A Semantic-enhanced Inference Framework for Heterogeneous Resources Management
1 like158 views
This document is the PhD dissertation of Silvia Giannini presented to the Dipartimento di Ingegneria Elettrica e dell'Informazione (DEI) at Politecnico di Bari in Italy. The dissertation addresses semantic integration of heterogeneous web resources and proposes using OWL as the representation language to manage these resources. It explores using inference services over the unified framework and how to exploit these services for heterogeneous resource management.
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
Common Subsumers: A logic-based approach
Example: Automatically extract Core Competence, by identifying a common
know-how in a company personnel [1]
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-22-320.jpg)
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
Common Subsumers: the Knowledge Compilation process
Example: Automatically extract Core Competence, by identifying a common
know-how in a company personnel [1]
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-23-320.jpg)
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
Common Subsumers: the Knowledge Compilation process
Example: Automatically extract Core Competence, by identifying a common
know-how in a company personnel [1]
Issues:
Computational diculties of deduction in knowledge bases expressed
through a logical formalism;
Combining the representation power of a logical language, with the
scalability and eciency of information processing in a DBMS.
Knowledge Compilation:
1 OFF-LINE REASONING
pre-processing of a company intellectual capital, described in a Description
Logics (DLs) Knowledge Base (KB), in an appropriate relational database
schema.
2 ON-LINE REASONING
querying of the data structure coming out from the rst phase through
standard SQL-queries for ecient Core Competence Extraction.
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-24-320.jpg)
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
I.M.P.A.K.T.3
Example: Automatically extract Core Competence, by identifying a common
know-how in a company personnel [1]
Knowledge Base
Mario Rossi: Cplusplus (5 years), Java (5 years), Visual Basic (5 years)
Daniela Bianchi: Cplusplus (2 years), Java (6 years), Visual Basic (1 years)
Elena Pomarico: CplusPlus, Java, Visual Basic
Carmelo Piccolo: VBScript, Process Performance Monitoring
Lucio Battista: DBMS (2 years)
Mariangela Porro: DBMS (2 years), Internet Technologies (2 years)
Nicola Marco: DBMS (5 years), Internet Technologies (5 years)
Domenico De Palo: OOprogramming (6 years), Articial intelligence (4 years), Internet technologies (4
years)
3
Information Management and Processing with the Aid of Knowledge-based Technologies
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-25-320.jpg)
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
Solution
Adaptation to RDF assertions:
adaptation of path and connectedness denition from Graph Theory to
RDF-graphs
denition of rooted RDF-graphs (r-graph)
denition of entailment between r-graphs
denition of Common Subsumer of pairs of RDF resources
Denition (Rooted Entailment)
Let R ∈ {S, RDF, RDF-S} be an entailment relation. R-graph r, Tr R-entails
s, Ts denoted by r, Tr |=R s, Ts when the following conditions hold:
1 if s is a blank node, then
1 if r is not a blank node, Tr |=R Ts[s → r] must hold;
2 if also r is a blank node, then Tr[r → u] |=R Ts[s → u] for a new URI u
occurring neither in Tr nor in Ts;
2 otherwise (i.e., s is not a blank node), if s = r, then Tr |=R Ts must hold;
3 otherwise (i.e., s is not a blank node and s = r) r, Tr never R-entails
s, Ts .
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-38-320.jpg)
![Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion
RDF Clustering
(c) How do we exploit them for managing heterogeneous web resources?
