesignofourthreecoreontologiestheEvent-Model-F,COMM,andX-COSIMOispresentedinSection5.
InSection6,wediscussthelessonslearnedwhendesigningandusingourcoreontolo-gies.
WeargueforthebeautyofourcoreontologiesinSection7,whichliesintheirabilitytobothformallyconceptualizetheirparticulareldsandtobeexiblycombinedtocovertheneedsofconcretescenarios,beforeweconcludethepaper.
2ModelingandSharingComplex,StructuredKnowl-edgeinEmergencyResponseIntheemergencyresponsescenariooftheEUprojectWeKnowIt2depictedinFigure1differentemergencyresponseentitiesareinvolvedusingdifferent,heterogeneoussys-tems.
Thesesystemsneedtoexchangecomplex,structuredknowledgethatneedstobesharedamongtheemergencyresponseentities.
Examplesofemergencyresponseentitiesaretheemergencyhotline,policedepartment,redepartment,emergencycon-trolcenter,andforwardliaisonofcers.
Theemergencycontrolcenterisinchargeofcoordinatingtheemergencyresponseentities.
Itreceiveseventdescriptionsfromtheemergencyhotline,processesthem,andcommunicateseventdescriptionswiththepo-licedepartmentandredepartment.
Inaddition,theemergencycontrolcenterforwards1http://ontologydesignpatterns.
org/wiki/Ontology:DOLCE+DnS_Ultralite2http://www.
weknowit.
eu/eventdescriptionstogetherwithtaskdescriptionstotheirforwardliaisonofcers.
For-wardliaisonofcersareoutintheeldtoreportaboutasituation,verifyingit,anddocumentingit,e.
g.
,bytakingphotosandnotes.
Asthefollowingscenarioshows,thissocio-technicalsystemforemergencyresponsebecomesveryactiveinthecaseofanincident.
Manydifferentpiecesofstructuredknowledgesuchasevents,tasks,andmultimediadatawithmetadataarecreated,combined,andcommunicatedbetweenthedifferentemergencyresponseentitiesinvolved.
Subsequently,wediscusshowourcoreontologiesareinvolvedinmodelingthiscomplex,structuredknowledge.
Figure1:Adistributedsystemforemergencyresponse2.
1ScenarioInanincidentofaheavystormamajoroodingmayhappen.
Duringtheoodingapoweroutageoccurs.
Somecitizensarelackingpowersupplyandarecallingtheemer-gencyhotlinetoreportaboutthepoweroutageevent.
Theofcersattheemergencyhotlinerecordthesecalls.
Inaddition,theytypeinatextualdescriptionforeachcalltodocumentthereportedeventintheirsystem.
Therecordedcallsareautomaticallyprocessedbysomespeechrecognitiontechniques,whichcreatesatranscriptofthecalls.
Thealgorithmalsoautomaticallydistinguishesthedifferentvoicesofthepar-ticipantsinthephonecallandcanautomaticallyassociateanddifferentiateutterancesmadebythecitizenandtheofcerattheemergencyhotline.
Subsequently,theeventdescriptiontogetherwiththeprocessedcallrecordinganditstranscriptareautomati-callytransferredtothesystemoftheemergencycontrolcenter.
Theemergencycontrolcenteralsoreceiveseventdescriptionsfromthesystemsofthepolicedepartmentandredepartmentthathappenduringtheincident.
Basedontheevidenceoftheeventdescriptions,theofcersintheemergencycontrolcenterusetheirsystemtoformu-latehypotheticaleventsthatmighthavecausedthepoweroutage.
Tothisend,eventdescriptionsareanalyzed,(semi-)automaticallyclustered,visualized,andputintore-lation.
Theofcersconcludethattherearetwopossibleinterpretationsthatmighthavecausedthepoweroutage,namelyasnappedpowerpoleclosetotheriverthatwasre-portedbyacitizencallingtheemergencyhotlineoraseriousproblemwiththepowerplantnearby.
Thecorrectassessmentofthesituationisverycrucialinordertomostef-fectivelydeploytheavailableemergencyresponseresources.
Thus,thedifferenteventinterpretationsmodeledinthesystemneedtobeveriedbytheofcersassoonaspos-sibleinordertoconrmorrejectthehypothesis.
Forthispurpose,theofcersintheemergencycontrolcentermaycontactthepersonnelofthepowerplant.
Atthesametime,thedescriptionofthehypotheticaleventofasnappedpowerpoletogetherwithataskdescriptionissenttothemobiledeviceofaforwardliaisonofcer.
Theforwardliaisonofcerreceivesthetaskdescription.
Shedrivestothelocationoftheeventstoverifyitandtodocumentitbytakingphotosandnotes.
Thephotoscanalsobetaggedorotherwiseannotated.
Theeventdescriptionextendedbyannotatedphotosaresendbacktothesystemoftheemergencycontrolcenter.
Asthescenarioshows,thedifferententitiesinvolvedinanemergencyresponsehavetosharestructuredknowledgeamongeachotherthroughthedifferentsystemstheyuse.
Thestructuredknowledgeisacombinationofeventdescriptionsthathappenduringanincident,e.
g.
,providedbythecitizenscallingtheemergencyhotlineoran-notatedbyforwardliaisonofcerswhenverifyingasituation.
Italsocomprisestaskdescriptionsthatarecommunicatedandshared,e.
g.
,theforwardliaisonofcersreceivetaskdescriptions(togetherwiththeeventdescription)toclarifyaspecicsituationatacertainplaceandtodocumentaboutitbytakingpicturesandnotes.
Finally,thesepicturesandnotestakenandthemetadatathatisattachedtothemediaisanotherkindofstructuredknowledgethatiscommunicatedwithinthedistributedsocio-technicalsystemforemergencyresponseshowninFigure1.
Asthediscussionshows,thestruc-turedknowledgethatistobecommunicatedisquitecomplex.
Inaddition,theentitiesinvolvedwiththisstructuredknowledge,i.
e.
,theevents,mediadata,andtasksappearandarerelevantindifferentcontextualsettingsthatneedtobemodeled.
Forexample,aneventreportedbyacitizentotheemergencyhotlineandrepresentedappropriatelymaybecomeanattachmentofamessagewithataskdescriptionthatissentbytheemergencycontrolcentertooneoftheirforwardliaisonofcers.
2.
2InvolvedOntologiesTomodelthecomplex,structuredknowledgeintheemergencyresponsescenario,i.
e.
,theevents,mediadata,andtasksweuseandcombineourthreecoreontologiesEvent-Model-F,COMM,andX-COSIMO.
Forrepresentingeventsandthemultiplerelation-shipsbetweenthem,weusetheEvent-Model-F(Scherpetal.
,2009a)thathasbeendevelopedintheWeKnowItproject.
TheEvent-Model-Fprovidesaformalrepresen-tationofthedifferentaspectsofeventsinwhichhumansparticipatesuchastimeandspace,composition,correlation,anddocumentation.
Comparedtoexistingmodels,theEvent-Model-Fdiffersinprovidingsophisticatedsupportformodelingcausality,correlation,andinterpretationofevents.
TheCoreOntologyonMultimedia(COMM)(Arndtetal.
,2009;Staabetal.
,2008;Arndtetal.
,2007)allowstorepresentarbitrarydigitalmediadatasuchasimages,videos,andaudio.
Itsupportsthedifferentkindsofannotationsofmediadataandtheirdecompositionintosegments.
COMMishighlyinuencedbyandspecicallydesignedtosupportthelow-leveldescriptorsofMPEG-7(MPEG-7,2001).
ItsrootsgobacktotheEUprojectaceMedia3,wherearstattempttomodelaMPEG-7ontol-ogyhasbeenundertaken.
ThisontologyisnotbasedonadirecttranslationofMPEG-7butonananalysisoftheMPEG-7standard.
Itfollowsaformalapproachformodel-ingthemultimediaannotationdomainbasedonDOLCE(Bloehdornetal.
,2005).
IntheEUNetworkofExcellenceK-Space4,thisideaofanalyzingMPEG-7inordertomodelaformalcoreontologyformultimediahasbeentakenupagain.
IncontrasttotheaceMediaapproach,theCOMMdevelopedinK-Spaceisfurtheraxiomatizedandbasedonontologydesignpatternsinordertoacquireaneasiertouseandformallymoresoundmodel.
Finally,theCross-ContextSemanticInformationManagementOntology(X-COSIMO)supportsformodelingsemanticinformationmanagementandcommuni-cation(Franzetal.
,2007).
X-COSIMOallowstorepresentthecommunicationtak-ingplacebetweendifferentpersonsandsystemsandtheinformationassociatedwiththiscommunicationlikeataskdescription.
ThecoreontologyhasbeendevelopedintheX-Mediaproject5,whichisdedicatedtoresearchonlargescaleandcross-mediaknowledgemanagementsolutions.
IntheX-MediaprojectalsoCOMMhasbeenusedandextended.
BothcoreontologiesCOMMandX-COSIMOplayakeyroleinthesharedrepresentationofautomaticallyextractedandnewlycreatedinformationintheX-Mediaproject.
Amongothers,thesharedrepresentationisexploitedinuserinter-facesthatenableuserstodealwiththediversityoftheknowledgerepresented.
2.
3SummaryWehaveshowntherequirementofsharingcomplex,structuredknowledgeattheex-ampleofasocio-technicalsystemforemergencyresponse.
Thestructuredknowledgetobemodeledandexchangedwithinthisscenarioaretherepresentationofeventsandobjects,multimediadataanditsannotations,andpersonalinformationmanagementsuchascommunicationandtasks.
Inthenextsection,wepresenthowsuchstructuredknowledgecanbemodeledwiththecoreontologiesEvent-Model-F,COMM,andX-COSIMOwehavedeveloped.
WedemonstratetheuseofourcoreontologiesattheexampleoftheemergencyresponsescenariooftheWeKnowItproject.
3ModelingtheEmergencyResponseScenarioReferringtothescenarioofthedistributedsocio-technicalsystemforemergencyre-sponsepresentedinSection2,weexemplifyinthissectionhowthestructuredknowl-edgethatisexchangedinthissystemcanbemodeled.
Wepickoutasmallpartofthescenarioandfullyrepresentitbyapplyingthecoreontologieswehavedeveloped,namelytheEvent-Model-Fformodelingeventsandobjects(Scherpetal.
,2009a),COMMforrepresentingmultimediaannotations(Arndtetal.
,2007),andX-COSIMOforpersonalinformationmanagementandcommunication(Franzetal.
,2007).
With3http://www.
acemedia.
org/4http://kspace.
qmul.
net/5http://www.
x-media-project.
org/modelingthescenario,weshowtheinterplayofthesecoreontologies,beforewedis-cussthepropertiesofcoreontologiesinSection4andtheirconcretedesigninSec-tion5.
Inthefollowing,weconsiderapoweroutagethathashappenedinthecourseofamajorooding.
ManycitizensarecallingtheemergencyresponsehotlinesuchasPaul.
Hecallstheemergencyhotlinetoreportaboutanobservationhemade,apowerpoleinhisstreethasjustsnapped.
Shortlyafter,thepoweroutagehappens.
Thus,Paulreportstothehotlinethathethinksthatthesnappedpowerpolehascausedthepoweroutage.
TheofcerRitaattheemergencyhotlineanswersPaul'scall.
ShetypesintohersystemwhatPaulreports,whilealsoanautomaticrecordingoftheconversationistaken.
Inourontologies,thecitizenPaulisrepresentedbytheindividualpaul-1andthepowerpoleisrepresentedbytheindividualpower-pole-1.
Wemodeltheeventwhenthepowerpolesnappedastheindividualsnapped-pp-1,theeventofthepoweroutageaspower-outage-1,andtheeventinwhichPaulcallsthehotlineascall-1.
TheofcerRitaworkingattheemergencyhotlineisrepresentedbytheindividualrita-1.
Sheisansweringthecall-1.
Theoveralloodingeventisreferencedasflooding-1.
