Semantic Wikis - Social Semantic Web in Action

Semantic Wikis - Social Semantic Web in Action

• 1.
  Semantic WikisSocial SemanticWeb In Action2011-03-25Specially Prepared for Tsinghua University Alumniin greater Seattle area for centennial celebration
• 2.
  About Me: JesseWang 王嘉欣2199819962005199719931988
• 3.
  Who is Vulcan3
• 4.
  What does Vulcando4Vulcan Inc. was established in 1986 by investor and philanthropist Paul G. Allen, co-founder of Microsoft, to manage his business and philanthropic efforts. Allen is chairman of Vulcan and his sister, Jody Allen, is president and CEO.
• 5.
  It all beganwith a vision…5
• 6.
  Now the VisionContinues as Project Halo6Automatic Question Answering SystemProject Halo is a staged, long-range research effort by Vulcan Inc. towards the development of a "Digital Aristotle"—a reasoning system capable of answering novel questions and solving advanced problems in a broad range of scientific disciplines and related human affairs. The project focuses on creating two primary functions: a tutor capable of instructing and assessing students in those subjects, and a research assistant with broad, interdisciplinary skills to help scientists and others in their work.
• 7.
  Project Halo’s FocusAreas7Knowledge AcquisitionPlus other related semantic technologies and commercial efforts
• 8.
  Project Halo’s GoalsAddressthe core problems in Knowledge BasesscalebrittlenessHave high impact8NowFuture
• 9.
  Crowdsourcing for BetterKnowledge Acquisition9
• 10.
  Success of Wikis10Oneof human’s greatest inventions
• 11.
  11OutlineWikis and SemanticWikisCrowdsourcing and Consensus on DataSemantic MediaWiki and its ExtensionsWiki-based Knowledge Management on a Larger ScaleEnterprise Knowledge ManagementSemantic EncyclopediaEvolving as a Web Application Development PlatformExamples: Semantic Football, Agile Project Management
• 12.
  A Key Featureof Wiki12ConsensusThis distinguishes wikis from other publication tools
• 13.
  Consensus in WikisComes fromCollaboration~17 edits/page on average in Wikipedia (with high variance)Wikipedia’s Neutral Point of View ConventionUsers follow customs and conventions to engage with articles effectively13
• 14.
  Software Support MakesWikis SuccessfulTrivial to editby anyoneTracking of all changes, one-step rollbackEvery article has a “Talk” page for discussionNotification facility allows anyone to “watch” an articleSufficient security on pages, logins can be requiredA hierarchy of administrators, gardeners, and editorsSoftware Bots recognize certain kinds of vandalism and auto-revert, or recognize articles that need work, and flag them for editors14
• 15.
  Finding Deeper InfoWikipediahas articles about…… all cities with info on their populations, locations and skyscrapers, etc.… all German cars with engine size, accelerating data…Can you find: Skyscrapers with 50+ floors and built after 2000 in Shanghai (or Chinese cities with 1,000,000+ people)?Or German(Porsche) cars that accelerate from 0-100km/h in 5 seconds?15
• 16.
  How Wikipedia Answers– List!16http://en.wikipedia.org/wiki/List_of_fastest_cars_by_acceleration
• 17.
  Going Deeper17http://en.wikipedia.org/wiki/List_of_German_cars
• 18.
  Deeper…18
• 19.
  And Deeper…19
• 20.
  And Now…20
• 21.
  Look into Listin Wikipedia21http://en.wikipedia.org/wiki/List_of_German_cars
• 22.
  Editing Standard WikiArticle – Static List22
• 23.
  Static List, Tables,…, Not Useable Enough23http://en.wikipedia.org/wiki/List_of_lists_about_Oregon
• 24.
  Semantics Come ToRescueTo find answers like:All Porsche vehicles made in Germany that accelerate from 1-100 km/h less than 4 seconds
• 25.
  Sci-Fi movies madeafter year 2000 that cost less than $10M and gross more than $30M
• 26.
  A map showingwhere all Mercedes-Benz vehicles are manufactured
• 27.
  All skyscrapers inChina (Japan, Thailand,…) of 50 (40/60/70) floors or more, and built in year 2000 (2001/2002) and after, sorted by built year, floors…, grouped by cities, regions…
• 28.