Clustering of Web resources with CS
Retrieving resources conveying the same information in their dierent RDF
descriptions1 Discover homogeneous groups of resources
2 Identify a cluster concept description
CS description → SPARQL queries:
WHERE { Tcs [blank nodes → variables] }
Silvia Giannini Ph.D. Final Dissertation](https://image.slidesharecdn.com/gianniniphddissertation-180224192201/85/A-Semantic-enhanced-Inference-Framework-for-Heterogeneous-Resources-Management-44-320.jpg)
A Semantic-enhanced Inference Framework for Heterogeneous Resources Management
- 1. Dottorato di Ricerca in Ingegneria Elettrica e dell'Informazione ING-INF/05 Ciclo XXVII A Semantic-enhanced Inference Framework for Heterogeneous Resources Management Ph.D. Final Dissertation Candidate: Silvia Giannini Tutor: Prof. Eugenio Di Sciascio Co-Tutor: Dr. Ph.D. Simona Colucci Co-ordinator of the Doctorate Course: Prof. Michele Antonio Trovato Dipartimento di Ingegneria Elettrica e dell'Informazione (DEI) Politecnico di Bari, Bari, Italy | 24 April 2015
- 2. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 3. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 4. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Scenario: Heterogeneous Resources Autonomous and independent information sources Heterogeneous schema and data-models Vocabulary Syntax Semantics User requests VS User satisfaction Silvia Giannini Ph.D. Final Dissertation
- 5. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Scenario: Semantic-enhanced Framework Semantic Web technologies Integration paradigms Adding structure to web resources Schema/ontology matching Hybrid strategies (GAV/LAV approaches) Scalability issues Silvia Giannini Ph.D. Final Dissertation
- 6. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Research Questions (a) What is the best representation language for the semantic integration of heterogeneous web resources? (b) Which inference-services can be enabled over such a unied framework? (c) How do we exploit them for managing heterogeneous web resources? Silvia Giannini Ph.D. Final Dissertation
- 7. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 8. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion The Web Ontology Language (OWL) (a) What is the best representation language for the integration of heterogeneous web resources? RDF is the language of the Web of Data OWL manages complex knowledge structures in well-dened domain Silvia Giannini Ph.D. Final Dissertation
- 9. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion A Semantic-enhanced Framework for Heterogeneous Resource Management (a) What is the best representation language for the integration of heterogeneous web resources? Silvia Giannini Ph.D. Final Dissertation
- 10. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion A Semantic-enhanced Framework for Heterogeneous Resource Management (a) What is the best representation language for the integration of heterogeneous web resources? Silvia Giannini Ph.D. Final Dissertation
- 11. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion RDF: the big picture DBpedia1 extract Graph-structured knowledge representation (data-model) Facts in Triple-form: subject - predicate - object 1 http://dbpedia.org Silvia Giannini Ph.D. Final Dissertation
- 12. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion RDF: the big picture DBpedia1 extract Graph-structured knowledge representation (data-model) http://dbpedia.org/resource/Bari http://dbpedia.org/ontology/country http://dbpedia.org/resource/Italy. 1 http://dbpedia.org Silvia Giannini Ph.D. Final Dissertation
- 13. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion RDF: the big picture DBpedia1 extract RDF Schema: Explicit semantics of nodes and links labels http://dbpedia.org/resource/Italy http://www.w3.org/1999/02/22-rdf-syntax-ns#type http://dbpedia.org/ontology/Country. 1 http://dbpedia.org Silvia Giannini Ph.D. Final Dissertation
