Big Data Big Media the new paradigm of multimedia content management with Perfect Memory at Big Media by Actuonda

  • Published on
    14-Apr-2017

  • View
    826

  • Download
    0

Transcript

Big Data, Big MediaTHE NEW PARADIGM OF THE MULTIMEDIA CONTENT MANAGEMENTSemtech top 10 startup of 2013 IBC Award for Content management 2013IBC Award for technology Who caught my eye looking for blue skies of IBC 2013@perfect__memory http://perfect-memory.comPerfect Memory The Team2BIG DATA - BIG MEDIAPerfect Memory The eco-systemRegistered in 2008, a Ltd with 245 K capitalFunded by SOFIMAC Partner a major investor in FranceDeploying for Media, Media-Trade, Archivers and Big CompaniesExpert in management, indexation, and of monetization of mass volumes of Multi MediaOwning a unique Middleware process transforming raw data into Knowledge3BIG DATA - BIG MEDIAThe context of Big DataBig Data is a buzz word that hide different realties :- Volume issue (data volume, media volume),- Interoperability that serves the ubiquity of the content (anywhere, anytime, anybody),- Diversity of sources (MAM, DB, internal, external, structured, unstructured).Deals with volume, ubiquity and diversityLe 09/10/2013 BIG DATA - BIG MEDIA 4The context of Big DataVolume is an old issue :- Upstream: Require to solve the administration, the exploitation and the indexing of the content,- Downstream: provide mapping and representation of the content. A posteriori, analytics, Data miningHealth, Oil, Retail (see the the diaper & beer case)Interoperability is a consequence of the raising of the Internet:- Cooperation, communities, coworking,- It requires standards (XML Schema, MPEG 4,7,21, MXF, OAIS, RDF. Deals with volume, ubiquity and diversity- It requires standards (XML Schema, MPEG 4,7,21, MXF, OAIS, RDF. A priori, structuration of the content MediaNew movie workflow (production to distribution)Diversity of sources:- Structuration, meaning,- Linked data. A priori, knowledge processingMedia, industrie, EducationWeb 3.0 paradigm (Semantic, LOD, Open Data)Le 09/10/2013 BIG DATA - BIG MEDIA 5Facts1.Multi media contents are growing massively2.Media inventories are managed by heterogeneous systems3. Indexation, if done, is mainly donemanually- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Media2000 2005 Media Asset Management Systemstime6BIG DATA - BIG MEDIAFacts generating deep archives issue . Main factsLost opportunities and management issueWaste of time7BIG DATA - BIG MEDIAMedia Challenge : New needs8BIG DATA - BIG MEDIAMedia Challenge : new comersThe Media-brands 9BIG DATA - BIG MEDIASummary of functional needsDuring our conversations we have identified the needs to: Structure the content using opened and documented standards, Link, enrich & index the massive volumes of Contents, Browse inside the massive volumes of Contents, Manage the content all along its life cycle, Monetize & Value the content. Become autonomous in the administration of the knowledge and its infrastructure Being flexible in term of strategy of knowledge management Avoid starting from scratch10BIG DATA - BIG MEDIASummary of our solutionThe semantic middleware is : Natively compliant to the main media standards (EBU Core, FIMS, OAIS,) Providing a media mapping manager (multiple instances of items handling), A non intrusive, scalable and flexible platform, Self learning, opened to other modules and functionalities, Transferable platform11BIG DATA - BIG MEDIASemantic layer cakeFrom modeling to exploitationUser InterfaceUSAGE360 RenderingPUBLISHINGInference rulesENRICHMENTSemantic DataPRODUCTION, INGESTOntology & knowledge baseMODELING12BIG DATA - BIG MEDIASemantic valorizationWhy? From DATA to information Understand information and build theknowledgeknowledge Provide solutions to value the content.13BIG DATA - BIG MEDIASemantic valorizationBring semantic to the mediaData InfoKnowledge2 persons, face to face, smiling