  And many moreWeneed structured data with clear and consistent semantics24
• 29.
  What is aSemantic WikiA wiki that has an underlying model of the knowledge  described in its pages.To allow users to make their knowledge explicit and formalSemantic Web Compatible25Semantic WikiHybrid ... Better Gas Mileage!
• 30.
  Two Perspectives26
• 31.
  Characteristics of SemanticWikisSemantic Wikis27
• 32.
  List of SemanticWikisAceWikiArtificialMemoryWagn - Ruby on Rails-basedKiWi – Knowledge in a WikiKnoodl – Semantic Collaboration tool and application platformMetaweb - the software that powers FreebaseOntoWikiOpenRecordPhpWikiSemantic MediaWiki - an extension to MediaWiki that turns it into a semantic wikiSwirrl - a spreadsheet-based semantic wiki applicationTaOPis - has a semantic wiki subsystem based on Frame logicTikiWiki CMS/Groupware integrates Semantic links as a core featurezAgile Wikidsmart - semantically enables Confluence28
• 33.
  Basics of SemanticWikisStill a wiki, with regular wiki featuresCategory/Tags, Namespaces, Title, Versioning, ...Typed Content (built-ins + user created, e.g. categories)Page/Card, Date, Number, URL/Email, String, …Typed Links (e.g. properties)“capital_of”, “contains”, “born_in”…Querying Interface SupportE.g. “[[Category:Member]] [[Age::<30]]” (in SMW)29
• 34.
  Short History ofSemantic MediaWikiBorn at AIFBTyped links and types and moreExport articles as RDFMaximally flexible for the wiki userSMW 0.1 released by AIFB in Sept 2005Parser/storage support for typed links – [[type::link | label]]FactBox for semantic relations at end of articleSpecial:SearchSemantic, with basic auto-completion for link typesSimple query language (“ask”)Vulcan kicks off Halo Extensions to SMW project in August 2007SMW 1.0 released by AIFB in Dec 2007, Ontoprise releases Halo Extension 1.0 in parallel“Property” instead of “Relation” and “Attribute”Many new datatypes/special pages/UI features30
• 35.
  Semantic MediaWiki (SMW)Markup Syntax31[[Property::Value | Display]]Tsinghua is a university located in [[Has location::Beijing]], with[[Has population::27000|about 27 thousands]] students.In page "Property:Haslocation":[[Has type::Page]]In page "Property:Haspopulation":[[Has type::number]]
• 36.
  Special Properties“Has Type”is a pre-defined “special” property for meta-dataExample: [[Has type::String]]“Allowed Values” is another special property[[Allows value::Low]], [[Allows value::Medium]], [[Allows value::High]]In Halo Extensions, there are domain and range supportRDFs expressivitySemantic Gardening extension also supports “Cardinality”32
• 37.
  Define Classes33Beijing isa city in [[Has country::China]], with population [[Has population::2,200,000]].[[Category::Cities]]Categories are used to define classes because they are better for class inheritance.The Jin Mao Tower (金茂大厦) is an 88-story landmarksupertallskyscraper in …[[Categories: 1998 architecture | Skyscrapers in Shanghai | Hotels in Shanghai | Skyscrapers over 350 meters | Visitor attractions in Shanghai | Landmarks in Shanghai | Skidmore, Owings and Merrill buildings]]Category: Skyscrapers by country Category:Skyscrapers in China
• 38.
  Database-style Query overWiki Data34Example: Skyscrapers in China higher than 50 stories, built before 2000ASK/SPARQL query target{{#ask: [[Category:Skyscrapers]] [[Located in::China]] [[Floor count::>50]] [[Year built::<2000]] …}}Data via DBpedia
• 39.
  35Why Semantic Wiki?Annotationof existing structures with machine readable metadatalinks carry meaning, typing of links, typing of pages
• 40.
  Context dependent adaptationand presentationdifferent domains have different ways of presenting content, personal preferences, etc.
• 41.
  Improved, “intelligent”, searchand navigationqueries to the structure, visualisation of structure, derived information
• 42.
  Improved interoperability betweensystemsexchange of content, integration of different systems, agents, etc.What is the Promise of Semantic Wikis?Semantic Wikis promise Consensus over DataCombine low-expressivity data authorship with the best features of traditional wikisUser-governed, user-maintained, user-definedEasy to use as an extension of text authoring36The ultimate data aggregator
• 43.
