overview ag aksw
DESCRIPTION
TRANSCRIPT
Vorstellung Forschungsgruppe Agile Knowledge Engineering &
Semantic Web (AKSW)Dr. Sören Auer
Research Group “Agile Knowledge Engineering and Semantic Web”
Founded 2006Initially hosted by the chair for Business Information Systems (Prof. Klaus-Peter
Fähnrich)
Now transition to Institute for Applied Informatics (InfAI)• An-institute at Universität Leipzig• Combines competences and resources of 8 University chairs from Computer
Science and Economics faculties as well as industry and sponsors
AKSW aims:• Contributing to the advancement of science in Semantic Web, Knowledge
Engineering, Software Engineering• Cost efficient, high-impact R&D, which proves usefulness at an early stage• Bridge the gap between research results and applications
AKSW actively educates students in Semantic Technologies and serves the community by (co-) organizing events such as Conference on Social Semantic Web, I-Semantics, Scripting for The Semantic Web workshop series etc.AKSW Vorstellung
AKSW Team
• Dr. Sören Auer, Head, everything currently especially DBpedia, Cofundos, Triplify• Thomas Riechert, wiss. Mitarbeiter, Software Engineering and teaching• Jens Lehmann, doctoral student, DL-Learner, DBpedia, Machine Learning• Sebastian Dietzold, doctoral student, OntoWiki, xOperator, RDF-LDAP and Data integration• Axel Ngonga, doctoral student (2006), Text Mining, Knowledge Management• Thorsten Berger, doctoral student (2007), Software Engineering• Michael Martin, doctoral student (2008), Semantic Web Applications• Sebastian Hellmann, doctoral student (2008), Machine Learning• Jörg Unbehauen, doctoral student (2008), Software EngineeringAlumni• Dr.-Ing. Muhammad Ahtisham Aslam, Ass. Prof. COMSATS Institute of Information Technology (CIIT),
Lahore, PakistanPermanently ca. 10 student assistants, bachelor/master/diplom studentsAKSW Vorstellung
AKSW Funded ProjectsSCMS: – Semantic Content Management Systems for Enterprise Knowledge Management and News Mining• Cooperative research project; 24 month / 2009-2011• Funding agency: Eurostars• Participants: Semantic Web Company, Digital Trowel, OpenLink Software Ltd., NetresearchOntoWiki: – Semantic Collaboration for Knowledge Management, E-Learning and E-Tourism• Cooperative research project; 24 month / 2008-2010• Funding agency: European Union FP7 / Research for the benefit of the SME program• Participants: OpenLink Software Ltd., Business Intelligence GmbH, B2 d.o.o., VakantielandLE4SW - Regionale Technologieplattform OntoWiki für soziale, semantische Kollaboration• Cooperative research project; 24 month / 2009-2011• Funding agency: BmbF (German Ministry for Education and Research), Programme “Regionale
Wachstumskerne / Potential”• Participants: Universität Leipzig, Business Intelligence GmbH, Netresearch GmbH & Co. KG, Ebrosia GmbHSoftWiki: End-user driven, distributed Requirements Engineering for agile Software Development• Cooperative research project; 42 months / 2006-2009• Funding agency: BmbF (German Ministry for Education and Research)• Participants: Universität Duisburg-Essen, T-Systems MMS, ProDV AG, LeCoS GmbH, QA Systems GmbH, ISA
Tools GmbHVakantieland – Semantic Collaboration Platform for Tourist Information• Industry / public funding; 36 months / 2006-2008• Funding agency: SenterNovem (Dutch Ministry of Economic Affairs)• Participants: Universität Leipzig, Vakantieland
AKSW Vorstellung
Impact beyond the Scientific Community
DBpedia.org – knowledge extraction from Wikipedia• Impact: 300 posts about DBpedia in the Blogosphere (according to Technorati), ca. 100.000
visitors of the DBpedia website in 2007, on average 350 daily visitors. DBpedia became the most popular dataset used with Semantic Web applications and research prototypes and is for example extensively referred in official W3C standards such as RDFa.
Cofundos.org – open-source innovation and resource pooling• Impact: more than 10.000 visitors in Oct 2007, 62 Blog posts (according to Technorati), Cofundos
news announcements in major news channels (e.g. Heise.de, Golem.de, Linux.com), more than 300 registered users, 100 projects, 10k€ pledged donations.
OntoWiki.net – social, semantic collaboration platform• Impact: more than 8.000 downloads of the OntoWiki software (since 2004) / 773 in Sep 2007, on
average 3.000 monthly visitors of http://Ontowiki.net, users include SAP, OpenLink SW, T-Systems MMS, ProDV AG.
