1 Typical query on a memory Request language Data and type - PDF document

ACACIA in short… � Objectives: Offer methodological and software support (i.e. models, methods and tools) for Managing enterprise construction, management and diffusion of applications as dynamic corporate memories. resources in corporate semantic � Corporate memory : Explicit and persistent webs materialization of crucial knowledge and information of an organization to ease access, an application scenario for sharing and reuse by the members of the semantic web services. organization in individual and collective tasks. � Corporate memories as corporate semantic Fabien Gandon, Moussa Lo, Olivier Corby, webs Rose Dieng-Kuntz Corporate web & intranet Corporate semantic Web � Past and Current work on Corporate Memories � Resources: persons , documents (XML, (10 minutes) HTML...), services, software, hardware, etc. � Current and future work with Semantic WS � Ontologies: describing the conceptual (10 minutes) vocabulary shared by the organisation communities web server � Semantic annotations: on these resources intranet mail (e.g. persons’ skills, document contents , characteristics of services/software/hardware), using the vocabulary defined in the ontologies � Diffusion on the intranet / corporate web. Corporate semantic web CORESE semantic search engine <accident> <date> 19 Mai 2000 </date> <description> corese <facteur>le facteur </description> </accident> Ontologies Documents XML Legacy sys. Users CG semantic <ns:article rdf:about="http://intranet/articles/ecai.doc"> rules <rdfs:Class rdf:ID="thing"/> annotations ontologies <ns:title>MAS and Corporate Semantic Web</ns:title> <rdfs:Class rdf:ID="person"> web server <ns:author> <rdfs:subClassOf decisions/ push RDFS <ns:person rdf:about="http://intranet/employee/id109" /> rdf:resource="#thing"/> OWL answer query </ns:author> </rdfs:Class> RDF </ns:article> Schemata Annotations in RDF formed by web server in RDFS instances of schemata in RDFS intranet mail CORESE Semantic Web stack RDFS CG Support QUERIES Web server RULES PROJECTION RDF CG Base ONTOLOGY RDFS CG Results RDF/S INFERENCES Rules CG Rules RDF XML NAMESPACES Queries CG Query URI UNICODE 1

Typical query on a memory Request language � Data and type operators: = <= ~ != … Find the documents about Java and return the titles and the authors : � XML Schema Datatypes : number, boolean, string, date, etc. select ?doc c:title ?person where � Natural language: xml:lang=“en-us” ?doc rdf:type c:Document ?doc c:concern ?topic � Boolean expressions and / or ?topic rdf:type c:Java � Query the Schema ?doc c:title ?title ?title ~ “web” � Negation of arc, optional arc, paths, ?doc c:author ?person approximation, variables on properties, etc. Discover paths between resources Approximate search � Request: Technical Report about Java written by an engineer ? Find a link between two persons (with maximal � Approximation : Technical Report → Handout Engineer → Team length of 4 relations) Object ?x rdf:type ex:Person ?y rdf:type ex:Person Document Actor ?x ex:relation[4] ?y Person Team Report Course Engineer Researcher Research R. Technical R. Handout Production rules Architecture HTTP Response Classify a ressource XHTML, CSS, SVG JavaScript If a person wrote a Ph.D. thesis on a suject the s/he is a doctor and an expert on that subject. Join Projection engine engine Type ?person author ?doc Notio HTTP inference Request engine PhDThesis Person CG author ?doc rdf:type PhDThesis Manager ?person ?person ?doc concern ?topic Topic concern → ?topic ?person expertIn ?topic ?person rdf:type PhD JDBC PhD expertIn ?person 2