  37Challenges on DataConsensusData modeling is (seemingly) a specialized skillFinding disagreements in data is difficultConsistently revising data schemas is difficultConsistency of schema information (“Population”, “Pop”, “Number_of_inhabitants”, etc...)Consistency of types, units of measure, application of rules…Semantics/interpretation of properties need explanation for humans…
• 44.
  One Key HelpfulFeature of Semantic Wikis38Semantic Wikis are “Schema-Last”Databases require DBAs and schema design; Semantic Wikis develop and maintain the schema in the wiki
• 45.
  Semantic MediaWiki CommunityOpensource (GPL)Well documentedActive mailing listCommercial support availableWorld-wide communityRegular ConferencesNext SMWCon 4/28-30, 2011 Arlington, VA39http://semantic-mediawiki.org/Very stable SMW coreMature while still growing, slowly but steadily
• 46.
  SMW Extensions –Help Build Great Things40
• 47.
  Example: Ultrapedia –Semantic WikipediaUltrapedia: An SMW demo built to explore general knowledge acquisition in a wikiWikipedia merged with the power of a databaseHelp Readers and Writers Be More Productive41An Analytical Encyclopedia
• 48.
  Data Flow inthe Ultrapedia Prototype Real-time feed of WP changesNote most WP page changes will be text and have no semantic importEnglish Wikipedia subsetDynamic extraction of WP semantic data into RDFDBpedia update streamWP page text updates
• 49.
  DBpedia data updatesWPupdatesUser-created page updates in WikipediaUltrapediaEnhanced Ultrapedia UsabilityFamiliar WP page text and layout
• 50.
  Exhibit-based visualizations
• 51.
  Dynamic tables/categories
• 52.
  Faceted navigation
• 53.
  Queries (both standingand ad-hoc)
• 54.
  Linked to relevantexternal dataWikipedia-based CorrectionsUP shows the user where to correct data in WP so that DBpedia will extract the correction
• 55.
  Ultrapedia exposes thedata source in terms of where the data was extracted from WP
• 56.
  WP changes andcorrections get quickly propagated to UPUltrapedia: An Analytic EncyclopediaGoal: Prototype a small semantic encyclopediaCreate an semantic version of a part of WikipediaSoftware is SMW+, Ontobrokertriplestore, DBpediaShow what a data-aware encyclopedia might look likeUltrapedia Prototype DetailsTest domain is German cars~2500 Wikipedia pages, ~40000 triplesFeaturesSimilar look and feel to WikipediaDynamic tables and chartsPowerful queriesNavigation beyond searchEdit, discuss and rate dataSPARQL-based queriesDerived assertions (via OntoBroker)
• 57.
  Better Views ofthe Wiki Datahttp://wiking.vulcan.com/up/index.php/Porsche_996
• 58.
  Dynamic Views ofthe Acceleration Data
• 59.
  Graph Views ofthe Acceleration Data
• 60.
  Dynamic Mapping andCharting
• 61.
  Information Discovery viaVisualization48
• 62.
  The InspirationWe startedwith a We could have an49wiki siteweb application
• 63.
  50Video: SemanticWikis for A New ProblemSemantic Entertainment WikiSocial database-style characterization
• 64.
  Database search +wiki text search
• 65.
  Semantic consistency viawiki mechanisms
• 66.
  Easy to engineerIncreasingtechnical complexity -> ← Increasing User ParticipationSocial tag-based characterizationKeyword search over tag dataInconsistent semanticsEasy to engineerAlgorithm-based object characterization
• 67.
  Database-style search
• 68.
  Consistent semantics
• 69.
  Extremely difficult toengineerSemantic Seahawks Football Wiki51
• 70.
  Based on SimpleTemplates and Forms52
• 71.
