1 / 41

The IRIS Risk Knowledge Portal

The IRIS Risk Knowledge Portal. http://irisportal.csd.auth.gr. Georgios Meditskos , Nick Bassiliades Logic Programming & Intelligent Systems group Dept. of Informatics Aristotle University of Thessaloniki, Greece. Outline. Introduction The Risk Ontology The Risk Knowledge Portal

adonis
Télécharger la présentation

The IRIS Risk Knowledge Portal

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The IRIS Risk Knowledge Portal http://irisportal.csd.auth.gr GeorgiosMeditskos, Nick Bassiliades Logic Programming & Intelligent Systems group Dept. of Informatics Aristotle University of Thessaloniki, Greece

  2. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  3. The Role of Semantics in IRIS • All the variables associated with the Risk Assessment Process are represented in terms of a Risk Ontology • An upper-level schema that describes terms, relationships and restrictions of risk identification and assessment • It is a formal representation of the IRIS Risk Glossary that is used for defining Risk Case Studies (Risk Identification)

  4. Benefits of Semantics • Formal vocabulary of terms publicly available for use by different organizations • integration of risk assessment practices from different domains • Derivation of implicit/hidden relationships through reasoning • semantic integration and consistency checking using state-of-the-art ontology reasoners • Ability to “link” risk terms with existing semantic descriptions in the Web of Data • building a network of semantically interconnected concepts (Linked Data)

  5. The Role of the Risk Knowledge Portal • To better manage and disseminate the case studies of the risk registry • Simplifies the definition and management of risk knowledge • form-based editing vs. Microsoft Excel worksheets • easier representation of risk relationships • Searching/browsing capabilities based on semantic relationships • To make the underlying risk registry public • A common web reference for risk management • Continuous updating/refinement of risk knowledge • To provide basic social features for the active contribution of users to the risk definition/refinement process • comments, discussions, etc. • Future -> Integrator for Risk Assessment

  6. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  7. Ontology Requirements • Should be able to represent all the needed terms, relationships and restrictions of the domain • e.g. case studies, risk cases, risk variables and their relationships • Should be simple • a complex ontology affects reasoning and querying performance • an important requirement in our case since there is a need for online (real-time) reasoning tasks

  8. The Ontology • OWL 2 RL Language • 27 classes • 31 properties • A revision of a previous risk ontology (2010) • modified to meet the requirements of the portal http://irisportal.csd.auth.gr/ontology/iris.owl

  9. Top-level Class Hierarchy

  10. Case Study • There are no subclasses • Properties • acronym • editor • hasRiskCases() • version • versionDate

  11. Risk Case • Properties • appearsInCaseStudy() • code • consistsOf (Risks) () • hasFactors() • hasComponents() • hasMechanisms() • hasImpacts() • editor

  12. Risk • Properties • belongsTo(Risk Case) () • hasComponent() • hasFactor () • hasImpact () • hasMechanism () • rickClass(Category)

  13. Risk Variable • Properties • isVariableOf(Risk) () • with subproperties • isFactorOf • isImpactOf • isComponentOf • isMechanismOf

  14. Category • There are no subclasses • Specific–only instances (owl:oneOf) • It is used in order to define instances relevant to categories

  15. Basic Ontology Restrictions • A Case Study may have zero or more Risk Cases • Each Risk Case belongs to a single Case Study • A Risk Case has one or more Risks • Each Risk belongs to a single Risk Case • A Risk has one or more Risk Variables (Factors, Components, Mechanisms, Impacts) • A Risk Variable may belong to multiple Risks

  16. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  17. Architecture • Risk Knowledge Portal • Web application (.NET 4 Web Forms)

  18. Risk Ontology Service • A WSDL Web Service (in JAVA) that supports operations for querying and updating the risk ontology • Front-end • receives SOAP messages from the IRIS Portal • Back-end • it communicates with Sesame (the triple store) and OWLIM Lite (rule-based ontology reasoner)

  19. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  20. Overview • User accounts • role-based authentication • Risk Ontology Management • Faceted Browsing/Filtering/Searching • Favorites