UsingourEvent-Model-F,wemodeltheparticipationofthepersonpaul-1intheeventofasnapped-pp-1asshowninFigure2byapplyingthecoreontology'spar-ticipationpattern.
ThepatternisbasedonthegenericontologydesignpatternDescrip-tionsandSituations(DnS)(Gangemi,2008;GangemiandMika,2003).
TheontologydesignpatternDnSprovidesanontologicalformalizationofcontext(Oberle,2006;GangemiandMika,2003).
Itallowsforaformallypreciserepresentationofdifferent,contextualizedviewsbydeningroles.
Thus,besidesrepresentingtheparticipationofapersoninanevent,theparticipationpatternshowninFigure2alsodenesthatPaulplaystheroleofacitizeninthisparticipation,indicatedwithpaul-citizen-1whichisofconceptCitizenRole.
Inanothersituation,Paulmighthaveadifferentrole,e.
g.
,ifheisbesidesbeingacitizenalsoaprofessionalreghter.
Thus,PaulcanplaytheroleofaFiremanRoleinotherevents.
ItisimportanttonotethatboththeCitizenRoleandtheFiremanRolearenotdenedwithinthepartici-pationpatternoftheEvent-Model-F.
However,theyareprovidedfromsomeexternal,domain-specicontologies.
Thus,theparticipationpatternandtheEvent-Model-Fingeneralallowstoreuseexistingdomainknowledge.
TheuseoftheDnSpatterninanontologysuchastheEvent-Model-Fcanbeeasilyrecognized.
Italwaysdenesasituationthatsatisesadescription.
Thesituationincludestheeventsandobjectsofaconcretecontextualsituation,i.
e.
,thereal-worldentitiesthatcanbeobservedinaconcretesituation.
Thedescriptiondenestherolesoftheseeventsandobjectsintheobservedsituation.
Intheconcreteexample,thesituationpart-sit-snapped-pp-1isanEventParticipationSituationthatsatisfiesthedescriptionpart-desc-snapped-pp-1,whichisanEventParticipation-Description.
Theindividualdesc-ev-snapped-pp-1classiesthereal-worldeventofthesnappedpowerpolesnapped-pp-1,whichisofinterest,i.
e.
,describedintheconsideredsituation.
Inaddition,wecanmodelthetimeoftheeventandlocationofPaulwhenparticipatingintheevent.
Thisisnotshowninthegure,butavailableonlineasOWLontologyfromourontologieswebsite:http://west.
uni-koblenz.
de/Research/ontologies/.
Figure2:CitizenPaulparticipatingintheeventofasnappedpowerpoleThephonecallbetweenPaulandRitaisrecordedtodocumenttheevent.
Theinfor-mationthatthecallhasbeenrecordedisrepresentedbytheindividualaudio-rec-1.
Thisdocumentaryevidencethatthephonecalleventcall-1actuallyhappenedcanbemodeledusingthedocumentationpatternoftheEvent-Model-FasshowninFig-ure3.
Itrepresentsthedocumentationoftheeventcall-1byarecordingofthecallaudio-rec-1.
Theindividualaudio-rec-1isoftypeAudioData,whichisaconcepttakenfromtheCOMMcoreontology.
TheAudioDatarepresentsthein-formationthatisrealized,i.
e.
,containedintheaudiorecording.
However,itisnotarepresentationoftheaudioartifactsuchasadigitalmediastreamthatactuallyhasbeencapturedduringthecall.
Figure3:DocumentationofPaul'scalltotheemergencyhotlinewithanaudiorecord-ingInadditiontotheaudiorecording,alsoadescriptionofthesnapped-pp-1eventistypedintothecomputerbytheoperatorsattheemergencyhotline.
Thiseventde-scriptioncapturesdetailedinformationabouttheincident.
Forexample,inthecaseofaoodedcellartheeventdescriptionmaycontaindetailedinformationfortheemer-gencyresponseentitiesliketheredepartmentandprovidespecicinstructionsabout,e.
g.
,howtobestreachthecellar,size,waterlevel,andothers.
Inourexample,theeventdescriptioniscapturedbytheindividualtext-description-1,whichisofconceptTextData.
Itrepresentsatextualdescriptionofwherethesnappedpowerpoleislocatedandhowitcanbebestreached.
Thistextualdescriptiondocumentsthesnapped-pp-1eventasdepictedinFigure4usingtheEvent-Model-F'sdocumenta-tionpattern.
TheTextDataisaspecializationoftheconceptDigitalDatathatisprovidedbyCOMM.
ItisaspecializationofCOMMtowardsrepresentingtextualdataandisdenedintheOntologyforKnowledgeAcquisition(OAK)(Iria,2009).
ThespecializationofCOMMbyOAKisdiscussedindetailinSection6.
3.
Figure4:DocumentationofthesnappedpowerpoleeventwithtextualinstructionsfortheemergencyentitiesForthecourseofthismodelingexample,itdoesnotmakeadifferencewhetheroneconsidersAudioDataorTextDatatounderstandtheroleofmediadatafordocumentingeventsinemergencyresponse.
Inthefollowingweconcentrateontheaudiorecording,i.
e.
,theaudio-rec-1indicatingthattherewasarecordingtakenduringthecall.
Besidesitsmeredocumentationpurpose,therecordingofthephonecallaudio-rec-1betweenPaulandRitacanbereplayedbytheemergencyresponsepersonnelinordertolistenagaintotheconversation.
Inordertoprovideamoreef-cientaccesstotheinformationcommunicatedinthephonecall,theaudiorecordingaudio-rec-1isprocessedbyautomaticclassicationalgorithmssuchasasegmen-tationintosmaller,distinctpartsinwhicheitherPaulorRitaarespeaking.
Eachpartisautomaticallyannotatedwiththespeaker'snameusingautomaticclassicationmeth-ods.
Boththesegmentationandtheannotationwiththespeaker'snamearemodeledusingthecoreontologyCOMM.
Inordertoconductsuchasophisticatedannotationofanaudiorecording,werstneedtoprovideabasicrepresentationoftheaudiorecord-ingandthedigitalartifactcreatedforitasshowninFigure5.
AnunderlyingparadigmofCOMMisthedistinctionoftheinformationrepre-sentedbyanykindofmultimediaassetanditsdigital(orevennon-digital)real-ization.
TherecordingismodeledinCOMMasAudioData,asubconceptofInformationObjectandrepresentstheinformationthatexpressessomestateofaffairs(inDULmodeledasaDescription).
Thisinformationisclearlyseparatedfromtheactualrealizationofthatrecording.
Thesameinformationmightbestoredas.
wavor.
mp3le,orevenbestoredontape.
ThisseparationoftheinformationobjectanditsdigitalrealizationisbasedonthegenericontologyofInformationOb-jects(IO)(BorgoandMasolo,2009).
Themodeledinformation,i.
e.
,thefactthatthecallhasbeenrecordedanditsdigitalrealizationisbroughttogetherbytheindividualaudio-mp-1oftheconceptMediaProfile.
Theconceptofamediaproleorig-inatesfromthemultimediaannotationstandardMPEG-7(MPEG-7,2001).
Itallowsformodelingvariousmetadatasuchascolorhistograms,letype,andlelocation.
Thismetadatacanbecomplexincludingnestingofmetadata.
InCOMM,werepresentFigure5:RepresentationofanaudiorecordingandthelethatrealizestherecordingthisnestedmetadatastructurewithStructuredDataDescriptions(seedigitaldatapatterninSection5.
2.
1).
IntheexampleofFigure5,weseetwoofsuchStructuredDataDescriptions,namelyaudio-mide-1(aMediaInstanceDescriptor)andaudio-mlde-1(aMediaLocator-Descriptor).
Asshowninthegure,theindividualaudio-mida-1ofthetypeMediaInstanceDataexpressestheMediaInstanceDescriptorandisabouttheindividualaudio-mlda-1oftypeMediaLocatorData.
Theindividualaudio-mlda-1expressesaMediaLocatorDescriptor,representedbytheindividualaudio-mlde-1.
Thisaboutrelationbetweenaudio-mida-1andaudio-mlda-1modelsthenestingoftheMediaInstanceDescriptor,whichmight,amongothers,containaMediaLocatorDescriptorassubelement.
Thelocationofthedigitalrealizationinformofsomeaudioleisrepre-sentedasanUrinamedaudio-rec-file-1.
ThisUriservesasqualityoftheMediaLocatorDatarepresentedbyaudio-mlda-1asdiscussedabove.
Aqualityisalwayslocatedinsomeregionthatrepresentsthespaceofallpossiblevalues.
SincethelocationisidentiedbysomeUri,thecorrespondingregionisthespaceofallUris,representedbytheindividualus-1oftypeUriSpace.
Sincetheaudio-mlda-1expressestheMediaLocatorDescriptor,welinkthequalitytothedescriptionusingtheMediaUriParameter,whichparametrizestheUriSpace.
Theaudio-rec-1itselfisfurtherclassi-edasRootAudioSegmentRole,whichindicatesthattheinformationobjectaudio-rec-1referstothewholerecording,incontrasttootherAudioDatarep-resentingpartsofthewholerecording.
InCOMM,audiodataisalwaysoftypeAudioDataregardlessofwhetheritreferstoawholeorapart.
Thelatterisrep-resentedusingthedecompositionpattern.
AstherecordingcontainsthevoicesofbothPaulandRitaparticipatinginthecall-1,anautomaticspeakerchangedetectionisemployedtoautomati-callysegmenttherecordingintoseveralpartsthatrefertothedifferentspeak-ers.
Tomodelthetwovoices,weapplytheCOMMdecompositionpatterntoaudio-rec-1inordertorepresentthedifferentsegmentsoftherecording.
Asanexample,Figure6depictsthedecompositionofaudio-rec-1intothreeseg-mentsidentiedasaudio-segment-1toaudio-segment-3.
Thewholerecordingaudio-rec-1playstheInputSegmentRole,whilethepartsplayanAudioSegmentRoleinthispattern.
Pleasenotethatinarealphoneconversationatanemergencyhotlinetherewilltypicallybemorethanthreespeakerchanges,i.
e.
,morethanthreesegmentsdetected.
However,forthepurposeofdemonstrationitisal-readysufcienttoconsideronlythreesegments.
PleasenotefurtherthatintheCOMMdecompositionpattern,wedonotidentifytheconcretepersonspeakinginasegment.
Weonlyrepresentthenumberofdifferentsegmentsinthecall.
Theannotationoftheindividualsegmentswiththeconcretepersonspeakingisdoneonlyinthesubsequentstep.
Figure6:DecompositionofthecallrecordingintosegmentsToannotatetheindividualsegmentsoftheaudiorecordingwiththepersonspeak-ing,analgorithmforspeechdetectionisapplied.
Itleveragestheknowncharacter-isticsoftheofcer'svoiceRitatodistinguishherfromPaul'svoice.
Attheexam-pleofaudio-segment-2,Figure7showsthesemanticannotationofthesecondaudiosegmentwiththespeaker'svoice,paul-1.
FollowingtheDnSpattern,theSemanticAnnotationsatisesaMethod,whichrepresentsthesemi-automaticmethodaudio-speaker-method-1forassigningapersontoasegment.
Theaudio-speaker-method-1denesanAnnotatedDataRolewhichclassiestheAudioDatatobeannotated,inourexampleaudio-segment-2.
Itfurtherde-nesaSemanticLabelRole,whichclassiesthesemanticannotationlabelfortheaudiosegment.
Inthiscase,theindividualpaul-1representingthecitizenPaul.
Figure7:AnnotationofacallsegmentwiththespeakerPaulSimilar,alsothesegmentsoftheaudiorecordingareannotatedwithrita-1speaking.
Figure8showsattheexampleofaudio-segment-1theannotationofasegmentwithrita-1.
ThesystemattheemergencyhotlinehasitsownrepositoryfortheoperatorsatthehotlineandusesitsownconceptPersoninsteadofDUL'sconceptNaturalPerson.
Thisexampleshowsthatourcoreontologiescanreuseexistingdomainontologies.
ThisissueisdiscussedlaterinSection6.