HCI and portal generation Corporate distributed knowledge � Build a list with sub-classes of Person corese <select name=‘ ihm _ person’ title= 'Profession' > sparql <query> CG ?x rdfs:subClassOf c:Person </query> semantic rules annotations ontologies web server </select> RDFS OWL RDF � HTML rendering: web server intranet mail agents � Request associated to the list : ?p rdf:type get: ihm_person DAI & DFK � archives distributed all over organisation � find best archive for new annotation � Contract-net (CfP, Proposal, Accept/Reject): :protocol fipa contract net :content <RDF Annotation> :language CoMMA-RDF :ontology CoMMA Ontology 5:accept/ 6:accept/ 1:cfp reject 2:cfp reject AMLocal:Med *:Med *:Arch 8:inform 4:propose 3:propose 7:inform 6:accept/ 2:cfp reject *:Arch 7:inform 3:propose :protocol fipa contract net :content <propose bid = distance archive / refuse / not understood> :language CoMMA-RDF :ontology CoMMA Ontology Corporate web services Web services & Enterprise application � Transversal use of enterprise modeling corese � End of 90’s: enterprise modeling for KM sparql CG � In the past 2 years: technology and application semantic integration can benefit from these models too rules annotations ontologies web server RDFS OWL � Evolution of KM scenarios RDF � Until end of 90’s focus on: knowledge capture, web server storage, access and diffusion intranet mail � More and more often: computation, decision, web enterprise routing, transformation,… knowledge workflow agents services � Unified and integrated access to knowledge applications sources and corporate applications 3

Memories with a broaden scope Corporate semantic web services � Corporate memories including: corese � information storage services; sparql CG � information capture services; semantic rules annotations ontologies � computation and inference services; web server RDFS OWL � information flows management services; RDF � information mediation services; web server intranet mail � information presentation services; � Resources may be internal or external web service enterprise agents services � external standard library // online service; annotations applications � interoperate smoothly and integrate workflows at the business layer. Corporate semantic web services Discover and invoke a service � Consider service just like other corporate semantic web resources: annotate them � Sub set of OWL-S (profile, grounding, ++) � Rely on CORESE to provide a corporate semantic UDDI Corese Requester Provider � Access resources (documents, services) search input form run & display Corese Webapp Form servlet Invocation servlet select submit Service description Input description & extension <process:AtomicProcess rdf:ID="PosteSecretaire"> <service:Service rdf:ID=" PosteService_Secretaire "> <service:describes rdf:resource="#PosteService_Secretaire"/> <service:presents <process:hasInput> rdf:resource="#Profile_Poste_Service_Secretaire"/> <process:Input rdf:ID="PosteSecr_input"> <service:describedBy rdf:resource="#PosteSecretaire"/> <process:parameterType>&xsd;#string</process:parameterType> <service:supports rdf:resource="#PosteGrounding_Secretaire"/> <process: semanticType rdf:resource="&doc; #EmployeeName "/> </service:Service> </process:Input> </process:hasInput> <profile:Profile rdf:ID="Profile_Poste_Service_Secretaire"> <service:presentedBy rdf:resource="#PosteService_Secretaire"/> <process:hasOutput> <profile:has_process rdf:resource="#PosteSecretaire"/> <process:Output rdf:ID="PosteSecr_output"> <profile:serviceName>PosteSecretaire</profile:serviceName> <process:parameterType>&xsd;#string</process:parameterType> <profile: hasInput rdf:resource=" #PosteSecr_input "/> <process:semanticType rdf:resource="&doc;#AssistantPhone"/> <profile:hasOutput rdf:resource="#PosteSecr_output"/> </process:Output> </profile:Profile> </process:hasOutput> </process:AtomicProcess> 4

Extension parameters OWL-S Composing with memory <owl:ObjectProperty rdf:ID=" semanticType "> <rdfs:domain rdf:resource=" #Parameter "/> </owl:ObjectProperty> (…) <cos:rule> <cos:if> ?x rdf:type c:Employee ?x c:Name ?n </cos:if> <cos:then> ?x c:EmployeeName ?n </cos:then> </cos:rule> (…) <c:Employee rdf:ID='ML'> <c: Name >Moussa Lo</c:Name> </c:Employee> Corporate semantic web applications Ongoing… � Composing services corese � Manual: Classic scenario = IT manager describes a sparql composed service and publishes its description CG semantic � Macro-recording interface rules annotations ontologies web server RDFS OWL � Semi-automatic: finding a path from a given input RDF to a given output web server � Fully-automatic ??? intranet mail � Capture and decompose end-users’ needs? service web � Even small examples seem to require a lot of domain service enterprise composition agents services description knowledge annotations applications � Controlled workflow description Composite services Automatic “sequences” using paths 5