Umberto Straccia Andrea Calì (Eds.) 0 Scalable Uncertainty 2 7 8 AI Management N L 8th International Conference, SUM 2014 Oxford, UK, September 15–17, 2014 Proceedings 123 Lecture Notes in Artificial Intelligence 8720 Subseries of Lecture Notes in Computer Science LNAISeriesEditors RandyGoebel UniversityofAlberta,Edmonton,Canada YuzuruTanaka HokkaidoUniversity,Sapporo,Japan WolfgangWahlster DFKIandSaarlandUniversity,Saarbrücken,Germany LNAIFoundingSeriesEditor JoergSiekmann DFKIandSaarlandUniversity,Saarbrücken,Germany Umberto Straccia Andrea Calì (Eds.) Scalable Uncertainty Management 8th International Conference, SUM 2014 Oxford, UK, September 15-17, 2014 Proceedings 1 3 VolumeEditors UmbertoStraccia IstitutodiScienzaeTecnologiedell’Informazione(ISTI-CNR) Pisa,Italy E-mail:[email protected] AndreaCalì Birkbeck,UniversityofLondon DepartmentofComputerScienceandInformationSystems London,UK E-mail:[email protected] ISSN0302-9743 e-ISSN1611-3349 ISBN978-3-319-11507-8 e-ISBN978-3-319-11508-5 DOI10.1007/978-3-319-11508-5 SpringerChamHeidelbergNewYorkDordrechtLondon LibraryofCongressControlNumber:2014948263 LNCSSublibrary:SL7–ArtificialIntelligence ©SpringerInternationalPublishingSwitzerland2014 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped.Exemptedfromthislegalreservationarebriefexcerptsinconnection withreviewsorscholarlyanalysisormaterialsuppliedspecificallyforthepurposeofbeingenteredand executedonacomputersystem,forexclusiveusebythepurchaserofthework.Duplicationofthispublication orpartsthereofispermittedonlyundertheprovisionsoftheCopyrightLawofthePublisher’slocation, inistcurrentversion,andpermissionforusemustalwaysbeobtainedfromSpringer.Permissionsforuse maybeobtainedthroughRightsLinkattheCopyrightClearanceCenter.Violationsareliabletoprosecution undertherespectiveCopyrightLaw. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Whiletheadviceandinformationinthisbookarebelievedtobetrueandaccurateatthedateofpublication, neithertheauthorsnortheeditorsnorthepublishercanacceptanylegalresponsibilityforanyerrorsor omissionsthatmaybemade.Thepublishermakesnowarranty,expressorimplied,withrespecttothe materialcontainedherein. Typesetting:Camera-readybyauthor,dataconversionbyScientificPublishingServices,Chennai,India Printedonacid-freepaper SpringerispartofSpringerScience+BusinessMedia(www.springer.com) Preface Information systems are becoming increasingly complex, involving massive amounts of data coming from different sources. Information is often inconsis- tent, incomplete, heterogeneous, and pervaded with uncertainty. The International Conference on Scalable Uncertainty Management (SUM) conferences series provides an international forum about the management of uncertain, incomplete, or inconsistent information. This volume contains the papers presented at the 8th International Confer- ence on Scalable Uncertainty Management (SUM 2014), which was held at St Anne’s College, Oxford, UK, from the 15th to the 17th of September, 2014. The call for papers solicited submissions in two categories: regular research papersandshortpapers,where the latter reportoninterestingworkinprogress or provide system descriptions. The call for papers resulted in 47 submissions. Based on the review reports and discussions, 26 papers were accepted for pub- lication andpresentationat the conference, among which 20 regular papers and 6 short papers. The conference program also included invited lectures by three leadingresearchers:AnthonyHunter(DepartmentofComputerScience,Univer- sityCollegeLondon),JensLehmann(DepartmentofComputerScience,Univer- sityofLeipzig),andDanOlteanu(DepartmentofComputerScience,University of Oxford). Aconferencesuchasthiscanonlysucceedasateameffort. Wewouldliketo thank:theauthorsofsubmittedpapers,theinvitedspeakers,andtheconference participants;the members ofthe ProgramCommittee andthe externalreferees; Alfred Hofmann and Springer for providing assistance and advice in the prepa- ration of the proceedings; the University of Oxford for providing local facilities; the creators and maintainers of the conference management system EasyChair. All of them made the success of SUM 2014 possible. July 2014 Umberto Straccia Andrea Cal`ı Organization Program Committee Leila Amgoud IRIT - CNRS, France Nahla Ben Amor InstitutSuperieurdeGestiondeTunis,Tunisia Leopoldo Bertossi Carleton University, Canada Isabelle Bloch ENST - CNRS UMR 5141 LTCI, France Fernando Bobillo University of Zaragoza, Spain Loreto Bravo Universidad de Concepcio´n, Chile Andrea Cal`ı University of London, Birkbeck College, UK Laurence