- 14. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Linked Open Data (LOD) project Linked Datasets as of August 2014 Uniprot Alexandria Digital Library Gazetteer lobid Organizations chem2 bio2rdf Multimedia Lab University Ghent Open Data Ecuador Geo Ecuador Serendipity UTPL LOD GovAgriBus Denmark DBpedia live URI Burner Linguistics Social Networking Life Sciences Cross-Domain Government User-Generated Content Publications Geographic Media Identifiers Eionet RDF lobid Resources Wiktionary DBpedia Viaf Umthes RKB Explorer Courseware Opencyc Olia Gem. Thesaurus Audiovisuele Archieven Diseasome FU-Berlin Eurovoc in SKOS DNB GND Cornetto Bio2RDF Pubmed Bio2RDF NDC Bio2RDF Mesh IDS Ontos News Portal AEMET ineverycrea Linked User Feedback Museos Espania GNOSS Europeana Nomenclator Asturias Red Uno Internacional GNOSS Geo Wordnet Bio2RDF HGNC Ctic Public Dataset Bio2RDF Homologene Bio2RDF Affymetrix Muninn World War I CKAN Government Web Integration for Linked Data Universidad de Cuenca Linkeddata Freebase Linklion Ariadne Organic Edunet Gene Expression Atlas RDF Chembl RDF Biosamples RDF Identifiers Org Biomodels RDF Reactome RDF Disgenet Semantic Quran IATI as Linked Data Dutch Ships and Sailors Verrijktkoninkrijk IServe Arago- dbpedia Linked TCGA ABS 270a.info RDF License Environmental Applications Reference Thesaurus Thist JudaicaLink BPR OCD Shoah Victims Names Reload Data for Tourists in Castilla y Leon 2001 Spanish Census to RDF RKB Explorer Webscience RKB Explorer Eprints Harvest NVS EU Agencies Bodies EPO Linked NUTS RKB Explorer Epsrc Open Mobile Network RKB Explorer Lisbon RKB Explorer Italy CE4R Environment Agency Bathing Water Quality RKB Explorer Kaunas Open Data Thesaurus RKB Explorer Wordnet RKB Explorer ECS Austrian Ski Racers Social- semweb Thesaurus Data Open Ac Uk RKB Explorer IEEE RKB Explorer LAAS RKB Explorer Wiki RKB Explorer JISC RKB Explorer Eprints RKB Explorer Pisa RKB Explorer Darmstadt RKB Explorer unlocode RKB Explorer Newcastle RKB Explorer OS RKB Explorer Curriculum RKB Explorer Resex RKB Explorer Roma RKB Explorer Eurecom RKB Explorer IBM RKB Explorer NSF RKB Explorer kisti RKB Explorer DBLP RKB Explorer ACM RKB Explorer Citeseer RKB Explorer Southampton RKB Explorer Deepblue RKB Explorer Deploy RKB Explorer Risks RKB Explorer ERA RKB Explorer OAI RKB Explorer FT RKB Explorer Ulm RKB Explorer Irit RKB Explorer RAE2001 RKB Explorer Dotac RKB Explorer Budapest Swedish Open Cultural Heritage Radatana Courts Thesaurus German Labor Law Thesaurus GovUK Transport Data GovUK Education Data Enakting Mortality Enakting Energy Enakting Crime Enakting Population Enakting CO2Emission Enakting NHS RKB Explorer Crime RKB Explorer cordis Govtrack Geological Survey of Austria Thesaurus Geo Linked Data Gesis Thesoz Bio2RDF Pharmgkb Bio2RDF Sabiork Bio2RDF Ncbigene Bio2RDF Irefindex Bio2RDF Iproclass Bio2RDF GOA Bio2RDF Drugbank Bio2RDF CTD Bio2RDF Biomodels Bio2RDF DBSNP Bio2RDF Clinicaltrials Bio2RDF LSR Bio2RDF Orphanet Bio2RDF Wormbase BIS 270a.info DM2E DBpedia PT DBpedia ES DBpedia CS DBnary Alpino RDF YAGO Pdev Lemon Lemonuby Isocat Ietflang Core KUPKB Getty AAT Semantic Web Journal OpenlinkSW Dataspaces MyOpenlink Dataspaces Jugem Typepad Aspire Harper Adams NBN Resolving Worldcat Bio2RDF Bio2RDF ECO Taxon- concept Assets Indymedia GovUK Societal Wellbeing Deprivation imd Employment Rank La 2010 GNU Licenses Greek Wordnet DBpedia CIPFA Yso.fi Allars Glottolog StatusNet Bonifaz StatusNet shnoulle Revyu StatusNet Kathryl Charging Stations Aspire UCL Tekord Didactalia Artenue Vosmedios GNOSS Linked Crunchbase ESD Standards VIVO University of Florida Bio2RDF SGD Resources Product Ontology Datos Bne.es StatusNet Mrblog Bio2RDF Dataset EUNIS GovUK Housing Market LCSH GovUK Transparency Impact ind. Households In temp. Accom. Uniprot KB StatusNet Timttmy Semantic Web Grundlagen GovUK Input ind. Local Authority Funding From Government Grant StatusNet Fcestrada JITA StatusNet Somsants StatusNet Ilikefreedom Drugbank FU-Berlin Semanlink StatusNet Dtdns StatusNet Status.net DCS Sheffield Athelia RFID StatusNet Tekk Lista Encabeza Mientos Materia StatusNet Fragdev Morelab DBTune John Peel Sessions RDFize last.fm Open Data Euskadi GovUK Transparency Input ind. Local auth. Funding f. Gvmnt. Grant MSC Lexinfo StatusNet Equestriarp Asn.us GovUK Societal Wellbeing Deprivation Imd Health Rank la 2010 StatusNet Macno Oceandrilling Borehole Aspire Qmul GovUK Impact Indicators Planning Applications Granted Loius Datahub.io StatusNet Maymay Prospects and Trends GNOSS GovUK Transparency Impact Indicators Energy Efficiency new Builds DBpedia EU Bio2RDF Taxon StatusNet Tschlotfeldt Jamendo DBTune Aspire NTU GovUK Societal Wellbeing Deprivation Imd Health Score 2010 Lotico GNOSS Uniprot Metadata Linked Eurostat Aspire Sussex Lexvo Linked Geo Data StatusNet Spip SORS GovUK Homeless- ness Accept. per 1000 TWC IEEEvis Aspire Brunel PlanetData Project Wiki StatusNet Freelish Statistics data.gov.uk StatusNet Mulestable Enipedia UK Legislation API