  Template:Run Source Code<noinclude> This is the 'Run' template. It should be called in the following format:<pre> {{Run |Running Back= |Run Direction Type= |Yardage= |Run of X Yards= |Result of Run Type= }}</pre>Edit the page to see the template text.</noinclude> <includeonly>{| class="wikitable"{{#if:{{{Running Back|}}}|! Running Back{{!}} {{#arraymap:{{{Running Back|}}}|,|x|[[Running Back::x]]}} {{!}}- }}{{#if:{{{Run Direction Type|}}}|! Run Direction Type{{!}} {{#arraymap:{{{Run Direction Type|}}}|,|x|[[Run Direction Type::x]]}}{{!}}- }}{{#if:{{{Yardage|}}}|! Yardage{{!}} {{#arraymap:{{{Yardage|}}}|,|x|[[Yardage::x]]}}{{!}}- }}{{#if:{{{Run of X Yards|}}}|! Run of X Yards{{!}} [[Run of X Yards::{{{Run of X Yards|}}}]]{{!}}- }}{{#if:{{{Result of Run Type|}}}|! Result of Run Type{{!}} {{#arraymap:{{{Result of Run Type|}}}|,|x|[[Result of Run Type::x]]}}}}|}[[Category:Play]]</includeonly>53
• 72.
  Semantic Entertainment: QueryResult  Highlight ReelCommercial Look/Feel
• 73.
  Play-by-play video search
• 74.
  Highlight reel generation
• 75.
  Search on crowd-definedpatterns (“touchdowns with big hits”)
• 76.
  Tree-based navigation widget
• 77.
  Very favorable economicsDemo
• 78.
  We CAN BuildApplications (Fairly) EasilyWith all the extensions of Semantic MediaWiki.55Social Semantic Web Applications
• 79.
  Showcase: NPS Wiki:Browse People56
• 80.
  View of Datain NPS Wiki57
• 81.
  Show case: WorkOrder Processing Wiki (NGT)58
• 82.
  Collaborative Proposal Managementat BT with SMW+59Active Bid Viewer Service Desk Selector
• 83.
  Showcase: RPI Map60RPIMaphttp://map.rpi.eduA mash-up map application based on Semantic MediaWikiProvides location-based information in the RPI campusIntegrates data from various external sourcesVisualizes integrated data using Google Map
• 84.
  Social Semantic WebApplications61Omitting x examples, y pictures and z lines of text…
• 85.
  Case Study andDemo: Project Management with SMW+62Automatically populate tables
• 86.
  Just the datayou want,
• 87.
  At the levelyou want
• 88.
  Calendars and timelines
• 89.
  Workflows
• 90.
  Personal menus
• 91.
  Form-oriented inputs
• 92.
  Notifications via email/RSS
• 93.
  MS Office integration
• 94.
  SVN integrationVulcan ProjectManagement Wiki (Story)Template and style sheet customizationsRelated content automatically included
• 95.
  Vulcan Project ManagementWiki (Task)64Color codes to indicate types and statusSVN Integration automatically “Completed” task and relate to repository
• 96.
  Vulcan Project ManagementWiki (Visualizations)65Demo
• 97.
  Screenshot of aSprint page66Data automatically generated via template queries on pagehttp://wiking.vulcan.com/dev/index.php/Sprint_101020
• 98.
  Requirements for Wiki“Developers”67One need notWrite code like a hardcore programmerDesign, setup RDBMS or make frequent schema changesPossess knowledge of a senior system adminInstead one needConfigure the wiki with desired extensionsDesign and evolve the data model (schema)Design ContentCustomize templates, forms, styles, skin, etc.
• 99.
  Effectiveness of SMWas a Platform ChoiceSMW + ExtensionsPackaged SoftwareCustom Development☺ Still quick to program☺ Easy to customize☺ Low-moderate costVulcan Project WikiB.L.S.RPI map☺Very quick to obtainN Hard to customizeN ExpensiveMicrosoft ProjectVersion OneMicrosoft SharePointN Slow to develop☺Extremely flexibleN High cost to develop and maintain.NET FrameworkJ2EE, …Ruby on rails68
• 100.
  Openness of SMWas a Platform69Open SourceOpen ContentOpen Metadata
• 101.
  Other SMW+ use?Collaborationapplications were conceived as desktop appsThen wikis made the web collaborativeNow the action is in mobile apps70
• 102.
  Potential to BuildMany More AppsWhy are there 300,000+ apps in the iPhone App Store?UI limitations drive specificity in appsPeople personalize their phonesBut better browser technologies are shrinking the gap between native apps and web pagesHTML5, JavaScript, etc.SMW is a tool to build apps!Collaborative: social semantic in natureData flow and report drivenCheap to customize and rapidly deployableHigh signal-to-noise ratio for the usersVulcan is investigating this concept71
• 103.