Cholvy ONERA - Toulouse, France Jan Chomicki University at Buffalo, USA Fabio Cozman Universidade de Sao Paulo, Brazil Alfredo Cuzzocrea ICAR - CNR, Italy Michael Dekhtyar Tver State University, Russia Juergen Dix Clausthal University of Technology, Germany Didier Dubois IRIT - CNRS, France Thomas Eiter Vienna University of Technology, Austria Zied Elouedi LARODEC, ISG de Tunis, Tunisia Ronald Fagin IBM Research - Almaden, USA Minos Garofalakis Technical University of Crete, Greece Lluis Godo Artificial Intelligence Research Institute, IIIA - CSIC, Spain Nikos Gorogiannis Middlesex University, UK John Grant Towson University, USA Sergio Greco University of Calabria, Italy Stijn Heymans SRI International, USA Anthony Hunter University College London, UK Gabriele Kern-Isberner Technische Universitaet Dortmund, Germany Angelika Kimmig KU Leuven, Belgium Kathryn Laskey George Mason University, USA Jonathan Lawry University of Bristol, UK Sebastian Link The University of Auckland, New Zealand Peter Lucas Radboud University Nijmegen, The Netherlands Jianbing Ma Bournemouth University, UK Thomas Meyer Centre for Artificial Intelligence Research, UKZN and CSIR Meraka, South Africa Serafin Moral University of Granada, Spain Kedian Mu School of Mathematical Sciences, Peking University, China VIII Organization Jeff Z. Pan University of Aberdeen, UK Simon Parsons University of Liverpool, UK Rafael Pen˜aloza TU Dresden, Germany Olivier Pivert IRISA - ENSSAT, France Henri Prade IRIT - CNRS, France Andrea Pugliese DEIS - University of Calabria, Italy Guilin Qi Southeast University, China Fabrizio Riguzzi University of Ferrara, Italy Sebastian Rudolph Technische Universita¨t Dresden, Germany V´ıtor Santos Costa Universidade do Porto, Portugal Steven Schockaert Cardiff University, UK Guillermo Ricardo Simari Universidad Nacional del Sur in Bahia Blanca, Argentina Giorgos Stoilos National Technical University of Athens (NTUA), Greece Umberto Straccia ISTI - CNR, Italy Heiner Stuckenschmidt University of Mannheim, Germany Vicenc Torra IIIA - CSIC, Spain Peter Vojt´aˇs Charles University, Czech Republic Nic Wilson 4C, UCC, Cork, Ireland Ronald Yager Machine Intelligence Institute - Iona College, USA Additional Reviewers Britz, Arina Molinaro, Cristian Casini, Giovanni Spina, Cinzia Incoronata Garcia, Jhonatan Tsalapati, Eleni Klarman, Szymon Weinzierl, Antonius Meilicke, Christian Table of Contents Possibilistic Networks: A New Setting for Modeling Preferences ........ 1 Nahla BenAmor, Didier Dubois, H´ela Gouider, and Henri Prade Min-based Assertional Merging Approach for Prioritized DL-Lite Knowledge Bases ................................................ 8 Salem Benferhat, Zied Bouraoui, Sylvain Lagrue, and Julien Rossit On the Revision of Prioritized DL-Lite Knowledge Bases.............. 22 Salem Benferhat, Zied Bouraoui, and Karim Tabia Interval-Based Possibilistic Networks ............................... 37 Salem Benferhat, Sylvain Lagrue, and Karim Tabia Tractable vs. Intractable Cases of Query Answering under Matching Dependencies.................................................... 51 Leopoldo Bertossi and Jaffer Gardezi Analogical Classification: Handling Numerical Data .................. 66 Myriam Bounhas, Henri Prade, and Gilles Richard Lazy Analogical Classification: Optimization and Precision Issues ...... 80 William Correa Beltran, H´el`ene Jaudoin, and Olivier Pivert Polyhedral Labellings for Argumentation Frameworks ................ 86 Cosmina Croitoru Update Operators for Inconsistent Query Answering: A New Point of View......................................................... 100 Madalina Croitoru and Ricardo Oscar Rodriguez Conflicts of Belief Functions: Continuity and Frame Resizement........ 106 Milan Daniel and Jianbing Ma Improving Inconsistency Resolution by Considering Global Conflicts.... 120 Cristhian Ariel David Deagustini, Maria Vanina Mart´ınez, Marcelo A. Falappa, and Guillermo Ricardo Simari Probabilistic Argumentation Frameworks – A Logical Approach........ 134 Dragan Doder and Stefan Woltran Computing Skyline from Evidential Data ........................... 148 Sayda Elmi, Karim Benouaret, Allel Hadjali, Mohamed Anis Bach Tobji, and Boutheina Ben Yaghlane X Table of Contents A Two-Level Approach to Maximum Entropy Model Computation for Relational Probabilistic Logic Based on Weighted Conditional Impacts......................................................... 162 Marc Finthammer and Christoph Beierle Solving Hidden-Semi-Markov-Mode Markov Decision Problems ........ 