4.
Pleasenotethatinthecasewheremultipledomainontologiesarereused,analignmentbetweentheseontologiesisnecessary.
Figure8:AnnotationofacallsegmentwiththespeakerRitaInhisphonecalltotheemergencyhotline,PaultellsRitathathethinksthattheeventofasnappedpowerpolerepresentedbytheindividualsnapped-pp-1hascausedtheeventofapower-outage-1thatshortlyafteroccurred.
ThissubjectivebutplausiblecausalrelationshipbetweeneventscanbemodeledusingthecausalitypatternoftheEvent-Model-FasindicatedinFigure9.
Thisisrepresentedbytworoles,oneclassifyingthecausingeventcause-snapped-pp-1andtheotherclassifyingtheeffecteventeffect-power-outage-1.
Thecausalitypatternalsoforeseestoattachajusticationtothecausalrelationship,herethelaws-of-physics-1.
Figure9:CausalrelationshipbetweenasnappedpowerpoleandapoweroutageInordertorepresentthatthecausalrelationshipmodeledaboveisPaul'sinter-pretationofhowthepower-outage-1happened,theEvent-Model-Fallowsforrepresentingdifferentcontextualviewsoneventsbyusingtheinterpretationpat-ternasshowninFigure10.
Weinterpretthepower-outage-1eventbyassem-blingthedifferentinstantiationsoftheEvent-Model-Fpatterns.
Theseinstantiationsofthepatternsareidentiedbytheindividualspart-sit-snapped-pp-1andcaus-sit-poweroutage-1fromFigure2andFigure9.
Figure10:RepresentingPaul'sinterpretationthatasnappedpowerpolehascausedthepoweroutageSubsequentlytothecallbetweenPaulandRita,thedescriptionofthecall-1eventcreatedbyRitaintheemergencyhotlinesystemandtheannotationoftheeventwiththeaudio-rec-1istransferredtotheemergencycontrolcenter.
TheemergencycontrolcenteralsoreceivesPaul'sinterpretationofthecauseforthepower-outage-1,namelytheeventofasnappedpowerpoleheobserved.
Attheemergencycontrolcenter,Henryrepresentedashenry-1worksasemergencycoordi-nationmanager.
Amongothers,hesupervisesandltersincomingeventdescriptions.
HereadstheeventdescriptionjustcreatedbyRitaandlistenstoherconversationwithPaul.
Asthesnappedpowerpolemightindeedhavecausedalargepoweroutageinthecity,Henrydecidestosendanurgentemailmessagerepresentedasmessage-1withataskdescriptiontask-1tohisforwardliaisonofcerMarie.
Henryattachestheeventinterpretationinter-sit-paul-1ofPaultothemessageaswellastheaudioleoftherecordedcallaudio-rec-1.
TheforwardliaisonofcerMarierep-resentedbymarie-1receivesthemessageanddrivestothelocationwherethepowerpolesnapped.
Mariechecksthepowercablesofthesnappedpoleandseesthatitonlyservesafewhousesintheneighborhoodwithelectricity.
Thus,sheconcludesthatthesnapped-pp-1eventcannothavecausedthelargepower-outage-1ofthecitythathappenedandthatPaul'sinterpretationislikelywrong.
Marietakesouthercameratotakeapictureofthesnappedpowerpoletodocumentherobservation.
Sheattachesthepicturetogetherwithsomemanuallyaddedtagsasdocumentarysupportforthesnapped-pp-1event.
TheresultsofMarie'sinvestigationaresendbacktoHenry,whodecidesthatMarie'sinterpretationofthesnapped-pp-1istherightoneandthatthepoweroutagemusthavebeencausedbysomeotherreason.
WithourcoreontologyX-COSIMO,werepresentcommunications,processes,andassociatedtasks.
Themessageexchangedbetweenhenry-1andhisforwardliai-sonofcermarie-1ismodeledusingthecommunicationpatternandisshowninFigure11.
Itprovidesacontextualviewoncommunicationwherehenry-1andmarie-1areplayingtherolessender-1andrecipient-1oftheemailmes-sage,respectively.
AttachedtothemessageisPaul'sinterpretationofthepoweroutageinter-sit-paul-1andtheaudio-recording-1isalsorepresented.
Thetask-1assignedtoMarieisofconceptInspectionTaskandisabouttoinspectthereportedincidentofasnappedpowerpole.
TheconceptInspectionTaskisdenedinadomainontologyfortasksinemergencyresponse,whichisreusedhere.
TheassignmentoftheInspectionTasktoMarieismodeledinX-COSIMOasshowninFigure12.
Theactioninspect-pp-1isassociatedtotask-1andassignedtomarie-1,whoisconsideredasthetaskownerwhileHenryisconsideredintheroleofaninformationprovider.
Severalinputstothetaskareavailableandaccordinglyrepresentedaswell:Paul'sinterpretationofthepowerout-ageinter-sit-paul-1,therecordedphonecallaudio-recording-1,andthemessage-1thatisdescribedabove.
Finally,fromtheviewpointofHenrybeingamemberoftheemergencycontrolcenter,alleventsthathappenarepartsofthemajorflooding-1event.
Forexam-ple,thepartsoftheflooding-1incidentaretheeventsofasnappedpowerpolesnapped-pp-1,apoweroutagepower-outage-1thatoccurred,phonecallstheemergencyhotlinereceivessuchascall-1,andmanyothers.
Thiscompositionofthelargeflooding-1eventbysmallereventsisdepictedinFigure13.
ItusesthemereologypatternoftheEvent-Model-F.
Thecompositeeventisflooding-1withitsthreeexamplecomponenteventssnapped-pp-1,power-outage-1,andcall-1.
ThetemporalconstraintsbywhichthecompositeeventiscomposedintothethreeandfurthersubeventsisnotshowninFigure13.
Itisimportanttonotethattheflooding-1isofconceptFlooding,whichisnotdenedwithinthemereologyFigure11:HenrysendsMarieamessagewithataskdescriptionpattern.
Rather,adomainontologyforemergencyresponsedevelopedbyapartnerintheWeKnowItprojectisbeingreusedherewithintheEvent-Model-F.
Bytheconcreteexampleofaoodingevent,wehaveshownhowthecomplex,structuredknowledgerequiredinourscenariocanbemodeled.
Multiplecoreontolo-giesareinvolvedandcombinedtorepresenttheknowledgesharedinthescenario,namelytheEvent-Model-F,X-COSIMO,andCOMM.
Theyaredesignedsuchthattheyallowforanoptimalinterplayandintegration.
Bythis,weallowforaneffec-tiveexchangeofemergencyresponseinformationbetweenthedifferentemergencyresponseentitiesinourscenarioandthesystemstheseentitiesuse.
Forthedesignofourcoreontologies,wefollowaspecicontologydesignapproach,whichwepresentinthefollowingsection.
4DesigningCoreOntologiesInordertoprovideaproperdesignapproachforcoreontologies,itisnecessarytoclarifywhatdistinguishesacoreontologyfromfoundationalontologiesanddomainontologies.
Thus,inSection4.
1,webrieydiscriminatetheconceptofcoreontologiesfromfoundationalontologiesanddomainontologies.
InSection4.
2,wedescribeourdesignapproachforbuildingcoreontologies.
EssentialchoiceforthisdesignapproachFigure12:DescriptionofthetasksenttoMarieFigure13:Modelingthelargeoodingeventbycompositionofsmallereventsaretheuseofafoundationalontologyasmodelingbasisandtheuseofontologyde-signpatterns.
TheapproachhasbeensuccessfullyappliedtodevelopourthreecoreontologiesEvent-Model-F,COMM,andX-COSIMO.
4.
1DenitionofFoundationalOntologies,CoreOntologies,andDomainOntologiesAnontologyisaspecialkindofinformationobjectthatallowsforformallyrepresent-ingtherelevantconceptsandrelationsofaconsidereddomaininamachinereadableformat(Oberleetal.
,2009b;Oberle,2006).
Thus,ontologiesareameanstoexplic-itlyspecifyconceptualmodelswithlogic-basedsemantics(Oberle,2006).
Inaddition,ontologiesareoftenreferredtohaveacollaborativeaspect(Oberleetal.
,2009b;Pintoetal.
,2009),i.
e.
,theformalconceptualizationshouldbeexpressedbyasharedviewandconsensusbetweenseveralparties.
Tobeasharedconceptualizationisofhighim-portanceforontologiesintendedtosupportlarge-scaleinteroperability(Oberleetal.
,2009b)suchasthecoreontologiesEvent-Model-F,COMM,andX-COSIMOusedintheemergencyscenarioinSection2.
Intheliterature,wenddifferenttypesandclas-sicationsofontologiessuchas(Oberle,2006;Gangemietal.
,2002).
Inthecontextofthiswork,wefollowthethree-layeredarchitectureofontologylibraries(Gangemietal.
,2004)anddiscriminatebetweenfoundationalontologies,coreontologies,anddomainontologies(Oberle,2006).
Coreontologiesmayusefoundationalontologiesaswellasleveragedomainontologies.
Thus,inthefollowingsections,webrieyan-alyzeanddenethenatureoffoundationalontologies,coreontologies,anddomainontologiesanddiscusstheirrelationtoeachother.
4.
1.
1FoundationalOntologiesFoundationalontologiesaregenericacrossmanyelds(Oberle,2006).
Theyhavealargescopeandarehighlyreusableindifferentmodelingscenarios(BorgoandMa-solo,2009).
Thus,foundationalontologiesservereferencepurposes(Oberle,2006)andaimatmodelingtheverybasicandgeneralconceptsandrelations(BorgoandMa-solo,2009;Oberle,2006)thatmakeupourworld,e.
g.
,objects,events,participation,andparthood.
Foundationalontologiesareheavyweightastheyarerichinaxioma-tization(BorgoandMasolo,2009),preciselydeningtheconceptsintheontologyandtheirrelations(Oberle,2006).
Synonymsofthetermfoundationalontologyaregenericontology,upperlevelontology,andtop-levelontology(EuzenatandShvaiko,2007;Oberle,2006).
ExamplesoffoundationalontologiesaretheABContologyandmodel(LagozeandHunter,2001),theBasicFormalOntology(BFO)6(Masoloetal.
,2003),DOLCE(BorgoandMasolo,2009;Gangemietal.
,2002;Masoloetal.
,2003),theObject-CenteredHigh-levelREferenceontology(OCHRE)(Schneider,2003),theGeneralFormalOntology(GFO)(Herreetal.
,2006),theOpenCycontology7(Lenatetal.
,1990),andtheSuggestedUpperMergedOntology(SUMO)(NilesandPease,2001).
Adetaileddiscussionofmostofthesefoundationalontologiescanbefoundin(Oberle,2006).
Heavyweightfoundationalontologiescanhavelightweightones(Oberle,2006),e.
g.
,DOLCEanditslightweightversiontheDOLCE+DnSUltralight(DUL)ontol-ogy8.
Themainpurposeofheavyweightfoundationalontologiesistoserveasref-erenceontologiesduringdevelopmenttime(Oberle,2006).
Alightweightversionofthefoundationalontologyisappliedtofacilitatereasoningatruntime(Oberle,2006).
Foundationalontologiescanbeusedasstartingpointforbuildingcoreontologiesanddomainontologies(Oberle,2006).
6http://www.
ifomis.
org/bfo7http://www.
cyc.
com/cyc/opencyc/8Availablefrom:http://ontologydesignpatterns.
org/wiki/Ontology:DOLCE+DnS_Ultralite4.
1.
2CoreOntologiesIncontrasttofoundationalontologiesthatspanacrossmanyeldsandmodeltheverybasicandgeneralconceptsandrelations(BorgoandMasolo,2009;Oberle,2006)thatmakeupourworld,coreontologiesprovideadetailedabstractdenitionofstructuredknowledgeinoneoftheseelds,e.
g.
,medicine,law,softwareservices,personalinfor-mationmanagement,multimediaannotations,andothers.
Bytheirnature,foundationalontologiesaremuchbroaderthancoreontologies.