  Summary: Application Platformby SMW+ ExtensionsSemantic MediaWiki + wide range of extensions make it a potential application development platform for social semantic webSMW + extensions provide a choice that fits into cost-effective sweet spot SMW + extensions could become a great platform for social semantic web application development, with moreExtensions, Widgets and Applications72There is an app for it!
• 104.
  73Conclusions: Semantic MediaWikiis a Powerful ToolSemantic MediaWiki+ (http://smwforum.ontoprise.com) Open-source, growing semantic wiki software systemWiki-style text + semantic markupsCollaborative, user-governed subject models and data curationSimple and extensible data models with easy import/exportSMW+ has many government and industry usersPeople built applications with it Knowledge Management viacrowds can workA way to leverage and exploit web-collected dataA lightweight collaborative knowledge management toolA new platform for lightweight web application developmentNowFuture
• 105.
  AcknowledgementThank you!74
• 106.
  75(End of Slides)Backupsstarts here
• 107.
  Case Study: Battle-spaceLuminary System Discover when New Information represents a change in understanding of entitiesDiscovery of explicit entity links, implicit relationshipsLarge Volumes of Data in various formatsUnstructured news articlesTactical Reports, Field IntelligenceStructured Database InformationUse Wiki Pages to represent current knowledge about an entity – “what we know”Domain Ontology to represent domain of information – “what we want to know”Issue Alerts when Significant Events occurNew information according to categoryChanging information on topics of interestNeed to send information to various devices – cell phones, email, etc.76
• 108.
  System DesignWiki ConfigurationSemanticMediaWiki: Large developer community, active development, open source. Wikipedia uses MediaWiki, so scalability and performance are important.Semantic Results Format: Provides various rich media displays of semantic information, including graphs, timelines, mapsSemantic Forms: Provides convenient user interface for entering semantic data into wiki, avoiding cumbersome wikitextSemantic Notifications: Enables sending of notifications when results of semantic query change.Domain OntologyCreated OWL Ontology for TerrorismSemantic Parsing, Extraction, ReasoningJava Process using various Open-Source ToolkitsRapid plugin of new technologiesMultiple Data Sources supported77
• 109.
  Sample Content Page78
• 110.
  Wiki Content DesignUseTemplates to Ensure Consistent Look-and-FeelTemplates Correspond to Ontology ClassesFields within Templates correspond to Properties within OntologyRich Content Visualizations derived in consistent wayHierarchical Categories match Class Hierarchy within OntologyEnsures Validity for PropertiesCategory included on each Template page to ensure consistencyFormsProvide ability for users to enter data directly into wiki without knowing Wiki TextEach form corresponds to a TemplateFields within forms correspond to the fields/properties within the TemplateGUI can include auto-completionCreated Page immediately linked semantically to rest of Wiki79
• 111.
  Sample Visualizations80UI enablesnotifications based on results of query – message sent when visualization changesVisualizations automatically created w/o user edit(tables, timelines, maps, social networks…)
• 112.
  Wikipedia for Porsches(Acceleration Data Example)Information Need: All Porsche models that accelerate 0-100kph in under 5, 6, and 7 secondsMore Porsche Acceleration Data in Wikipedia
• 113.
  Main PageUltrapedia MainPage
• 114.
  Tree View ControlAbstract/Summaryquick previewSemantics for Improved Wiki Navigation
• 115.
  The Porsche 996Acceleration Table In Ultrapedia
• 116.
  Same Table asa Query
• 117.
  Which Porsches acceleratefast?Dynamically-Generated Tables for QueriesInformation Need: All Porsche models that accelerate 0-100kph in under 5, 6, and 7 secondsGraph Views of the Acceleration Data
• 118.
  External Data viaa Live Ebay Query
• 119.
  Linking to ExternalEbay Data
• 120.
  Mercedes-Benz E-class W212Gallery SectionPhotos in Wiki Articles as Data
• 121.
  Volkswagen Production TimelineViewTimelines from Data
• 122.
  Dynamic Mapping andCharting
• 123.
  Editing Wiki DataIn PlaceReturn