Linked MDB StatusNet Qth Sider FU-Berlin DBpedia DE GovUK Households Social lettings General Needs Lettings Prp Number Bedrooms Agrovoc Skos My Experiment Proyecto Apadrina GovUK Imd Crime Rank 2010 SISVU GovUK Societal Wellbeing Deprivation Imd Housing Rank la 2010 StatusNet Uni Siegen Opendata Scotland Simd Education Rank StatusNet Kaimi GovUK Households Accommodated per 1000 StatusNet Planetlibre DBpedia EL Sztaki LOD DBpedia Lite Drug Interaction Knowledge Base StatusNet Qdnx Amsterdam Museum AS EDN LOD RDF Ohloh DBTune artists last.fm Aspire Uclan Hellenic Fire Brigade Bibsonomy Nottingham Trent Resource Lists Opendata Scotland Simd Income Rank Randomness Guide London Opendata Scotland Simd Health Rank Southampton ECS Eprints FRB 270a.info StatusNet Sebseb01 StatusNet Bka ESD Toolkit Hellenic Police StatusNet Ced117 Open Energy Info Wiki StatusNet Lydiastench Open Data RISP Taxon- concept Occurences Bio2RDF SGD UIS 270a.info NYTimes Linked Open Data Aspire Keele GovUK Households Projections Population W3C Opendata Scotland Simd Housing Rank ZDB StatusNet 1w6 StatusNet Alexandre Franke Dewey Decimal Classification StatusNet Status StatusNet doomicile Currency Designators StatusNet Hiico Linked Edgar GovUK Households 2008 DOI StatusNet Pandaid Brazilian Politicians NHS Jargon Theses.fr Linked Life Data Semantic Web DogFood UMBEL Openly Local StatusNet Ssweeny Linked Food Interactive Maps GNOSS OECD 270a.info Sudoc.fr Green Competitive- ness GNOSS StatusNet Integralblue WOLD Linked Stock Index Apache KDATA Linked Open Piracy GovUK Societal Wellbeing Deprv. Imd Empl. Rank La 2010 BBC Music StatusNet Quitter StatusNet Scoffoni Open Election Data Project Reference data.gov.uk StatusNet Jonkman Project Gutenberg FU-Berlin DBTropes StatusNet Spraci Libris ECB 270a.info StatusNet Thelovebug Icane Greek Administrative Geography Bio2RDF OMIM StatusNet Orangeseeds National Diet Library WEB NDL Authorities Uniprot Taxonomy DBpedia NL L3S DBLP FAO Geopolitical Ontology GovUK Impact Indicators Housing Starts Deutsche Biographie StatusNet ldnfai StatusNet Keuser StatusNet Russwurm GovUK Societal Wellbeing Deprivation Imd Crime Rank 2010 GovUK Imd Income Rank La 2010 StatusNet Datenfahrt StatusNet Imirhil Southampton ac.uk LOD2 Project Wiki DBpedia KO Dailymed FU-Berlin WALS DBpedia IT StatusNet Recit Livejournal StatusNet Exdc Elviajero Aves3D Open Calais Zaragoza Turruta Aspire Manchester Wordnet (VU) GovUK Transparency Impact Indicators Neighbourhood Plans StatusNet David Haberthuer B3Kat Pub Bielefeld Prefix.cc NALT Vulnera- pedia GovUK Impact Indicators Affordable Housing Starts GovUK Wellbeing lsoa Happy Yesterday Mean Flickr Wrappr Yso.fi YSA Open Library Aspire Plymouth StatusNet Johndrink Water StatusNet Gomertronic Tags2con Delicious StatusNet tl1n StatusNet Progval Testee World Factbook FU-Berlin DBpedia JA StatusNet Cooleysekula Product DB IMF 270a.info StatusNet Postblue StatusNet Skilledtests Nextweb GNOSS Eurostat FU-Berlin GovUK Households Social Lettings General Needs Lettings Prp Household Composition StatusNet Fcac DWS Group Opendata Scotland Graph Simd Rank DNB Clean Energy Data Reegle Opendata Scotland Simd Employment Rank Chronicling America GovUK Societal Wellbeing Deprivation Imd Rank 2010 StatusNet Belfalas Aspire MMU StatusNet Legadolibre Bluk BNB StatusNet Lebsanft GADM Geovocab GovUK Imd Score 2010 Semantic XBRL UK Postcodes Geo Names EEARod Aspire Roehampton BFS 270a.info Camera Deputati Linked Data Bio2RDF GeneID GovUK Transparency Impact Indicators Planning Applications Granted StatusNet Sweetie Belle O'Reilly GNI City Lichfield GovUK Imd Rank 2010 Bible Ontology Idref.fr StatusNet Atari Frosch Dev8d Nobel Prizes StatusNet Soucy Archiveshub Linked Data Linked Railway Data Project FAO 270a.info GovUK Wellbeing Worthwhile Mean Bibbase Semantic- web.org British Museum Collection GovUK Dev Local Authority Services Code Haus Lingvoj Ordnance Survey Linked Data Wordpress Eurostat RDF StatusNet Kenzoid GEMET GovUK Societal Wellbeing Deprv. imd Score '10 Mis Museos GNOSS GovUK Households Projections total Houseolds StatusNet 20100 EEA Ciard Ring Opendata Scotland Graph Education Pupils by School and Datazone VIVO Indiana University Pokepedia Transparency 270a.info StatusNet Glou GovUK Homelessness Households Accommodated Temporary Housing Types STW Thesaurus for Economics Debian Package Tracking System DBTune Magnatune NUTS Geo- vocab GovUK Societal Wellbeing Deprivation Imd Income Rank La 2010 