Triplify.org - “semantification” of Web applications• Impact: more than 1.000 visitors on the Triplify.org Website in the first week alone, Numerous
blog posts e.g. in the no 1 Web technology blog ReadWriteWeb, Triplify configurations for major Web applications such as Drupal, Joomla!, Wordpress, WackoWiki
AKSW Vorstellung
Ausgewählte Partner und ProjekteLinkedGeoData.org
AKSW Vorstellung
Semantic Leipzig
Knowledge ManagementLogical Foundations & ReasoningService Engineering & ManagementMachine Learning & Text Mining
Semantic Web InfrastructureSemantic SearchSocial Software & Web 2.0eGovernment
Applied ResearchTechnology TransferBusiness Scenarios
Applied ResearchProduct Development
Basic ResearchApplied Research
AKSW Vorstellung
DBpedia“Semantification” of Wikipedia
AKSW Linked Data Web Bausteine
AKSW Vorstellung
Triplify“Semantification” of (small) Web Applications
OntoWikiCollaborative creation of explicit knowledge via Semantic Wikis
OWLDBExtending DBs for ontology handling / revealing implicit information
VakantielandBuilding Data Web applications
SoftWikiDistributed, stakeholder driven Requirements Engineering
GrundlagenMarrying databases with RDFand ontologies
WerkzeugeAnwendungenBringing the Data Web to end users
RDF Query Subsumption & View MaintenanceScaling database backed Triple Stores
xOperatorCombining Instant Messaging with the Data Web
OpenResearch.orgA semantic Wiki for the sciences
…
DL-LearnerMachine Learning for Ontologies
The Semantic Data Wiki• Agile, distributed knowledge engineering• Not a Wiki with semantic extensions (Semantic MediaWiki,
IkeWiki), but an ontology editor using Wiki• Concepts:
– Make it easy tocorrect mistakes(ant intelligence)
– Activity can bewatched andreviewed
– Everything canbe undone
AKSW Vorstellung
SoftWiki
AKSW Vorstellung
Problem: Requirements Engineering with large, spatially distributed stakeholder groups
Solution: comprehensive ontology for representing RE relevant knowledge + adapted OntoWiki application
Application of text-miningmethods for duplicate detection
Lohmann, Heim, Auer, Dietzold, Riechert: Semantifying Requirements Engineering – The SoftWiki Approach. In I-SEMANTICS 2008.
Vakantieland*One of the largest tourist information sites in NL
(>100.000 daily page views, >20.000 points of interest)Traditional relational DB system was to inflexible to capture the increasingly
heterogeneous content types• Development of an OntoWiki based Data Web application• Geo-data integration from OpenStreetMaps• Semantic-Search• Integration of
DBpedia data• Comprehensive
performance tuning
* work with Ceriel Jakobs,Michael Martin partiallyfunded by SenterNovem
AKSW Vorstellung
OpenResearch.org – Semantic Wiki for the SciencesBased on SMWSupport for scientific
content types• Events (Conferences,
Workshops, etc.)• People, research
groups, sciencegenealogy
• Journals• Funding callsAdditional categorization schemes include scientific field (not limited to CS) and
location/regionSemantic annotation and structuring of these facilitate search (e.g. SE conferences by
acceptance rate)Already one of the largest KB’s of science meta-information more than 7.000
pages/entitiesAKSW Vorstellung
xOperator–connecting IM & Data WebSemantic overlay network for Instant messagingNaturally solves some provenance, trust issues and
context awareness
AKSW Vorstellung
Dietzold, Unbehauen, Auer: xOperator - Interconnecting the Semantic Web and Instant Messaging Networks. In ESWC 2008.
• Knowledge base derived from Wikipedia• One of the largest ontologies• Multi-domain, multi-language• Joint work with FU Berlin and OpenLink• Extract RDF/OWL from Wikipedia, e.g. From
Infoboxes, categories, Geo-Coordinates, Images, ...• 274 million triples, 213.000 persons, 328.000
places, 57.000 music albums, 36.000 films, 20.000 companies
• 2500 manual mappings for infobox attributes to DBpedia Ontology (175 classes, 384 object properties, 336 data properties)
AKSW Vorstellung
Auer; Bizer, Lehmann, Kobilarov, Cyganiak, Ives: DBpedia: A Nucleus for a Web of Open Data. In ISWC 2007.Bizer, Lehmann, Kobilarov, Auer, Becker, Cyganiak, Hellmann: DBpedia - A Crystallization Point for the Web of Data. In Journal of Web Semantics, 2009.
Extraction Job
Extraction Manager
PageCollections
Destinations
N-TripleDumps
WikipediaDumps
WikipediaOAI-PMH
Database Wikipedia
LiveWikipedia
N-TripleSerializer
SPARQL-UpdateDestination
Extractors
Generic Infobox
Label
Geo
Redirect Disambiguation
Image
Abstract Pagelink
Parsers
DateTime Units
Ontology-Mappings
Mapping-based Infobox
String-List Numbers
Geo
SPARQL endpoint Linked Data
The Web RDF browser HTML browserSPARQL clientsDBpedia apps
Triple StoreVirtuoso
UpdateStream
Article-Queue
Wikipedia
Category
AKSW Vorstellung
Triplify
AKSW Vorstellung
Auer, Dietzold, Aumueller, Lehmann, Hellmann: Triplify - Light-weight Linked Data Publication from Relational Databases. In WWW 2009.