Coreontologiescanbebasedonfoundationalontologiesandprovidearenementtofoundationalontologiesbyaddingdetailedconceptsandrelationsintheirspeciceld.
However,coreontologiesarestillverygenericandspanacrossasetofdomainsinaspeciceld(Oberle,2006).
ExamplesofcoreontologiesaretheMPEG7ontology9formultimediaanno-tations(Hunter,2005),acoreontologyforsoftwarecomponentsandwebser-vices(Oberleetal.
,2009a;Oberle,2006;Mikaetal.
,2004),ourcoreontologyEvent-Model-Fforcapturinghumanexperiencesintermsofeventsandobjects(Scherpetal.
,2009a),theX-COSIMOontologyforpersonalinformationmanagement(Franzetal.
,2007),andtheCoreOntologyforMultimedia(COMM)formodelingmultimediaan-notations(Arndtetal.
,2009;Staabetal.
,2008;Arndtetal.
,2007).
Coreontologiesaresituatedinbetweenthetwoextremesoffoundationalontolo-giesanddomainontologies(Oberle,2006),describednext.
Asfoundationalontologiesserveasagoodmodelingbasisforcoreontologies,sodocoreontologiesfordomainontologies.
4.
1.
3DomainOntologiesFinally,withdomainontologieswendrepresentationofknowledgethatisspecicforaparticulardomain(EuzenatandShvaiko,2007;Oberle,2006).
Domainontolo-giesusetermsinasensethatisrelevantonlytotheconsidereddomainandwhichisnotrelatedtosimilarconceptsinotherdomains(EuzenatandShvaiko,2007).
Do-mainontologiescanbeverycomplex,i.
e.
,theycancompriseaverylargenumberofconceptsandrelations.
Theycanmakeuseofandcanbebasedonfoundationalontolo-giesorcoreontologiesbyspecializingtheirconceptsinthedomainontology(Oberle,2006).
Domain-specicontologiescanbeusedasexternalsourcesofbackgroundknowledge(EuzenatandShvaiko,2007),e.
g.
,incombinationwithcoreontologies.
ExamplesofdomainontologiesareasoccerontologydevelopedintheSmartWebproject(Oberleetal.
,2007),theFoundationalModelofAnatomy10asadomain-specicmedicalontologydescribingtheanatomyofthehumanbody(RosseandMejino,2003),theRadLexLexiconforRadiology11(Kunduetal.
,2009),andothermedicalontologiessuchasSnomed(Coteetal.
,1993),theGeneOntology(Ashburner,2000),andGalen(RectorandHorrocks,1997).
9http://metadata.
net/mpeg7/10http://sig.
biostr.
washington.
edu/projects/fm/index.
html11http://www.
radlex.
org/Figure14:Ontologystackoffoundationalontologies,coreontologies,anddomainontologies4.
1.
4SummaryTherelationbetweenfoundationalontologies,coreontologies,anddomainontologiesisillustratedbytheontologystackinFigure4.
1.
4.
Thebottomofthegureshowsthefoundationalontologies.
Theymaybeusedandrenedbycoreontologiesinthemiddlelayer.
Thefurther,coreontologiesaswellasfoundationalontologiescanbeusedfordeningsemanticallyprecisedomainontologies.
Theborderlinefromcoreontologiestodomainontologiesisnotclearlydened.
Coreontologiesintendtobegenericwithinaeldthatspansacrossmultipledomains(Oberle,2006).
Similarly,alsothedistinctionbetweenfoundationalontologiesandcoreontologiesisnotclearlydened(Oberle,2006).
However,thedistinctionismeaningfulandusefulforbuildingontologylibrariesasfoundationalontologies,coreontologies,anddomainontologiesservedifferentpurposes(Gangemietal.
,2004).
4.
2DesignApproachforCoreOntologiesWehavedevelopedaguidelinedescribingthedesignapproachofcoreontologies.
Thisdesignapproachforcoreontologiesisdescribedinthefollowingsectionsalongthenon-functionalpropertiesidentied,namelyaxiomatizationandformalprecision,modularity,extensibility,reuseability,andseparationofconcerns.
4.
2.
1AxiomatizationandFormalPrecisionWhendesigninganontology,itisdesirabletouseasolidandsoundmodelingba-sis(Oberle,2006).
Thus,ourapproachforeseestheuseofafoundationalontologyfordesigningthecoreontology.
Wealigntheconceptsandrelationsdenedinthecoreontologytothebasiccategoriesofhumancognitioninvestigatedinphilosophy,linguistics,andpsychology(Oberle,2006).
Thesebasiccategoriesaremanifestedinthefoundationalontology.
Thealignmentofthefoundationalontologywiththecoreontologyalsoincludestheadoptionandspecializationoftheformalsemanticsofthefoundationalontologyinthecoreontology.
Whiletheaxiomatizationofafoundationalontologyenablesvalidatingtheupper-levelsemanticsoftheknowledgeexpressedwithit,thealignmentofthecoreontologywiththefoundationalontologyprovidessupportforvalidatingthemorespecicsemanticsoftheconceptsandrelationsdenedinthecoreontology.
Awell-designedfoundationalontologyisverydiligentwithrespecttotheontolog-icalchoicestowhichitcommits(Oberle,2006),e.
g.
,theselectionofthemostabstractconceptsthataremodeled.
Thus,whenreusingsuchafoundationalontologyforde-signingacoreontologytheengineerisalsorequestedtocarefullythinkabouthisorherdesignchoices(Oberle,2006).
Developingfoundationalontologiesisextremelyhardbutitneedstobeconductedonlyonce(Oberle,2006).
Anontologyengineershouldstriveforapplyingafoundationalontologythathasprovenitsusefulnesswhendesigninganewcoreontologyordomainontology(Oberle,2006).
Fordesigningourcoreontologies,wehavebasedthemonthefoundationalontol-ogyDOLCEandhavecarefullyalignedourcoreontologieswithDOLCE'slightweightversionDOLCE+DnSUltralight(DUL).
DOLCEaimsatcapturingtheontologicalcat-egoriesunderlyingnaturallanguageandhumancommonsense.
Ithasaminimalcorethatincludesonlythemostgeneralconceptsandpatternsandiswellsuitedformod-ularization.
AdditionaltheoriessuchasDescriptionsandSituations,OntologyonIn-formationObjects,ortheOntologyofPlanscanbeintegratedwhennecessary(Oberle,2006).
WechoseDOLCEasitalreadyhasproventoserveasgoodmodelingbasisforcoreontologies.
DOLCEhasbeensuccessfullyappliedtodesignontologiesindiffer-entdomainssuchaslaw,biomedicine,agriculture,andsoftwarecomponentsandwebservicesOberle(2006).
4.
2.
2ModularityEvenwelldesignedcoreontologiesareusuallylargeandcovermoreknowledgethanmightberequiredinaspecicapplicationdomain(GangemiandPresutti,2009).
Thus,concretesystemswillcommonlyuseonlyportionsofit.
Inthiscase,itishardtoreuseonlythe"useful"piecesofsuchamonolithiccoreontology(GangemiandPresutti,2009).
Coreontologies(anddomainontologies)haveabetterdesignwhenapplyingon-tologydesignpatternscapturedbythefoundationalontologyituses(Oberle,2006).
Byusingontologydesignpatterns,itallowsforselectingthepartsoftheontologyinaconcreteapplicationthatareactuallyneededandused(GangemiandPresutti,2009).
Thus,ourdesignapproachforcoreontologiesbuildsuponafoundationalontologythatsupportsapattern-orienteddesignapproach(Oberleetal.
,2006).
DULprovidessuchapattern-orientedapproach.
Ontologydesignpatternsshallnotbetoospecicortooclosetoaparticulardo-main.
Thiswoulddisallowtheapplicationofthepatterninotherdomainsoftheeldcoveredbythecoreontology.
Ontheotherhand,thepatternsshallalsonotbetoogenericasreuseinaconcretedomainwouldbehampered.
Atoogenericpatternishardtoapplyinaspecicdomain.
Thescopeofacoreontologyitselfisdenedthroughthescopeofitspatterns.
4.
2.
3ExtensibilityFoundationalontologiesprovideahigh-level,abstractvocabularyofconceptsandre-lationsthatarelikelytobeusedincurrentandfutureapplicationdomains.
Inordertoprovideasolidbasisforfutureextensionsofcoreontologies,aprecisealignmentoftheconceptsdenedinacoreontologywiththehigh-levelconceptsofafoundationalon-tologyisconductedusingourdesignapproach.
Bythisprecisealignment,newpatternscanbeaddedtothecoreontologywithoutaffectingtheexistingpatterns.
Inadditiontoaddingnewpatterns,theexistingpatternsofacoreontologycanbeextended.
Thisistypicallyconductedbyspecializingtheexistingconceptsandpropertiesdenedinthepatterns.
Finally,besidesthealreadyconnectedcoreontologieswithinaspecicapplicationalsofurthercoreontologiescanbedevelopedandintegratedifnecessary.
4.
2.
4ReuseabilityFormodelingcomplex,structuredknowledge,reusecanhappenondifferentlevels,e.
g.
,onthelevelofontologydesignpatterns,coreontologies,anddomain-specicontologies.
Differentpatternsinthecoreontologyprovidedifferentdescriptionsofconceptsdenedinit.
Bysplittingupcoreontologiesintodifferentpartstheyallowforreusingthestructuredknowledgedenedwithinthecoreontology'sdesignpat-ternsamongdifferentapplications.
ThisreferstotheissueofextensibilitydiscussedinSection4.
2.
3.
Inaddition,thecoreontologiescanbecombinedwithdomain-specicknowledge.
Intheidealcase,domainontologiesreusetheontologydesignpatternsdenedincoreontologiesbyspecializationsoftheontologydesignpatterns(GangemiandPresutti,2009).
However,ourapproachdoesnotrequirethatdomainontologiesarebasedonafoundationalontologyoracoreontology.
Infact,weexplicitlyconsiderbothoptionsasitcannotbeassumedthatalldomainontologiesarealignedwithafoundationalontologyorcoreontology.
ThisisachievedbyusingtheDescriptionsandSituationsontologydesignpatternofDOLCE.
Here,therolesdenedwithinacontextualsit-uationcanrefertoadomainontologythatiseithercarefullyalignedwithDOLCE,alignedtothecoreontology,orthatiscompletelyindependent.
4.
2.
5SeparationofConcernsOurdesignapproachsupportstheseparationofconcernsbydeningthestructuredknowledgeinthecoreontologyandleavingalldomain-specicaspectsoutofit.
ThisisachievedagainwiththeDescriptionsandSituationsontologydesignpattern.
Thestructuredknowledgeoftheconcreteeldofthecoreontologyiscapturedbyitson-tologydesignpatterns,e.
g.
,participation,causality,anddocumentationfortheEvent-Model-F,theannotationanddecompositioninCOMM,andthecommunicationpatterninX-COSIMO.
Thisstructuredknowledgeisspeciedbytherolesdenedwithinthesepatterns,i.
e.
,thedefinesrelationsoftheDescriptions.
Thedomainknowledgeisonlyreferredtobytherolesclassifyingtheeventsandobjectsused.
Bythis,thecoreontologiesareindependentofanyconcretedomainthatmakesuseoftheconceptsdenedbythem.
Inaddition,coreontologiesprovidesupporttoincludeindividualsdenedinsomedomainontologies.
5ExamplesofCoreOntologiesInthissection,wedescribethedesignofourcoreontologiesEvent-Model-F,COMM,andX-COSIMOwithfocusonthepartsrelevanttomodeltheemergencyresponsescenarioinSection3.
OurcoreontologiesarebasedonthefoundationalontologyDOLCE+DnSUltralight(DUL).
DULdenestheclassDUL:Eventnexttothedis-jointupperclassesDUL:Object,DUL:Abstract,andDUL:Quality.
Thedef-initionofEventhasbeenspecializedfromtheformaldenitioninDOLCEasanentitythatexistsintime.