BBC Wildlife Finder StatusNet Mystatus Miguiad Eviajes GNOSS Acorn Sat Data Bnf.fr GovUK imd env. rank 2010 StatusNet Opensimchat Open Food Facts GovUK Societal Wellbeing Deprivation Imd Education Rank La 2010 LOD ACBDLS FOAF- Profiles StatusNet Samnoble GovUK Transparency Impact Indicators Affordable Housing Starts StatusNet CoreyavisEnel Shops DBpedia FR StatusNet Rainbowdash StatusNet Mamalibre Princeton Library Findingaids WWW Foundation Bio2RDF OMIM Resources Opendata Scotland Simd Geographic Access Rank Gutenberg StatusNet Otbm ODCL SOA StatusNet Ourcoffs Colinda Web Nmasuno Traveler StatusNet Hackerposse LOV Garnica Plywood GovUK wellb. happy yesterday std. dev. StatusNet Ludost BBC Program- mes GovUK Societal Wellbeing Deprivation Imd Environment Rank 2010 Bio2RDF Taxonomy Worldbank 270a.info OSM DBTune Music- brainz Linked Mark Mail StatusNet Deuxpi GovUK Transparency Impact Indicators Housing Starts Bizkai Sense GovUK impact indicators energy efficiency new builds StatusNet Morphtown GovUK Transparency Input indicators Local authorities Working w. tr. Families ISO 639 Oasis Aspire Portsmouth Zaragoza Datos Abiertos Opendata Scotland Simd Crime Rank Berlios StatusNet piana GovUK Net Add. Dwellings Bootsnall StatusNet chromic Geospecies linkedct Wordnet (W3C) StatusNet thornton2 StatusNet mkuttner StatusNet linuxwrangling Eurostat Linked Data GovUK societal wellbeing deprv. imd rank '07 GovUK societal wellbeing deprv. imd rank la '10 Linked Open Data of Ecology StatusNet chickenkiller StatusNet gegeweb Deusto Tech StatusNet schiessle GovUK transparency impact indicators tr. families Taxon concept GovUK service expenditure GovUK societal wellbeing deprivation imd employment score 2010 Linking Open Data cloud diagram (Richard Cyganiak and Anja Jentzsch, http://lod-cloud.net/) Datasets: 1014 as of April 2014 Silvia Giannini Ph.D. Final Dissertation
- 15. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Semantic Web and LOD (b) Which inference-services can be enabled over such a unied framework? Silvia Giannini Ph.D. Final Dissertation
- 16. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Reasoning on RDF (b) Which inference-services can be enabled over such a unied framework? Basic reasoning rules in four entailment regimes2 : Simple Entailment RDF Entailment RDFS Entailment D-Entailment 2 P.J. Hayes, P.F. Patel-Schneider, http://www.w3.org/TR/rdf11-mt/ Silvia Giannini Ph.D. Final Dissertation
- 17. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Reasoning on RDF (b) Which inference-services can be enabled over such a unied framework? Basic reasoning rules in four entailment regimes2 : Simple Entailment RDF Entailment RDFS Entailment D-Entailment 2 P.J. Hayes, P.F. Patel-Schneider, http://www.w3.org/TR/rdf11-mt/ Silvia Giannini Ph.D. Final Dissertation
- 18. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Reasoning on RDF (b) Which inference-services can be enabled over such a unied framework? Basic reasoning rules in four entailment regimes2 : Simple Entailment 2 P.J. Hayes, P.F. Patel-Schneider, http://www.w3.org/TR/rdf11-mt/ Silvia Giannini Ph.D. Final Dissertation
- 19. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion The inference framework (b) Which inference-services can be enabled over such a unied framework? Description Logics (DLs) as baseline for non-standard reasoning services - Least Common Subsumer - Matching - Rewriting - Unication - Concept Abduction - Concept Contraction - ... Silvia Giannini Ph.D. Final Dissertation
- 20. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion The inference framework (b) Which inference-services can be enabled over such a unied framework? Description Logics (DLs) as baseline for (non)-standard reasoning services - Least Common Subsumer - Matching - Rewriting - Unication - Concept Abduction - Concept Contraction - ... Denition (Least Common Subsumer (LCS)) Let C1, . . . , Cn be a collection of n concepts in a DL L. The Least Common Subsumer (LCS) of C1, . . . , Cn is a concept D in L such that D is the most specic concept subsuming all the elements of the collection. Silvia Giannini Ph.D. Final Dissertation
- 21. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion The inference framework (b) Which inference-services can be enabled over such a unied framework? Description Logics (DLs) as baseline for (non)-standard reasoning services - Least Common Subsumer - Matching - Rewriting - Unication - Concept Abduction - Concept Contraction - ... Identication of subsets of resources related to a common informativecontent - Cluster search (approximate matching) - Entity disambiguation - Missing values identication - Personalization Silvia Giannini Ph.D. Final Dissertation