Triplify
Relational Database
Web Browser Keyword-basedSearch Engines
Web Application
Semantic-basedSearch Engines
HTML pagesRDF triple-based descriptions(Linked Data, RDF, JSON)
Triplify script
Endpoint registry
Configuration repository
Webserver
Triplify
AKSW Vorstellung
LinkedGeoData.orgHow to publish geo-data using Triplify?
http://linkedgeodata.org/near/48.213056,16.359722/1000/amenity=Hotel
http://linkedgeodata.org/node/212331http://linkedgeodata.org/node/944523http://linkedgeodata.org/node/234091http://linkedgeodata.org/way/56719
node/150760824 amenity "pub"; created_by "JOSM"; distance "5995";name "La friolera";geo#lat "40.4474";geo#long "-3.7173".
AKSW Vorstellung
Lon Lat RadiusAttribute Value
Faceted Linked-Geo-Data Browser
AKSW Vorstellung
Ontology Learning (DL-Learner)• Framework for Supervised
Machine Learning for OWL and Description Logics
• Application Areas:– “Classical” Machine Learning,
e.g. predicting Carcinogenesis– Ontology Engineering– recommendation/navigation
• Works on OWL Files and SPARQL Endpoints
• Supports different reasoner interfaces
• Accessible via command-line, GUI, web service
AKSW Vorstellung
Hellmann, Lehmann, Auer: Learning of OWL Class Descriptions on Very Large Knowledge Bases. International Journal On Semantic Web and Information Systems, 2009.Lehmann, Hitzler: Concept Learning in Description Logics. Machine Learning Journal, 2009
Participatory Research Idea• Engaging the wisdom of the crowds for research
project definition and assessment• Outsource idea evaluation and progress review
to the stakeholder community• Organize research funding like an information
market – the best known instruments for aggregating (asymmetrically distributed) information
• Facilitate involvement of private endowments, foundations, individuals
AKSW Vorstellung
Auer, Braun-Thürmann: Towards Bottom-Up, Stakeholder-Driven Research Funding – Open Source Funding, Open Peer Review. In Peer Review Reviewed: The International Career of a Quality-control Instrument and New Challenges, 2008.
Open Science Platform Concept• Research ideas: published by Researchers or SMEs on an open-science Web
platform as early as possible• Project definition:
– All participants of the platform, i.e. researchers and stakeholders (e.g. SMEs, NGOs) are equipped with a virtual cash budget for pledging
– Stakeholders comment on the ideas, add requirements and pledge a certain amount of money they would be willing to “pay” for a successful realization.
• Selection: A funding agency can select and fund the highest ranked proposals in a certain area or application domain
• Project runs: involved investigators report publicly (e.g. Weblog) => enables stakeholders to influence the projects (e.g. changed requirements or alternative approaches appear)
• Results are published, everybody is invited to comment on the success, only the stakeholders (i.e. those who pledged) are eligible to vote about the success / write an evaluation report (publically available => track record of a researcher)
AKSW Vorstellung
Cofundos.org• Application of the concept for open-
source software development• Funding is provided by individuals• Stakeholder community interested in a
certain software (feature) decides collaboratively about requirements, who to entrust, project success
• All contributions licensed under Creative Commons
• Based on: Reputation & community, fairness & trust, open-knowledge & open source, iterative methodology
AKSW Vorstellung
Sören Auer: Endanwendergetriebene Open-source Softwareentwicklung mit Cofundos. In Open Source Jahrbuch 2008.
Portal
• Web app• Cofundos.org• Uses tagging/
folksonomies,OpenID, RSSfeeds, AJAX,REST,Semantic Webtechnologies
AKSW Vorstellung
DBpedia“Semantification” of Wikipedia
AKSW Linked Data Web Bausteine
AKSW Vorstellung
Triplify“Semantification” of (small) Web Applications
OntoWikiCollaborative creation of explicit knowledge via Semantic Wikis
OWLDBExtending DBs for ontology handling / revealing implicit information
VakantielandBuilding Data Web applications
SoftWikiDistributed, stakeholder driven Requirements Engineering
GrundlagenMarrying databases with RDFand ontologies
WerkzeugeAnwendungenBringing the Data Web to end users
RDF Query Subsumption & View MaintenanceScaling database backed Triple Stores
xOperatorCombining Instant Messaging with the Data Web
OpenResearch.orgA semantic Wiki for the sciences
…
DL-LearnerMachine Learning for Ontologies