TheclassObjectstandsforentitiesthatexistinspacesuchaslivingthingsaswellasnon-livingandsocialandcognitiveentities.
AQuality12isacharacteristicofanobjectoranevent.
Ithasavaluethatisrepre-sentedasapointorareainsomeAbstract.
TheclassAbstractrepresentsvaluespaces,e.
g.
,thespaceofnaturalnumbersorthetimeofaday.
Typically,wedonotprescribespecicAbstractsthataretobeused.
WeratherrefertothegenericAbstractsalreadydenedinDULsuchastheregionsDUL:TimeInterval,DUL:SpatioTemporalRegion,andDUL:SpaceRegion.
Formodelingourcoreontologies,wemakeuseoftheontologydesignpatternsDescriptionsandSituations(DnS)andtheontologyofInformationObject(IO)(BorgoandMasolo,2009).
TheDnSpatternprovidesanontologicalformalizationofcon-text(Oberle,2006;GangemiandMika,2003).
WithDnSonecanreifyEventsandObjectsanddescribethen-aryrelationthatexistsbetweenmultipleindividualsofthem.
Thus,itallowsforaformallypreciserepresentationofdifferent,contextualizedviewsonevents.
TheIOontologypatterndescribestherelationbetweenaninforma-tionobjectsuchasapoem,song,andastoryandtheiractualphysicalrealizationinformofaprintedbook,recordedtrack,andamovietaken(Oberle,2006).
WedescribethedesignofourthreecoreontologiestheEvent-Model-F,COMM,andX-COSIMOinthefollowingsections.
5.
1Event-Model-F—CoreOntologyofEventsforRepresentingHumanExperienceThecoreontologyEvent-Model-Fforrepresentinghumanexperienceallowsformod-elingthedifferentrelationshipsbetweeneventsandobjects.
Therequirementstothecoreontologyforeventshavebeenderivedfromexistingeventmodelsinvariousdo-mainssuchasmusic,journalism,multimedia,news,culturalheritage,andknowledgerepresentation(Wangetal.
,2007;RaimondandAbdallah,2007;IPTC,2008;Doerretal.
,2007;Mueller,2008;Francoisetal.
,2005;Jain,2008;Ekinetal.
,2004).
Identi-edrequirementsarerepresenting(1)participationoflivingandnon-livingobjectsinevents,(2)temporaldurationofevents,and(3)spatialextensionofobjects.
Inaddition,threekindofeventrelationshipsshallbesupported,namely(4a)mereological(com-positionofevents),(4b)causal,and(4c)correlation.
Thecommonmodelshallalsosupporttheexperientialaspect,i.
e.
,the(5)annotationofeventswithsensordatasuchasmediadata,andallowfor(6)differentinterpretationsofevents.
Existingmodelsalmostfullysupportparticipation,timeandspace,andtheexperientialaspect.
How-12Alsocalledtrope,seehttp://plato.
stanford.
edu/entries/tropes/ever,theysubstantiallylackinthemereological,causal,andcorrelationrelationships,andeventinterpretations.
Here,wenddifferentlimitationsorevennosupport,e.
g.
,onlysimplemereologicalrelationshipsin(RaimondandAbdallah,2007)andcausalrelationshipsin(Doerretal.
,2007).
Correlationisnotconsideredatallandeventinterpretationsareonlymentionedin(Jain,2008)butremainfuturework.
Withrespecttotherequirements,weintroducedspecializedinstantiationsoftheDnSpattern.
Here,theparticipationofobjectsinevents(1)isimplementedbytheparticipationpattern.
Italsoprovidesformodelingtheabsolutetimeandlocationofevents(2)andobjects(3).
Themereologypattern,causalitypattern,andcorrelationpatternimplementthestructuralrelationshipsbetweenevents(4a-4c).
Inaddition,themereologypatternallowsformodelingtherelativetemporalrelationsandrelativespa-tialrelationsbetweenevents(2)andobjects(3).
Inordertoexpresssuchrelativetem-poralrelationsbetweenevents,onecanfacilitatetheprovidedmeansofDOLCEsuchastheformalizationofAllen'sTimeCalculus13.
Thedocumentationpatternprovidesforannotatingevents(5).
Itcanbeseamlesslylinkedwithotherontologies,e.
g.
,theCoreOntologyforMultimedia(Arndtetal.
,2007)forpreciselydescribingdigitalme-diadatalikeimagesandvideos.
Finally,theinterpretationpatternsupportsdifferenteventinterpretations(6).
WeusetheDnSpatternforrepresentingoccurrencesintherealworld,i.
e.
,theeventsandobjectswearemodeling.
Theseoccurrencesaresubjecttodiscussionandinterpretationandmaynotbeobjectivelyobservable.
TheDnSpatternallowsforrep-resentingdifferentopinionsabouteventsandtheirparticipatingobjects.
ThisfeatureisnotprovidedbyDOLCE'sparticipationrelation.
Inthefollowing,wepresenttheontologypatternsoftheEvent-Model-Fthathavebeenemployedtomodelthesce-narioinSection3andillustratethemindiagrams.
ThiscomprisesalmostallpatternsoftheEvent-Model-F,expectfromthecorrelationpattern.
AcompletedescriptionoftheEvent-Model-Fcanbefoundin(Scherpetal.
,2009a).
5.
1.
1ParticipationPatternTheparticipationpatternoftheEvent-Model-Fenablestoformallyexpresstheparticipationofobjectsinevents.
AsshowninFigure15,participationisexpressedbyanF:EventParticipationSituationthatsatisfiesanF:EventParticipationDescription.
ThesituationincludestheEventbeingdescribedandtheObjectsparticipatinginthisevent.
TheEventParticipationDescriptionclassiesthedescribedeventanditsparticipantsbyusingtheconceptsF:DescribedEvent(specializedfromDUL:EventType)andtheobjectroleF:Participant(specializedfromDUL:Role).
TheconceptDescribedEventclassifiestheEventthatisdescribedbytheparticipationpattern,e.
g.
,theeventofaoodedcellar.
Likewise,instancesofParticipantclassifyobjectsasparticipantsoftheevent.
Forexample,thecitizencallingtheemergencyhotlinetoreportabouttheoodedcellar.
InstancesofParticipantcanberolesdenedinsomedomainontologyasindicatedinFig-ure15.
Forexample,anemergencyresponseontologythatdenestheroleofaperson13Availablefrom:http://wiki.
loa-cnr.
it/index.
php/LoaWiki:Ontologiesbeingaffected,i.
e.
,theemergencysubjectsuchasaCitizenRole,andtherolede-scribingtherescuestaffsuchasaFiremanRole.
Besidestheroleanobjectcanplayinaspecicparticipationpattern,alsothedescribedeventanditsparticipatingobjectsthemselvescanbedenedinsomedomainontologyasindicatedinFigure15.
TheparameterF:LocationParameterdescribesthegeneralspatialre-gionwheretheobjectsarelocated.
ItparametrizesaDUL:SpaceRegionanddenesapropertyDUL:isParameterFortotheParticipantrole.
TheObjectthatisclassiedbytheParticipanthasaQualitywiththepropertyDUL:hasRegionofaDUL:SpaceRegion.
Thus,usingtheF:LocationParameterwecandenethelocation(s)representedbyDUL:SpaceRegionsthatarerelevantfordescribingtheeventinagivencontext.
Forexample,whenquenchingahousereallremenhavetheirspeciclocationwithinandaroundthebuilding.
TheF:LocationParametercanthenbeusedtode-scribeingeneralthattheremenwhereatthatspecichouse,e.
g.
,informofsomelongitude-latituderectangular.
Thus,wedonotneedtoexplicitlystatewheretheindi-vidualremenare.
TheF:TimeParameterdescribesthegeneraltemporalregionwhentheeventhappened.
ItparametrizesaDUL:TimeIntervalanddenesapropertyDUL:isParameterFortotheDescribedEventrole.
Forexample,onecanstatethatthehouserehappenedonJune13,2006.
Figure15:Participationpattern5.
1.
2MereologyPatternEventsarecommonlyconsideredatdifferentabstractionlevelsdependingontheviewandtheknowledgeofaspectator.
Forinstance,theeventofaoodedcel-larmaybeconsideredassuchoraspartofthelargereventofaoodinginwhichmanyother(smaller)incidentsoccur.
ThemereologypatternshowninFig-ure16enablesexpressingsuchmereologicalrelationsascompositionofevents.
Thecompositeeventisthe"whole"andthecomponenteventsareits"parts".
Formally,aF:EventCompositionSituationincludesoneinstanceofaneventthatisclassiedbytheconceptF:Compositeandmanyeventsclassi-edasitsF:Component(s).
Accordingly,anEventCompositionSituationsatisfiesaF:CompositionDescriptionthatdefinestheconceptsCompositeandComponentforclassifyingthecompositeeventanditscomponentevents.
Figure16:MereologypatternEventsthatplaytheComponentrolemaybefurtherqualiedbytemporal,spatial,andspatio-temporalconstraints.
Aseventsareformallydenedasentitiesthatexistintimeandnotinspace(Scherpetal.
,2009a),constraintsincludingspatialrestrictionsareexpressedthroughtheobjectsparticipatinginthecomponentevent.
Forinstance,aComponenteventmayberequiredtooccurwithinacertaintime-interval,e.
g.
,thesecondweekofJune2009.
Dependingonitsobjects,aComponenteventmayalsohappeninacertainspatialregion.
Forexample,theoodingofatownshouldbecomposedofeventsthathaveobjectsassociatedtoit,whichhavesomecertainrangeoflongitudeandlatitude.
Finally,eventsandtheobjectsbondtoitmaybequaliedbyaspatio-temporalqualityliketheprogressofaoodthatextentsovertimeandspace,startingwithahighwaterlevellocatedinsomeareaofariverandextendingspatiallyovertimeintootherareas.
AnysuchconstraintsareformallyexpressedbyoneormultipleinstancesoftheF:EventCompositionConstraint.
Thus,withthecompositionpattern,eventsmaybearbitrarilytemporallyrelatedtoeachother,i.
e.
,theymightbedisjoint,overlapping,orotherwiseordered.
Inordertoexpresssuchrelativetemporalrelationsbetweenevents,onecanfacilitatetheprovidedmeansofDOLCEsuchastheformalizationofAllen'sTimeCalculus14.
14Availablefrom:http://wiki.
loa-cnr.
it/index.
php/LoaWiki:Ontologies5.
1.
3CausalityPatternCausalityisthephilosophicalprobleminvestigatingtheexistenceofanyspecial"tie"bindingcausesandeffectstogether(Itkonen,1983).
Itcanbequestionedeitheras"Why"or"How"(Itkonen,1983).
Ananswertothisquestionis(orbettersaidclaimstobe)acausalexplanation.
Whatexplains,isthecauseandthatwhatisexplainedistheeffect(Itkonen,1983).
Eventsarethemostnaturalconcepttoservefordeningcausalrelations(Quinton,1979).
Infact,causesandeffectsaretwospecictypesofevents(Itkonen,1983;Lombard,1986).
Acausalrelationshipisalwaysjustiedbysome(maybeimplicit)underlyingcausaltheory.
WedesignedacausalitypatternasdepictedinFigure17.
ThepatterndenestwoEventTypescalledF:CauseandF:EffectwhichclassifyEvents.
Itfur-therdenesaDUL:Description,whichisclassiedbyaF:Justification.
Bythis,thepatternexplicitlyexpressesthecausalrelationshipbetweenthecauseandtheeffectunderthejusticationofsometheory.
Atheorymightbeanopinion,ascien-ticlaw,ornotfurtherspecied.
Forexample,duringaheavystorm,apoweroutageeventmightoccurcausedbyasnappedpowerpoleevent.
TheJustificationofthiscausalrelationshipisthelawsofphysics.
Figure17:Causalitypattern5.
1.
4DocumentationPatternDocumentaryevidenceforaneventmaybegivenbyarbitraryobjects,e.
g.
,somesensordataormediadata,orbyotherevents.
Formally,thisrelationisexpressedinFbythedocumentationpatterndepictedinFigure18.