- 22. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Common Subsumers: A logic-based approach Example: Automatically extract Core Competence, by identifying a common know-how in a company personnel [1] Silvia Giannini Ph.D. Final Dissertation
- 23. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Common Subsumers: the Knowledge Compilation process Example: Automatically extract Core Competence, by identifying a common know-how in a company personnel [1] Silvia Giannini Ph.D. Final Dissertation
- 24. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Common Subsumers: the Knowledge Compilation process Example: Automatically extract Core Competence, by identifying a common know-how in a company personnel [1] Issues: Computational diculties of deduction in knowledge bases expressed through a logical formalism; Combining the representation power of a logical language, with the scalability and eciency of information processing in a DBMS. Knowledge Compilation: 1 OFF-LINE REASONING pre-processing of a company intellectual capital, described in a Description Logics (DLs) Knowledge Base (KB), in an appropriate relational database schema. 2 ON-LINE REASONING querying of the data structure coming out from the rst phase through standard SQL-queries for ecient Core Competence Extraction. Silvia Giannini Ph.D. Final Dissertation
- 25. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion I.M.P.A.K.T.3 Example: Automatically extract Core Competence, by identifying a common know-how in a company personnel [1] Knowledge Base Mario Rossi: Cplusplus (5 years), Java (5 years), Visual Basic (5 years) Daniela Bianchi: Cplusplus (2 years), Java (6 years), Visual Basic (1 years) Elena Pomarico: CplusPlus, Java, Visual Basic Carmelo Piccolo: VBScript, Process Performance Monitoring Lucio Battista: DBMS (2 years) Mariangela Porro: DBMS (2 years), Internet Technologies (2 years) Nicola Marco: DBMS (5 years), Internet Technologies (5 years) Domenico De Palo: OOprogramming (6 years), Articial intelligence (4 years), Internet technologies (4 years) 3 Information Management and Processing with the Aid of Knowledge-based Technologies Silvia Giannini Ph.D. Final Dissertation
- 26. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion I.M.P.A.K.T.3 Core Competence module GUI 3 Information Management and Processing with the Aid of Knowledge-based Technologies Silvia Giannini Ph.D. Final Dissertation
- 27. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 28. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Problem Denition Premises: DLs propose the LCS service for learning from examples RDF resources have no bounded description RDF has high-order feature not comparable to any DL Silvia Giannini Ph.D. Final Dissertation
- 29. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Problem Denition Premises: DLs propose the LCS service for learning from examples RDF resources have no bounded description RDF has high-order feature not comparable to any DL Silvia Giannini Ph.D. Final Dissertation
- 30. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Problem Denition Premises: DLs propose the LCS service for learning from examples RDF resources have no bounded description RDF has high-order feature not comparable to any DL Silvia Giannini Ph.D. Final Dissertation
- 31. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Problem Denition Premises: DLs propose the LCS service for learning from examples RDF resources have no bounded description RDF has high-order feature not comparable to any DL Silvia Giannini Ph.D. Final Dissertation
- 32. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Problem Denition Premises: DLs propose the LCS service for learning from examples RDF resources have no bounded description RDF has high-order feature not comparable to any DL Silvia Giannini Ph.D. Final Dissertation
- 33. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (RDF-path) Let T be a set of triples. A resource r is always RDF-connected to itself with an RDF-path of length 0 (independently of T ). A resource r is RDF-connected to another resource p with a path of length n + 1 if r is connected to a resource a with a path of length n, and either there is a triple a p s, or there is a triple a q p in T. Silvia Giannini Ph.D. Final Dissertation