  21. User Roles • Three role-based security levels • Users: • limited access • they are allowed to search, browse/navigate the risk ontology • able to comment, share ideas and participate in discussions with other users(not implemented yet) • Knowledge Engineers: • have all the privileges of Users • they are also capable of posting new case studies and risk cases or updating the ones that they have submitted • they are not allowed to modify posts of other users • Administrators: • have all the privileges of Users and Knowledge Engineers • they are also capable of modifying the entire risk ontology • they are responsible for the management of the user accounts (not implemented yet)

  22. Posting New Risks • Two types of post • New Case Studies • New Risk Cases • with their risk variables • In order to post a Risk Case, the Case Study should already exist • each Risk Case must belong to a single Case Study

  23. Faceted Searching/Browsing • Users are able to search and browse • Case Studies • Risk Case • Risk Variable • For each type, different facets/filters may be defined • e.g. give me all the Case Studies that are associated (indirectly) with a specific Risk Variable • A dynamic SPARQL query is generated

  24. Favorites • Users can add Case Studies, Risk Cases, Risks and Risk Variables to their favorites • easy access • keep track of updates and activities

  25. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  26. Repository Queries • Dynamically generated SPARQL 1.1 queries based on users activity • searching, filtering, browsing the ontology • Queries are submitted via the Risk Ontology Service to OWLIM • it supports non-trivial inference with tens of millions of statements

  27. Sample SPARQL Query 1 Retrieve all Case Studies SELECT DISTINCT ?ID ?Acronym ?Description ?Editor ?Version ?VersionDate WHERE { ?ID rdf:typeiris:CaseStudy; iris:acronym ?Acronym; iris:description ?Description; iris:version ?Version; iris:versionDate ?VersionDate; iris:editor ?Editor. }

  28. Sample SPARQL Query 2 Retrieve all Case Studies that are related to the “blades deflections” variable SELECT DISTINCT ?CaseStudy ?Acronym ?Description ?Editor ?Version ?VersionDate WHERE { ?CaseStudyrdf:typeiris:CaseStudy; iris:acronym?Acronym; iris:description?Description; iris:version?Version; iris:date?date; iris:versionDate?VersionDate; iris:editor?Editor. ?CaseStudyiris:hasRiskCases ?RiskCase0. ?RiskCase0 iris:hasVariables iris:component622427b2-a64a-4e0a-83e4-83c4b309b618. } variable id

  29. Updates • Currently OWLIM does not support the SPARQL 1.1 UPDATE construct • is used to update the ontology via SPARQL • Will be available in the next release • We use the native OWLIM API for adding and removing triples

  30. Custom Rules • Ability to enhance the reasoning results with custom inferences • OWLIM supports the definition of rules using a triple-based rule language • Both OWL entailments and custom rules are translated into JAVA byte-code

  31. Custom Rule Example Id: AssociateRiskCaseAndComponents x <rdf:type> <iris:RiskCase> x <iris:consistsOf> r r <iris:hasComponent> c ------------------------------- x <iris:hasComponents> c

  32. Summary of Technologies • Portal • ASP.NET 4 Web forms • jQuery (Ajax-based communication) • SQL SERVER 2008 R2 (for the social features) • Risk Ontology Service • JAVA 6 • Sesame (triple store) • OWLIM Lite 4 (reasoner) • SPARQL 1.1 • Apache Tomcat

  33. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  34. Outline • Introduction • The Risk Ontology • The Risk Knowledge Portal • Architecture • Functionality • Implementation • Demonstration • Next Steps

  35. Risk Assessment Calculation • Portal will become the integrator of various risk assessment tools • The following calculations are required to assess overall risk probability • “summation” of the probability of occurrence of the components • “summation” of the probability of occurrence of risk variables occurrence, i.e. factors, components, mechanisms, and impacts

  36. Risk Assessment • The ontology will evolve in order to “host” data for probabilities and various other numerical data values • Will be re-inserted from older ontology versions • Specialized custom rules will be used to summate the values provided by the users or other tools • Special APIs will be used to communicate with these other tools (e.g. Matlab) • Database will hold past risk assessment cases if needed

  37. More Social Features • User profiles • Comments • Connections among users

More Related