Itdenesthecon-ceptF:DocumentedEventthatclassiesthedocumentedeventandtheconceptF:Documenterthatclassiesthedocumentaryevidenceforthatevent.
Thisevi-dencecanbeexpressedbyanyspecializationofanObject,e.
g.
,adigitalphototakenwithacellphoneduringanincident,oraspecializationofEvent.
Forexample,digi-talmediadatalikeimagesandvideoscanbeclassiedasDocumenterandpreciselydescribedusingtheCoreOntologyforMultimedia(Arndtetal.
,2007)describedinSection5.
2.
Objectsaredocumentedviatheeventsinwhichtheyparticipate(seepar-ticipationpatterninSection5.
1.
1).
Figure18:Documentationpattern5.
1.
5InterpretationPatternTheperceptionofeventsasoccurrencesintherealworldheavilydependsonthecon-textandpointofviewoftheobserver.
Suchdifferent,context-dependenteventin-terpretationscanbedescribedformallybyinstantiatingthedifferentEvent-Model-Fpatternspresentedsofarandbindingthemtogetherwiththeinterpretationpatternde-pictedinFigure19.
Eachpatternmodelsasingle,specicinterpretationofaneventbyassociatingparticipations,mereological,causal,andcorrelativerelationships,aswellasdocumentationsrelevantinthecontextofaspecicinterpretation.
Intheemer-gencyscenario,twoemergencycontrolofcersmighthavedifferinginterpretationsofthepoweroutage.
Onemightbeconvincedthatthepoweroutageisduetoasnappedpowerpole,whiletheothermightthinkofamoreseriouscaseofadamagedpowerplant.
Bothconsiderthesameeventofapoweroutage,however,consideritfromdifferentpointsofviewthatinvolveothereventsandobjectsindifferentpatterns.
Figure19:InterpretationpatternFormally,theinterpretationpatternshowninFigure19denesaF:InterpretantthatisspecializedfromEventTypeandclassiesthein-terpretedEvent.
TheInterpretantmightbedenedinsomedomainontologyanddetermineshowaneventisinterpreted,e.
g.
,asemergencyincidentinthecaseoftheemergencycontrolcenterorasnewseventdescribedinanewspaper.
Withineachinterpretation,weclassifytheF:RelevantSituations,namelythesituationssatisfyingtheparticipation,mereology,causality,correlation,anddocumentation.
ThesearedenedasspecializationsofRelevantSituation.
5.
2COMM—CoreOntologyforMultimediaTheCoreOntologyforMultimedia(COMM)(Arndtetal.
,2007)modelsthedomainofmultimediacontentandannotation,andisbasedonMPEG-7.
IncontrasttootherapproachestomodelingMPEG-7asanontology(Hunter,2005;Bloehdornetal.
,2005;GarciaandCelma,2005;IsaacandTroncy,2004;Tsinarakietal.
,2004),COMMisnotdesignedasaone-to-onemapping,butprovidesasetofpatternsthatcoverthecoreandrepetitivebuildingblocksofMPEG-7.
Inthefollowing,wediscussthepatternsofCOMMusedfortheemergencyresponsescenarioinSection3.
ThreerepetitivestructureshavebeenidentiedinMPEG-7:Decomposition.
MPEG-7providesdescriptorsforspatial,temporal,spatio-temporalandmediasourcedecompositionsofmultimediacontentintosegments.
Aseg-mentisthemostgeneralabstractconceptinMPEG-7andcanrefertoaregionofanimage,apieceoftext,atemporalsceneofavideo,oreventoamovingobjecttrackedduringaperiodoftime.
Annotation.
MPEG-7denesaverylargecollectionofdescriptorsthatcanbeusedtoannotateasegment.
Thesedescriptorscanbelow-levelvisualfeatures,audiofeaturesormoreabstractconcepts.
Theyallowtheannotationofthecontentofmultimediadocumentsorthemediaassetitself.
NestedDataStructures.
DescriptorsinMPEG-7arenestedstructurescontainingdif-ferentkindsofdata.
Exceptforthesemanticannotation,datareferstostringsornumericalvaluessuchastheencodingofanimageorthevaluesofacolorhis-togram.
ThesethreestructuresaremodeledaspatternsofourontologyCOMM.
Thepat-ternsarethedecompositionpattern,annotationpattern,andthedigitaldatapattern.
Beforewediscussthepatternsindetail,wewillintroducesomecentralconceptsthatarepresentinallthepatterns:DigitalData.
Withinthedomainofmultimediaannotation,thenotionofdigitaldataiscentral—boththemultimediacontentbeingannotatedandtheannotationsthem-selvesareexpressedasdigitaldata.
WeconsiderDigitalDataentitiesofarbitrarysizetobeInformationObjects,whichareusedforcommuni-cationbetweenmachines.
TheIOdesignpatternstatesthatDescriptionsareexpressedbyInformationObjects,whichhavetobeaboutfacts(representedbyindividualsoftypeEntity).
ThesefactsaresettingsforSituationsthathavetosatisfytheDescriptionsthatareexpressedbyInformationObjects.
Thischainofconstraintsallowsthemodelingofcomplexdatastructurestostoredigitalinformation.
MultimediaData.
ThisencapsulatestheMPEG-7notionofmultimediacontentsuchasimages,audio,text,opendocumentformat(ODF)documents,andisasubcon-ceptofDigitalData.
MultimediaDataisanabstractconceptthathastobefurtherspecializedforconcretemultimediacontenttypes(e.
g.
AudioDatacorrespondstothedatarepresentingtherecordedaudiosignal).
AccordingtotheIOpattern,MultimediaDataisrealizedbysomephysicalMedia,e.
g.
,MediaProfile,whichcontainsinformationaboutthestoragelocation,letype,andothers.
Thisconceptisneededforannotatingthephysicalrealizationofmultimediacontent.
Media.
MediaobjectsrepresenttherealizationsoftheinformationrepresentedasMultimediaData.
Inthecontextofelectronicdevicesthisreferstolesac-cessibleviasomeprotocol,butmightinamoregeneralsettingalsorefertophysicalrealizationssuchasapainting.
Method.
AMethodreferstosomemanualprocess,anditssubclassAlgorithmtoa(semi-)automaticprocessthatprocessessomeinputandgeneratesoutput.
Examplesaretheextractionoffeatures,e.
g.
,aDominantColorExtractionAlgorithm,orthemanualannotationofcontentwithasemanticconcept.
Input/OutputRoles.
MethodsalwaysdeneInputRolesandOutputRoles.
Theinputroleclassiestheobjectthatisprocessed,whiletheoutputroleclassiestheresults.
Forexample,asegmentationdenesexactlyoneInputSegmentRolewhichclassiestheimagethatisbeingsegmentedandatleastoneOutputSegmentRolewhichclassiestheresultingsubsegments.
Inthefollowingsubsections,wepresentthethreepatternsofCOMM.
Forex-pressingconcretedescriptorsofMPEG-7aspecializationoftheconceptsinvolvedineachpatternisrequired,butforeasiercomprehensionwediscussthepatternonamoreabstractlevel.
Thedetailsarecomprehendedbytheontologyitselfwhichisavailablethroughourwebsite:http://west.
uni-koblenz.
de/Research/ontologies/.
5.
2.
1DigitalDataPatternThedigitaldatapatternisdenedasdepictedinFigure20.
DigitalDataex-pressesStructuredDataDescriptions,whichdenemeaningfullabelsfortheinformationcontainedbyDigitalData.
Thisinformationisrepresentedbyliteralssuchasscalars,matrices,strings,rectangles,orpolygons.
InDOLCEterms,thesevaluesarerepresentedasaQualitythatisassociatedwiththeDigitalDataandlocatedinRegions.
InthecontextofaDescription,theseRegionsareparametrizedbyParameters.
StructuredDataDescriptionsdeneStructuredDataParameters,forwhichthequalitieslocatedintheparametrizedRegionsassignvaluestotheDigitalData.
ReferringtotheexampleinFig-ure5(cf.
Section3),weseethattheformalizationofdatastructuressofarisnotsufcient.
ComplexMPEG-7typescanincludenestedtypesthatagainhavetoberepresentedbyStructuredDataDescriptions.
Inourexample,theMediaInstanceDescriptorcontainstheMediaLocatorDescriptor.
ThedigitaldatapatterncoverssuchcasesbyallowingaDigitalDatain-stancetobeaboutanotherDigitalDatainstancewhichexpressesthenestedStructuredDataDescription.
Figure20:Digitaldatapattern5.
2.
2DecompositionPatternFollowingtheDnSpattern,weconsiderthedecompositionofMultimediaDatatobeaSituationthatsatisesaMethod.
MorespecicallytheSituationisaSegmentDecomposition.
ThedescriptionreferstoanalgorithmsuchasaSegmentationAlgorithm(cf.
Figure6)ortoamanualsegmenta-tionMethodlikeatoolthatallowstheusertodrawaboundingboxaroundadepictedface.
OfparticularinterestwithrespecttothedecompositionpatternaretherolesthataredenedbyaSegmentationAlgorithmoraMethod.
TheinputtoasegmentationisMultimediaDatathatisclassiedbyanInputSegmentRole,whiletheMultimediaDatareferringtotheoutputseg-mentsareclassiedbyOutputSegmentRoles.
Thesedataentitieshaveasset-tingtheSegmentDecompositionsituation.
OutputSegmentRolesaswellasSegmentDecompositionsarethenspecializedforspecictypesofmedia(accord-ingtothesegmentanddecompositionhierarchiesofMPEG-7(MPEG-7(2001),part5,section11).
ThedecompositionpatternisdepictedinFigure21.
PleasenotethattheconceptTextDataisactuallynotprovidedbyCOMM.
IthasbeenaddedtothedecompositionpatternbytheOntologyforKnowledgeAcquisition(OAK)(Iria,2009)inordertosupporttheannotationanddecompositionoftextualdata.
IntermsofMPEG-7,unsegmented(complete)multimediacontentalsocor-respondstoasegment.
Consequently,annotationsofcompletemultimediacon-tentstartwitharootsegment.
InordertodesignateMultimediaDatain-stancesthatcorrespondtotheserootsegmentsthedecompositionpatternprovidestheRootSegmentRoleconcept.
NotethatRootSegmentRolesarenotdenedbyMethodswhichdescribeSegmentDecompositions.
TheyareratherdenedbyMethodswhichcausetheproductionofmultimediacontent.
Thesemethodsaswellasannotationmodeswhichallowthedescriptionoftheproductionprocess(e.
g.
MPEG-7(2001),part5,section9)arecurrentlynotcoveredbyourontology.
Nevertheless,theprerequisiteforenhancingtheCOMMintothisdirectionisalreadygiven.
Figure21:DecompositionpatternThedecompositionpatternalsoreectstheneedforlocalizingsegmentswithintheinputsegmentofadecompositionaseachOutputSegmentRolerequiresaMaskRolethatclassiessomeDigitalDatawhichexpressesoneLocalizationDescriptor.
Thelatterdescribesthelocationofasegment,e.
g.
,startandendtimeofanaudiosegment.
Inourexample,wedidnotincludethisrequiredinformationinordertoreducethecomplexityofthediagram.
5.
2.
3AnnotationPatternInthissection,wedescribetheattachmentofmetadata,i.
e.
,annotationstobothMultimediaDataorMediaasdepictedinFigures22,23,and24.
Ingeneral,wedistinguishthreeannotationpatterns.
ThecontentannotationpatternmodelstheannotationofMultimediaDatawithmetadatarepresentedasDigitalData,i.
e.
,ingeneralMPEG-7descriptorssuchasadominantcolordescriptor.
Themediaan-notationpatternisusedtoattachmetadatatothemedia,e.
g.
,thelenameorlesize.
Finally,thesemanticannotationpatternformalizesthesemanticannotationofMultimediaData.
Asemanticannotationreferstosomeindividualofadomainontology.
EachannotationpatternconsistsofanAnnotation(subclassofSituation)thatsatisesaMethod.