- 34. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (RDF-connection) A resource r is RDF-connected to a resource p in T if there is an RDF-path from r to p. Denition (RDF-distance) The distance between two RDF-connected resources r and p is the length of the shortest RDF-path from r to p. Silvia Giannini Ph.D. Final Dissertation
- 35. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources p t q r s p Silvia Giannini Ph.D. Final Dissertation
- 36. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (Rooted RDF-Graph (r-graph)) A Rooted RDF-Graph (r-graph for short) is a pair r, Tr , where: 1 r is either the URI of an RDF resource, or a blank node; 2 Tr is a subset of the global set of triples in the Web, such that r is RDF-connected to every resource in Tr. Silvia Giannini Ph.D. Final Dissertation
- 37. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (Characteristic function) σTr : TW → {false, true} is dened s.t. σTr (t) = true if t ∈ Tr false if t /∈ Tr , where σTr is a parameter tuned according to the application problem. Silvia Giannini Ph.D. Final Dissertation
- 38. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (Rooted Entailment) Let R ∈ {S, RDF, RDF-S} be an entailment relation. R-graph r, Tr R-entails s, Ts denoted by r, Tr |=R s, Ts when the following conditions hold: 1 if s is a blank node, then 1 if r is not a blank node, Tr |=R Ts[s → r] must hold; 2 if also r is a blank node, then Tr[r → u] |=R Ts[s → u] for a new URI u occurring neither in Tr nor in Ts; 2 otherwise (i.e., s is not a blank node), if s = r, then Tr |=R Ts must hold; 3 otherwise (i.e., s is not a blank node and s = r) r, Tr never R-entails s, Ts . Silvia Giannini Ph.D. Final Dissertation
- 39. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (Common Subsumer) Let a, Ta , b, Tb be two r-graphs. An r-graph x, Tx is an R-Common Subsumer (R-CS) of a, Ta , b, Tb i both a, Ta |=R x, Tx and b, Tb |=R x, Tx . Silvia Giannini Ph.D. Final Dissertation
- 40. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Denition (Least Common Subsumer) Let a, Ta , b, Tb be two r-graphs. An r-graph x, Tx is an R-Least Common Subsumer (R-LCS) of a, Ta , b, Tb i both conditions below hold: 1 x, Tx is an R-CS of a, Ta , b, Tb ; 2 for every other R-CS y, Ty of a, Ta , b, Tb : if y, Ty |=R x, Tx then x, Tx |=R y, Ty , (i.e., x, Tx and y, Ty are R-equivalent). Silvia Giannini Ph.D. Final Dissertation
- 41. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solution Adaptation to RDF assertions: adaptation of path and connectedness denition from Graph Theory to RDF-graphs denition of rooted RDF-graphs (r-graph) denition of entailment between r-graphs denition of Common Subsumer of pairs of RDF resources Properties of R-LCS Given R-LCS( a, Ta , a, Ta ), the following properties hold: Uniqueness (or equivalence) Algebraic properties: idempotency commutativity associativity Moreover, it can be proved that S-LCS is computable in polymonial time. Silvia Giannini Ph.D. Final Dissertation
- 42. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Solving Algorithm Main Features: anytime: if interrupted, it always returns a Common Subsumer of the input pair of RDF resources modular: it takes as input a function computing the sets of triples relevant for the input RDF resources Our current criterion for triples selection: triples within a given graph distance from the input resource triples having properties within to a selected set of signicant properties for the dataset/application of interest Output: A Common Subsumer of two r-graphs a, Ta and b, Tb : a pair made up by a resource (anonymous or not) and a set of triples stating facts about such a resource which are true for both a and b. Alternative cases: _ : cs, T : a blank node _ : cs together with a set of triples related to _ : cs. a, Ta , i and a, Ta = b, Tb _ : cs, ∅ if either Ta = ∅ or Tb = ∅ Silvia Giannini Ph.D. Final Dissertation
- 43. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 44. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion RDF Clustering (c) How do we exploit them for managing heterogeneous web resources? Clustering of Web resources with CS Retrieving resources conveying the same information in their dierent RDF descriptions1 Discover homogeneous groups of resources 2 Identify a cluster concept description CS description → SPARQL queries: WHERE { Tcs [blank nodes → variables] } Silvia Giannini Ph.D. Final Dissertation