and laughingSemanticsystemData InfoKnowledge2 persons, face to face, smiling and laughing Samuel L. Jackson (Person) Leonardo DiCaprio (Person) Thumbnail from Django unchained Quentin Tarantino behind the camera14BIG DATA - BIG MEDIAEnhancement & EnrichmentFrom flat to rich contentENTITY Organisation URI: MMC#RTBF#6756593Name : RTBFENTITY Person URI: MMC#RTBF#67554778Name : BartheFirst name: FranoisENTITY Person URI: MMC#RTBF#6753Name : ZidaneFirst name: Zinedine Works for Talk about Enrichment3Enhancementsemantic2Table Person Id : #67554778Name : BartheFirst name: FranoisEmployer : RTBFDescription : > Worked on Zidanesbio.1ENTITY Person URI: MMC#RTBF#67554778Name : BartheFirst name : FranoisDescription : > Worked on Zidanes bio.ENTITY Organisation URI: MMC#RTBF#6756593Name : RTBF Workfor for semanticEnrichmentsemantic3Le 09/10/2013 15BIG DATA - BIG MEDIAInferencePerfect Memorys data bases : Increasing of amount of information in time, Increasing of quality of links in time. Preserve, enrich and sharing of knowledge.Capitalisation of knowledgeSemanticInference4Semantic negentropic DBNews factsInference rules16BIG DATA - BIG MEDIAExploiting the knowledgeExploiting the knowledgeFrom Pr. Bachimont University of Technology of CompigneLe 09/10/2013 17BIG DATA - BIG MEDIALinked Open DataConsolidating the distributed knowledgeRTBF18BIG DATA - BIG MEDIAProcess managementOSB workersInterOp-WindowA service on the OSB1. Manager: Identification of request2. Manager: Main process instantation3. Manager: Sub process instantiation4. Manager :Tasks instantiation5. Manager: IOW calls6. Guichets#1 : Execution of tasks and works7. 8. Guichets#n : Execution of tasks and worksTreatment request8. Guichets#n : Execution of tasks and works19BIG DATA - BIG MEDIASemantic PlayerRendering the semantic linksInterOp-GuichetPlayer SmantiqueNetworkNetworkOSB OSBOSB OSBEnhancement, Repurposing & Exploitation of audiovisual contents.20BIG DATA - BIG MEDIAAn architecture scalable, distributedIntroducing the Semantic Middleware approachSEMANTICPLAYER(1) The Heart inludes the Knowledge base features, and the OAIS functionalities(1)(2)YOURAPPLICATION(2) The BUS, 100% compliant to EBUCore, becomes the backboneof the middleware(3) Any Bases ingested, or functionalitiesconnected via an InteOperability Windows (GIO) becomes a semantic ressource for the Middleware(2)(3)21BIG DATA - BIG MEDIAFlexibility & scalability of the middlewareControl EnrichmentExtractionExpressivityLe 09/10/2013 22BIG DATA - BIG MEDIARTBF Annotation & search interfacesFeatures: Breakthrough user friendly interface for big data visualization Graphical browsing in big data content (media and metadata)23BIG DATA - BIG MEDIARadio France Tablet InterfaceConnection: Building the contextualization of the display according to the Role and Skill ofthe connected user.24Big Data - Big MediaThe PROFILE : knowledge capitalizationYourYour Data StructureData StructureYourYour Media LibraryMedia Library Linked Open DataLinked Open DataYOURYOUR KNOWLEDGEKNOWLEDGE25BIG DATA - BIG MEDIAThe Middleware features Automatic linking with external related contents,BeforeBeforeAfter Automatic knowledge validation, Cross-browsing in broadcasters MAMs.AfterMedia Processing26BIG DATA - BIG MEDIAThe semantic middleware approachLe 09/10/2013 BIG DATA - BIG MEDIA 27BIG DATA BIG MEDIATHE NEW PARADIGM OF THE MULTIMEDIA CONTENT MANAGEMENTFrdric ColominaBusiness DevelopmentFrederic.colomina@perfect-memory.com@perfect__memorySteny SolitudeCEOSteny.solitude@perfect-memory.com@perfect__memorySemtech top 10 startup of 2013 IBC Award for Content management 2013IBC Award for technology Who caught my eye looking for blue skies of IBC 2013@perfect__memory http://perfect-memory.comOrganizadores Sponsor platinumSponsor Gold Con el apoyo de Socio tecnolgicoNicolas Moulard, Director de Actuondamoulard@actuonda.comTel : +34 699 248 200@Radio_20 www.bigmediaconnect.es