TheMethoddenesanInputRolethatisspecializedtoeitherAnnotatedDataRoleorAnnotatedMediaRole,dependingonwhatisannotated.
FurthermoreitspeciesatleastoneOutputRole,whichisspecializedtoAnnotationRoleforthecontentannotationpatternandmediaannotationpatternandspecializedtoSemanticLabelRoleincaseofthesemanticannotationpattern,respectively.
Inthefollowing,wediscusstheindividualpatternsinmoredetail.
Westartwiththecontentannotationpattern(cf.
Figure22).
AsContentAnnotationsweunderstandSituationsthatincludeannotatedme-diaasMultimediaDataclassiedbyanAnnotatedDataRole.
Themeta-dataisrepresentedbyDigitalData,whichisclassiedbyAnnotationRoles.
TherolesaredenedbyanAlgorithmoraMethod,respectively.
Theac-tualmetadatathatisrepresentedbyaDigitalDataentitydependsontheStructuredDataDescriptionthatthemetadatacomprises.
Thesestructureddescriptionsareformalizedusingthedigitaldatapattern(seeSection5.
2.
1)andtypi-callyrefertoMPEG-7descriptors,althoughthisisnotrequired.
Thecontentannota-tionpatternisusedtorepresentrathertechnicalmetadatasuchasextractedlow-levelfeatures.
Figure22:ContentannotationpatternThemediaannotationpatternformsthebasisfordescribingthephysicalinstancesofmultimediacontent(cf.
Figure23).
Itdiffersfromthecontentannotationpatterninonlyonerespect:itistheMediathatisbeingannotatedandthereforeplaysanAnnotatedMediaRole.
ThesituationisspecializedtoaMediaAnnotation.
AnexampleoftheapplicationofthispatternisgiveninFigure5inSection3.
Theexampledepictsalerealizinganaudiorecording.
TheaudiorecordingisannotatedwiththeMPEG-7mediaproledescriptor,whichintheexamplecontainstheURItothele.
Finally,COMMalsoprovidesthesemanticannotationpatterndepictedinFig-ure24.
MPEG-7providesthemeanstomodelsemanticsasMPEG-7descriptorsFigure23:Mediaannotationpattern(seeMPEG-7(2001),Part5,Section12).
However,inthecontextofanontology-basedapproachlikeCOMMtheintegrationofdomain-specicontologiesismoreappropri-atethenusingtheMPEG-7descriptorforsemantics.
Thus,forthesemanticannotationCOMMreliesondomain-specicontologiesthatrepresent,e.
g.
,realworldentitiesthataredepictedintheannotatedmultimediacontent.
Consequently,thesemantican-notationpatternspecializesthecontentannotationpatterntoallowtheconnectionofmultimediadescriptionswithdomaindescriptionsprovidedbyindependentdomainontologies.
5.
3X-COSIMO—CoreOntologyforPersonalInformationMan-agementTheCross-ContextSemanticInformationManagementOntology(X-COSIMO)pro-videsaformallypreciserepresentationofpersonalinformationandassociatedtaskstofosteritsreuseacrosspersonalinformationmanagementapplicationssuchastaskmanagers,emailclients,andlemanagementtools.
Itisusedamongothersforper-sonalinformationmanagementintheX-COSIMsemanticdesktop(Franzetal.
,2007).
Inthefollowing,wedetailontwoofthepatternsprovidedbyX-COSIMOthatareemployedtodescribetheemergencyscenario,namelythecommunicationpatternandthetaskpattern.
5.
3.
1CommunicationPatternTheaimofthecommunicationpatternistoenableauniedviewontocommunica-tionandtorepresentinformationdealtinthecontextofcommunicationforreuseinfurthercontextswhilemaintaininginformationlinkage.
Forachievingtherst,theFigure24:SemanticannotationpatterncommunicationpatternprovidesaconceptualviewofcommunicationasdevelopedbyJakobson(Jakobson,1960).
Inhiswork,JakobsondenestheconceptofaMessagethatisaboutsomethingwhichhecallsContext.
ThecontextistransmittedviaaContact—aconnectionbetweentheAddresserandAddressee—andthatisexpressedwithinaCode.
Suchamodelgeneralizescommunicationsothatarbitrarycommunicationmodessuchaschat,phone,andemailcanberepresentedconsistently.
Forachievingthesecondgoal,thecommunicationpatternisdesignedusingtheDnSpatternasaCommunicationDescription(cf.
Fig.
25)thatrepresentsthecom-municationmodelofJakobson.
TheCommunicationDescriptiondenestherolesAddresser,Addressee,Contact,andMessageascommonrolesforanykindofcom-municationasdenedbyJakobson.
Jakobson'sconceptofContextandCodearenotdenedbythedescriptionasexistingontologiesformediaannotationanddecompositionsuchasCOMMexpresstheseaspectsinamoregeneralwayandareapplicablealsotoothercontextsthancommunication.
Additionally,aCommunicationDescriptionalsodenestheCommunicationParameterConversationStartandAddressParameter,whichexpressconstraintsontheclassiedCommunicationEventandtheAgentsthatoccurintheCommunicationSituation.
Precisely,itrequiresthatAgentshaveanAddressqualitythatiscontainedintheAddressSpaceparameterizedbytheAddressParameter.
Similarly,CommunicationEventsarerequiredtohaveaTemporalQualitythatiswithintheTimeIntervalparameterizedBytheparameterConversationStart.
Thequalitythenexpressestemporalinformationaboutthetimeatwhichaconversationtookplace.
Agentshavetheroleofaddressersandaddressees,whiledescriptionsofprotocolsplaytheroleofContact,e.
g.
,aBishererschienenArbeitsberichteausdemFachbereichInformatik(http://www.
uni-koblenz-landau.
de/koblenz/fb4/publications/Reports/arbeitsberichte)AnsgarScherp,CarstenSaathoff,ThomasFranz,SteffenStaab,DesigningCoreOntologies,ArbeitsberichteausdemFachbereichInformatik5/2011OlegV.
Kryuchin,AlexanderA.
Arzamastsev,KlausG.
Troitzsch,Thepredictionofcurrencyexchangeratesusingartificialneuralnetworks,ArbeitsberichteausdemFachbereichInformatik4/2011KlausG.
Troitzsch,AnnaLantsberg,RequirementsforHealthCareRelatedWebsitesinRussia:ResultsfromanAnalysisofAmerican,BritishandGermanExamples,ArbeitsberichteausdemFachbereichInformatik3/2011KlausG.
Troitzsch,OlegKryuchin,AlexanderArzamastsev,Auniversalsimulatorbasedonartificialneuralnetworksforcomputerclusters,ArbeitsberichteausdemFachbereichInformatik2/2011KlausG.
Troitzsch,NataliaZenkova,AlexanderArzamastsev,Developmentofatechnologyofdesigningintelligentinformationsystemsfortheestimationofsocialobjects,ArbeitsberichteausdemFachbereichInformatik1/2011KurtLautenbach,APetriNetApproachforPropagatingProbabilitiesandMassFunctions,ArbeitsberichteausdemFachbereichInformatik13/2010ClaudiaSchon,LinklessNormalFormforALCConcepts,ArbeitsberichteausdemFachbereichInformatik12/2010AlexanderHug,Informatikhautnaherleben,ArbeitsberichteausdemFachbereichInformatik11/2010MarcSantos,HaraldF.
O.
vonKortzfleisch,SharedAnnotationModel–EinDatenmodellfürkollaborativeAnnotationen,ArbeitsberichteausdemFachbereichInformatik10/2010GerdGrner,SteffenStaab,CategorizationandRecognitionofOntologyRefactoringPattern,ArbeitsberichteausdemFachbereichInformatik9/2010DanielEiing,AnsgarScherp,CarstenSaathoff,IntegrationofExistingMultimediaMetadataFormatsandMetadataStandardsintheM3O,ArbeitsberichteausdemFachbereichInformatik8/2010StefanScheglmann,AnsgarScherp,SteffenStaab,Model-drivenGenerationofAPIsforOWL-basedOntologies,ArbeitsberichteausdemFachbereichInformatik7/2010DanielSchmei,AnsgarScherp,SteffenStaab,IntegratedMobileVisualizationandInteractionofEventsandPOIs,ArbeitsberichteausdemFachbereichInformatik6/2010RüdigerGrimm,DanielPhler,E-Mail-Forensik–IP-AdressenundihreZuordnungzuInternet-TeilnehmernundihrenStandorten,ArbeitsberichteausdemFachbereichInformatik5/2010ChristophRingelstein,SteffenStaab,PAPEL:SyntaxandSemanticsforProvenance-AwarePolicyDefinition,ArbeitsberichteausdemFachbereichInformatik4/2010NadineLindermann,SylviaValcárcel,HaraldF.
O.
vonKortzfleisch,EinStufenmodellfürkollaborativeoffeneInnovationsprozesseinNetzwerkenkleinerundmittlererUnternehmenmitWeb2.
0,ArbeitsberichteausdemFachbereichInformatik3/2010MariaWimmer,DagmarLück-Schneider,UweBrinkhoff,ErichSchweighofer,SiegfriedKaiser,AndreasWieber,FachtagungVerwaltungsinformatikFTVIFachtagungRechtsinformatikFTRI2010,ArbeitsberichteausdemFachbereichInformatik2/2010MaxBraun,AnsgarScherp,SteffenStaab,CollaborativeCreationofSemanticPointsofInterestasLinkedDataontheMobilePhone,ArbeitsberichteausdemFachbereichInformatik1/2010MarcSantos,EinsatzvonSharedIn-situProblemSolving"AnnotationeninkollaborativenLern-undArbeitsszenarien,ArbeitsberichteausdemFachbereichInformatik20/2009CarstenSaathoff,AnsgarScherp,UnlockingtheSemanticsofMultimediaPresentationsintheWebwiththeMultimediaMetadataOntology,ArbeitsberichteausdemFachbereichInformatik19/2009ChristophKahle,MarioSchaarschmidt,HaraldF.
O.
vonKortzfleisch,OpenInnovation:KundenintegrationamBeispielvonIPTV,ArbeitsberichteausdemFachbereichInformatik18/2009DietrichPaulus,LutzPriese,PeterDecker,FrankSchmitt,Pose-TrackingForschungsbericht,ArbeitsberichteausdemFachbereichInformatik17/2009AndreasFuhr,TassiloHorn,AndreasWinter,Model-DrivenSoftwareMigrationExtendingSOMA,ArbeitsberichteausdemFachbereichInformatik16/2009EckhardGromann,SaschaStrau,TassiloHorn,VolkerRiediger,AbbildungvongrUMLnachXSDsoamig,ArbeitsberichteausdemFachbereichInformatik15/2009KerstinFalkowski,JürgenEbert,TheSTORComponentSystemInterimReport,ArbeitsberichteausdemFachbereichtInformatik14/2009SebastianMagnus,MarkusMaron,AnEmpiricalStudytoEvaluatetheLocationofAdvertisementPanelsbyUsingaMobileMarketingTool,ArbeitsberichteausdemFachbereichInformatik13/2009SebastianMagnus,MarkusMaron,KonzepteinerPublicKeyInfrastrukturiniCity,ArbeitsberichteausdemFachbereichInformatik12/2009SebastianMagnus,MarkusMaron,APublicKeyInfrastructureinAmbientInformationandTransactionSystems,ArbeitsberichteausdemFachbereichInformatik11/2009AmmarMohammed,UlrichFurbach,Multi-agentsystems:ModelingandVirificationusingHybridAutomata,ArbeitsberichteausdemFachbereichInformatik10/2009AndreasSprotte,PerformanceMeasurementaufderBasisvonKennzahlenausbetrieblichenAnwendungssystemen:EntwurfeineskennzahlengestütztenInformationssystemsfüreinenLogistikdienstleister,ArbeitsberichteausdemFachbereichInformatik9/2009GwendolinGarbe,TobiasHausen,ProcessCommodities:EntwicklungeinesReifegradmodellsalsBasisfürOutsourcingentscheidungen,ArbeitsberichteausdemFachbereichInformatik8/2009PetraSchubertet.
al.