- 45. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: A use case The Italian Chamber of Deputies LOD 4 Running example: Find the commonalities between deputies Nilde Iotti and Tina Anselmi in the 10th Legislature Figure: Possible r-graphs for deputies Nilde Iotti and Tina Anselmi 4 Public SPARQL endpoint: http://dati.camera.it/sparql Silvia Giannini Ph.D. Final Dissertation
- 46. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: A use case The Italian Chamber of Deputies LOD 4 Running example: Find the commonalities between deputies Nilde Iotti and Tina Anselmi in the 10th Legislature Figure: A possible CS for deputies Nilde Iotti and Tina Anselmi r-graphs 4 Public SPARQL endpoint: http://dati.camera.it/sparql Silvia Giannini Ph.D. Final Dissertation
- 47. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: A use case The Italian Chamber of Deputies LOD 4 Running example: Clustering with a CS SELECT DISTINCT ?x0 WHERE{ ?x0 a http://dati.camera.it/ocd/deputato . ?x0 http://dati.camera.it/ocd/rif_leg http://dati.camera.it/ocd/legislatura.rdf/repubblica_10 . ?x0 http://dati.camera.it/ocd/rif_mandatoCamera ?x1 . ?x0 http://xmlns.com/foaf/0.1/gender female . ?x0 http://purl.org/dc/elements/1.1/description Laurea in lettere; insegnante@it . } 4 Public SPARQL endpoint: http://dati.camera.it/sparql Silvia Giannini Ph.D. Final Dissertation
- 48. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: A use case 10th Legislature clusters Silvia Giannini Ph.D. Final Dissertation
- 49. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: A use case 1st Legislature clusters: missing values Silvia Giannini Ph.D. Final Dissertation
- 50. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Clustering with CS: Some complexity results Silvia Giannini Ph.D. Final Dissertation
- 51. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Disambiguating with CS (c) How do we exploit them for managing heterogeneous web resources? An Entity Linking problem Silvia Giannini Ph.D. Final Dissertation
- 52. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Disambiguating with CS An Entity Linking problem Silvia Giannini Ph.D. Final Dissertation
- 53. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Disambiguating with CS An Entity Linking problem Silvia Giannini Ph.D. Final Dissertation
- 54. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Outline 1 Motivation 2 State-of-the-Art Technologies 3 Common Subsumers in the Web of Data 4 Proof-of-concept 5 Conclusion Silvia Giannini Ph.D. Final Dissertation
- 55. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion Research Answers Goal: Develop a framework able to manage distributed resources, heterogeneous in format, syntax and semantics (a) Identify RDF as the most adopted KR language for integration in the Web of Data (b) Explore (Least) Common Subsumer resoning service in DLs and dene an analog inferences for RDF with a proper computational algorithm (c) Analyse feasibility in possible application scenario (clustering, entity linking, drugs comparison, ...) Silvia Giannini Ph.D. Final Dissertation
- 56. Motivation State-of-the-Art Technologies Common Subsumers in the Web of Data Proof-of-concept Conclusion List of Publications 1 S. Colucci, E. Tinelli, S. Giannini, E. Di Sciascio, and F.M. Donini, Knowledge Compilation for Core Competence Extraction in Organizations In: Proc. of Business Information Systems 2013, Springer (2013) 163174. 2 E. Tinelli, S. Colucci, S. Giannini, E. Di Sciascio, and F.M. Donini, Large scale skill matching through knowledge compilation In: Proc. of ISMIS 2012, Springer-Verlag (2012) 192201. 3 S. Colucci, S. Giannini, F.M. Donini, and E. Di Sciascio, A deductive approach to the identication and description of clusters in Linked Open Data In: Proc. of the 21th European Conf. on Articial Intelligence (ECAI'14). IOS Press. 4 S. Giannini, RDF Data Clustering, In: Business Information Systems Workshops 2013, Springer (2013) 220-231. 5 S. Colucci, S. Giannini, F.M. Donini, E. Di Sciascio, Finding Commonalities in Linked Open Data, In: 29th Italian Conference on Computational Logic (CILC 2014), 37 - 42. 6 S. Giannini, Heterogeneous resources management through an RDF-based inference service, In: 1st SCORE@POLIBA workshop (2014). 7 S. Colucci, F.M. Donini, S. Giannini, and E. Di Sciascio, Dening and computing Least Common Subsumers in RDF, Journal of Web Semantics, Submitted and under review. Silvia Giannini Ph.D. Final Dissertation