176 Emmanuel Hadoux, Aur´elie Beynier, and Paul Weng Probabilistic Strategies in Dialogical Argumentation.................. 190 Anthony Hunter Analytics over Probabilistic Unmerged Duplicates.................... 203 Ekaterini Ioannou and Minos Garofalakis A PsychologicalAnalysis of Preference Semantics in Conditional Logics for Preference Representation...................................... 209 Souhila Kaci and Eric Raufaste Answering Ontological Ranking Queries Based on Subjective Reports... 223 Thomas Lukasiewicz, Maria Vanina Mart´ınez, Cristian Molinaro, Livia Predoiu, and Gerardo I. Simari A Petri Net Model of Argumentation Dynamics...................... 237 Diego C. Martinez, Maria Laura Cobo, and Guillermo Ricardo Simari Integrity Constraints for Probabilistic Spatio-Temporal Knowledgebases ................................................. 251 Francesco Parisi and John Grant Repairs and Consistent Answers for Inconsistent Probabilistic Spatio-Temporal Databases ....................................... 265 Francesco Parisi and John Grant Skyline Queries in an Uncertain Database Model Based on Possibilistic Certainty ....................................................... 280 Olivier Pivert and Henri Prade Querying Uncertain Multiple Sources............................... 286 Olivier Pivert and Henri Prade Population Size Extrapolation in Relational Probabilistic Modelling .... 292 David Poole, David Buchman, Seyed Mehran Kazemi, Kristian Kersting, and Sriraam Natarajan Improving PersonalizedSearchon the Social Web Based on Similarities between Users ................................................... 306 Zhenghua Xu, Thomas Lukasiewicz, and Oana Tifrea-Marciuska Author Index.................................................. 321 Possibilistic Networks: A New Setting for Modeling Preferences NahlaBenAmor1,DidierDubois2,He´laGouider1,andHenriPrade2 1 LARODEC,Universite´deTunis,ISGdeTunis,41ruedelaLiberte´,2000LeBardo,Tunisia 2 IRIT–CNRS,118,routedeNarbonne,Toulouse,France [email protected], {dubois,prade}@irit.fr, [email protected] Abstract. PossibilisticnetworksarethecounterpartofBayesiannetworksinthe possibilisticsetting.Possibilisticnetworkshaveonlybeenstudiedanddeveloped from a reasoning-under-uncertainty point of view until now. In thisshort note, for the first time, one advocates their interest in preference modeling. Beyond theirgraphicalappeal,theycanbeshowntoprovideanaturalencodingofpref- erencesagreeingwiththeinclusion-basedpartialorderappliedtothesubsetsof preferencesviolatedinthedifferentsituations.Moreovertheydonotencounter thelimitationsofCP-Netsintermsofrepresentationcapabilities.Theyalsoenjoy alogicalcounterpartthatmaybeusedforconsistencychecking.Thisshortnote provides a comparative discussion of the merits of possibilistic networks with respecttootherexistingpreferencemodelingframeworks. 1 Introduction Preferencesareusuallyexpressedbymeansoflocalpiecesofinformation,ratherthan as a complete preorder between the different possible states of the world. This state of facts has led AI researchers to propose compactrepresentation formats for prefer- encesandproceduresforcomputingaplausiblerankingbetweencompletelydescribed situations from such representations, in the last fifteen years. Conditional preference networks[6](CP-Netsforshort)haveemergedasapopularreferencesettingforrepre- sentingpreferences,leadingtodifferentrefinements[5,15],aswellassomealternative approaches[4,8,13] (see [7] for a brief overview). Inspired from Bayesian networks, CP-Netsinherittheirgraphicalnature,andbesides,relyonasimple,apparentlynatural principle,namedceterisparibus,whichallowstoextendanycontextualpreference“in contextc,Ipreferato¬a”(denotedforshortc:a(cid:2)¬a),toanyparticularspecification boftheothervariablesusedfordescribingtheconsideredsituations,i.e.,thepreference is understood as ∀b,cab is preferred to c¬ab. The CP-net approach perfectly exem- plifiestheingredientsneededforasatisfactoryrepresentationofpreferences,statedin a conditionalmanner,into a partial order useful for a user: i) a simple representation setting,preferablyhavingagraphicalcounterpartforelicitationease,ii)anaturalprin- cipleformakingexplicitthepreferencesbetweencompletelydescribedsituations,and iii)analgorithmfordetermininghowtocomparetwocompletesituationsaccordingto theexistenceofapathofworseningflipslinkingthem.Inspiteoftheirappealingfea- tures,CP-Nets havesomelimitations.First, there existpreordersthatmakesense and U.StracciaandA.Cal`ı(Eds.):SUM2014,LNAI8720,pp.1–7,2014. (cid:2)c SpringerInternationalPublishingSwitzerland2014