,Open-Source-SoftwarefürdasEnterpriseResourcePlanning,ArbeitsberichteausdemFachbereichInformatik7/2009AmmarMohammed,FriederStolzenburg,UsingConstraintLogicProgrammingforModelingandVerifyingHierarchicalHybridAutomata,ArbeitsberichteausdemFachbereichInformatik6/2009TobiasKippert,AnastasiaMeletiadou,RüdigerGrimm,EntwurfeinesCommonCriteria-SchutzprofilsfürRouterzurAbwehrvonOnline-berwachung,ArbeitsberichteausdemFachbereichInformatik5/2009HannesSchwarz,JürgenEbert,AndreasWinter,Graph-basedTraceability–AComprehensiveApproach.
ArbeitsberichteausdemFachbereichInformatik4/2009AnastasiaMeletiadou,SimoneMüller,RüdigerGrimm,AnforderungsanalysefürRisk-Management-Informationssysteme(RMIS),ArbeitsberichteausdemFachbereichInformatik3/2009AnsgarScherp,ThomasFranz,CarstenSaathoff,SteffenStaab,AModelofEventsbasedonaFoundationalOntology,ArbeitsberichteausdemFachbereichInformatik2/2009FrankBohdanovicz,HaraldDickel,ChristophSteigner,AvoidanceofRoutingLoops,ArbeitsberichteausdemFachbereichInformatik1/2009StefanAmeling,StephanWirth,DietrichPaulus,MethodsforPolypDetectioninColonoscopyVideos:AReview,ArbeitsberichteausdemFachbereichInformatik14/2008TassiloHorn,JürgenEbert,EinReferenzschemafürdieSprachenderIEC61131-3,ArbeitsberichteausdemFachbereichInformatik13/2008ThomasFranz,AnsgarScherp,SteffenStaab,DoesaSemanticWebFacilitateYourDailyTasks,ArbeitsberichteausdemFachbereichInformatik12/2008NorbertFrick,KünftigeAnfordeungenanERP-Systeme:DeutscheAnbieterimFokus,ArbeitsberichteausdemFachbereichtInformatik11/2008JürgenEbert,RüdigerGrimm,AlexanderHug,LehramtsbezogeneBachelor-undMasterstudiengngeimFachInformatikanderUniversittKoblenz-Landau,CampusKoblenz,ArbeitsberichteausdemFachbereichInformatik10/2008MarioSchaarschmidt,HaraldvonKortzfleisch,SocialNetworkingPlatformsasCreativityFosteringSystems:ResearchModelandExploratoryStudy,ArbeitsberichteausdemFachbereichInformatik9/2008BernhardSchueler,SergejSizov,SteffenStaab,QueryingforMetaKnowledge,ArbeitsberichteausdemFachbereichInformatik8/2008StefanStein,EntwicklungeinerArchitekturfürkomplexekontextbezogeneDiensteimmobilenUmfeld,ArbeitsberichteausdemFachbereichInformatik7/2008MatthiasBohnen,LinaBrühl,SebastianBzdak,RoboCup2008MixedRealityLeagueTeamDescription,ArbeitsberichteausdemFachbereichInformatik6/2008BernhardBeckert,ReinerHhnle,TestsandProofs:PapersPresentedattheSecondInternationalConference,TAP2008,Prato,Italy,April2008,ArbeitsberichteausdemFachbereichInformatik5/2008KlaasDellschaft,SteffenStaab,UnterstützungundDokumentationkollaborativerEntwurfs-undEntscheidungsprozesse,ArbeitsberichteausdemFachbereichInformatik4/2008RüdigerGrimm:IT-Sicherheitsmodelle,ArbeitsberichteausdemFachbereichInformatik3/2008RüdigerGrimm,HelgeHundacker,AnastasiaMeletiadou:AnwendungsbeispielefürKryptographie,ArbeitsberichteausdemFachbereichInformatik2/2008MarkusMaron,KevinRead,MichaelSchulze:CAMPUSNEWS–ArtificialIntelligenceMethodsCombinedforanIntelligentInformationNetwork,ArbeitsberichteausdemFachbereichInformatik1/2008LutzPriese,FrankSchmitt,PatrickSturm,HaojunWang:BMBF-Verbundprojekt3D-RETISEGAbschlussberichtdesLaborsBilderkennenderUniversittKoblenz-Landau,ArbeitsberichteausdemFachbereichInformatik26/2007StephanPhilippi,AlexanderPinl:Proceedings14.
Workshop20.
-21.
September2007AlgorithmenundWerkzeugefürPetrinetze,ArbeitsberichteausdemFachbereichInformatik25/2007UlrichFurbach,MarkusMaron,KevinRead:CAMPUSNEWS–anIntelligentBluetooth-basedMobileInformationNetwork,ArbeitsberichteausdemFachbereichInformatik24/2007UlrichFurbach,MarkusMaron,KevinRead:CAMPUSNEWS-anInformationNetworkforPervasiveUniversities,ArbeitsberichteausdemFachbereichInformatik23/2007LutzPriese:FiniteAutomataonUnrankedandUnorderedDAGsExtentedVersion,ArbeitsberichteausdemFachbereichInformatik22/2007MarioSchaarschmidt,HaraldF.
O.
vonKortzfleisch:ModularittalsalternativeTechnologie-undInnovationsstrategie,ArbeitsberichteausdemFachbereichInformatik21/2007KurtLautenbach,AlexanderPinl:ProbabilityPropagationNets,ArbeitsberichteausdemFachbereichInformatik20/2007RüdigerGrimm,FaridMehr,AnastasiaMeletiadou,DanielPhler,IlkaUerz:SOA-Security,ArbeitsberichteausdemFachbereichInformatik19/2007ChristophWernhard:TableauxBetweenProving,ProjectionandCompilation,ArbeitsberichteausdemFachbereichInformatik18/2007UlrichFurbach,ClaudiaObermaier:KnowledgeCompilationforDescriptionLogics,ArbeitsberichteausdemFachbereichInformatik17/2007FernandoSilvaParreiras,SteffenStaab,AndreasWinter:TwoUse:IntegratingUMLModelsandOWLOntologies,ArbeitsberichteausdemFachbereichInformatik16/2007RüdigerGrimm,AnastasiaMeletiadou:RollenbasierteZugriffskontrolle(RBAC)imGesundheitswesen,ArbeitsberichteauddemFachbereichInformatik15/2007UlrichFurbach,JanMurray,FalkSchmidsberger,FriederStolzenburg:HybridMultiagentSystemswithTimedSynchronization-SpecificationandModelChecking,ArbeitsberichteausdemFachbereichInformatik14/2007BjrnPelzer,ChristophWernhard:SystemDescription:"E-KRHyper",ArbeitsberichteausdemFachbereichInformatik,13/2007UlrichFurbach,PeterBaumgartner,BjrnPelzer:HyperTableauxwithEquality,ArbeitsberichteausdemFachbereichInformatik,12/2007UlrichFurbach,MarkusMaron,KevinRead:LocationbasedInformationsystems,ArbeitsberichteausdemFachbereichInformatik,11/2007PhilippSchaer,MarcoThum:State-of-the-Art:InteraktioninerweitertenRealitten,ArbeitsberichteausdemFachbereichInformatik,10/2007UlrichFurbach,ClaudiaObermaier:ApplicationsofAutomatedReasoning,ArbeitsberichteausdemFachbereichInformatik,9/2007JürgenEbert,KerstinFalkowski:AFirstProposalforanOverallStructureofanEnhancedRealityFramework,ArbeitsberichteausdemFachbereichInformatik,8/2007LutzPriese,FrankSchmitt,PaulLemke:AutomatischeSee-ThroughKalibrierung,ArbeitsberichteausdemFachbereichInformatik,7/2007RüdigerGrimm,RobertKrimmer,NilsMeiner,KaiReinhard,MelanieVolkamer,MarcelWeinand,JrgHelbach:SecurityRequirementsforNon-politicalInternetVoting,ArbeitsberichteausdemFachbereichInformatik,6/2007DanielBildhauer,VolkerRiediger,HannesSchwarz,SaschaStrau,grUML–EineUML-basierteModellierungssprachefürT-Graphen",ArbeitsberichteausdemFachbereichInformatik,5/2007RichardArndt,SteffenStaab,RaphalTroncy,LyndaHardman:AddingFormalSemanticstoMPEG-7:DesigningaWellFoundedMultimediaOntologyfortheWeb,ArbeitsberichteausdemFachbereichInformatik,4/2007SimonSchenk,SteffenStaab:NetworkedRDFGraphs,ArbeitsberichteausdemFachbereichInformatik,3/2007RüdigerGrimm,HelgeHundacker,AnastasiaMeletiadou:AnwendungsbeispielefürKryptographie,ArbeitsberichteausdemFachbereichInformatik,2/2007AnastasiaMeletiadou,J.
FelixHampe:BegriffsbestimmungunderwarteteTrendsimIT-Risk-Management,ArbeitsberichteausdemFachbereichInformatik,1/2007GelbeReihe"(http://www.
uni-koblenz.
de/fb4/publikationen/gelbereihe)LutzPriese:SomeExamplesofSemi-rationalandNon-semi-rationalDAGLanguages.
ExtendedVersion,FachberichteInformatik3-2006KurtLautenbach,StephanPhilippi,andAlexanderPinl:BayesianNetworksandPetriNets,FachberichteInformatik2-2006RainerGimnichandAndreasWinter:WorkshopSoftware-ReengineeringundServices,FachberichteInformatik1-2006KurtLautenbachandAlexanderPinl:ProbabilityPropagationinPetriNets,FachberichteInformatik16-2005RainerGimnich,UweKaiser,andAndreasWinter:2.
Workshop''ReengineeringProzesse''–SoftwareMigration,FachberichteInformatik15-2005JanMurray,FriederStolzenburg,andToshiakiArai:HybridStateMachineswithTimedSynchronizationforMulti-RobotSystemSpecification,FachberichteInformatik14-2005ReinholdLetz:FTP2005–FifthInternationalWorkshoponFirst-OrderTheoremProving,FachberichteInformatik13-2005BernhardBeckert:TABLEAUX2005–PositionPapersandTutorialDescriptions,FachberichteInformatik12-2005DietrichPaulusandDetlevDroege:Mixed-realityasachallengetoimageunderstandingandartificialintelligence,FachberichteInformatik11-2005JürgenSauer:19.
WorkshopPlanen,SchedulingundKonfigurieren/Entwerfen,FachberichteInformatik10-2005PascalHitzler,CarstenLutz,andGerdStumme:FoundationalAspectsofOntologies,FachberichteInformatik9-2005JoachimBaumeisterandDietmarSeipel:KnowledgeEngineeringandSoftwareEngineering,FachberichteInformatik8-2005BennoSteinandSvenMeierzuEien:ProceedingsoftheSecondInternationalWorkshoponText-BasedInformationRetrieval,FachberichteInformatik7-2005AndreasWinterandJürgenEbert:Metamodel-drivenServiceInteroperability,FachberichteInformatik6-2005JoschkaBoedecker,NorbertMichaelMayer,MasakiOgino,RodrigodaSilvaGuerra,MasaakiKikuchi,andMinoruAsada:Gettingcloser:HowSimulationandHumanoidLeaguecanbenefitfromeachother,FachberichteInformatik5-2005TorstenGippandJürgenEbert:WebEngineeringdoesprofitfromaFunctionalApproach,FachberichteInformatik4-2005OliverObst,AnitaMaas,andJoschkaBoedecker:HTNPlanningforFlexibleCoordinationOfMultiagentTeamBehavior,FachberichteInformatik3-2005AndreasvonHessling,ThomasKleemann,andAlexSinner:SemanticUserProfilesandtheirApplicationsinaMobileEnvironment,FachberichteInformatik2-2005HeniBenAmorandAchimRettinger:IntelligentExplorationforGeneticAlgorithms–UsingSelf-OrganizingMapsinEvolutionaryComputation,FachberichteInformatik1-2005