Piotr Kaminski University of Victoria September 24th, 2002 Integrating Information on the Semantic Web Using Partially Ordered Multi Hypersets Piotr Kaminski University of Victoria September 24th, 2002

This is not an Outline Outline Motivation Approach Solution Evaluation and Contributions Future work The Semantic Web as Hyper-pomsets

The Semantic Web Getting machines to understand the contents of the web I'd better talk to the bank before I spend it, because of what has happened to other people Nobody owes me that much money wisdom actions Agents Your bank balance has jumped 8087% to $1234567.89 knowledge reasoning Semantic Web World Wide Web is only understandable by humans; for computers, it's just data shades of AI, but emphasis is on getting useful results The Semantic Web: data is structured, making it into information information can be reasoned upon  knowledge agents can act on knowledge  wisdom Why? can make database-like queries, instead of natural language parsing/grepping e.g. "Am I allowed to take SEng 330 next term?", "How many train lines in Japan?", "Cheapest price for this list of books, including shipping?", "Doctors in my area specializing in sports injuries with whom I could make a mutually convenient appointment tomorrow?" 1234567.89 information structure World Wide Web data The Semantic Web as Hyper-pomsets Information Interchange on the Semantic Web

Existing Solutions Document analysis Yahoo! Google Data models Relational Databases Extensible Markup Language (XML) Information models Resource Description Framework (RDF) Topic Maps Syntax Object Semantic Source Level Target Level ? The Semantic Web as Hyper-pomsets

My Approach Integrate other metamodels object-level mappings no changes to source metamodels no loss of information identify provenance Flexible and elegant simple closed (self-describing) practical Not goals: serialization, code efficiency & complexity The Semantic Web as Hyper-pomsets

The Braque Metamodel Primitives: things relationships order Presents Piotr Motivation Approach Solution Evaluation Contributions Future Work Presents Piotr Motivation Solution Evaluation Contributions Future Work Approach A1 A2 B1 C1 C2 Presents Piotr Motivation Solution Evaluation Contributions Approach A1 A2 B1 C1 C2 Future Work Partial Sets Queries … Presents Piotr Motivation Solution Evaluation Contributions Approach A1 A2 B1 C1 C2 Future Work Partial Sets Queries … Piotr Motivation Approach Solution Evaluation Contributions Future Work Motivation Approach Solution Evaluation Contributions Future Work Motivation Approach Solution Evaluation Contributions Future Work Motivation Approach Solution Evaluation Contributions Future Work Partially Ordered Multi Hypersets things sets total order links – relationships as things partial order infinite hyper The Semantic Web as Hyper-pomsets Information Interchange on the Semantic Web

Braque in Action Naïve Upper Ontology type hierarchies, relations membership reification, roles names, identifiers metatype constraint Integration mappings XML: ordered elements, qualified names RDF: types, properties, containers, reification Topic Maps: identification, classes, scopes, names The Semantic Web as Hyper-pomsets

Evaluation and Contributions Braque metamodel collections explicit identification infinite reflection depth no formal interpretation, open vs. closed worlds Integration XML, RDF and Topic Maps complete, automatic, semantically aligned no schemas, ugly class punning workaround Proof of concept implementation The Semantic Web as Hyper-pomsets

Future Work Theory mereology, mereotopology, partial sets formal theory, algebra, and interpretation query & inference language, schema ontology Implementation developer evaluation studies semantic integration with host language visualization, distributed computing Applications integration: maintain current maps; UML, UDDI modelling multi-dimensional software adoption-centric information management The Semantic Web as Hyper-pomsets

Thank You The Semantic Web as Hyper-pomsets