Creating Relevance and Reuse With Targeted Semanti

Creating Relevance and Reuse withTargeted Semantics: 

Creating Relevance and Reuse with Targeted Semantics Michael C. Daconta Metadata Program Manager November 16, 2004

Slide2: 

Welcome & Introduction Michael C. Daconta Author/co-author of 10 Technical books C, C++, Java, XML, Semantic Web XML experience: First XML project: 1996 Wrote XML Courses: 1998 Lead for the DOD/IC Core Taxonomy Chief Architect of the DIA’s Virtual Knowledge Base Inventor of the Fannie Mae Electronic Mortgage Standard http://www.daconta.net High Five! XML has won!

The Department of Homeland Security: 

The Department of Homeland Security The “Mother of all Integrations” CIO Office and CIO Council leading the charge… Lee Holcomb, Steve Cooper Martin Smith, Ryan Cast, Lee Smith, Sean Thrash … many others! The Information Sharing Imperative! Metadata Center of Excellence Expose, Standardize and Federate DHS Metadata Working Group Presidential Orders EO 13356 and HSPD 11 We will aggressively execute this mission! The Train has left the station! Are you on board?

Slide4: 

An “Aggressive” Agenda Modeling and the State of Things Evolution and Revolution Entity-based XML XML Profile of FEA DRM The Three R’s of Semantics Targeted Semantics Semantic Bootstrapping DHS CIO Efforts Other Efforts The Challenge A lot to cover so … Hang On!

Modeling and the State of Things…: 

Modeling and the State of Things… The “Messy Reality Scale” affects how you view modeling. My View: There are stable subsets of reality that can be effectively modeled. We all use internal “maps”. Maps are not supposed to be reality, just effective. And … the need is great! And … Hardware is still improving: GPS, cell-phones and voice recognition will ratchet up the requirements for “real-time relevance!” 64 Bit Computing So … Subjective Unpredictable XML is necessary but not sufficient! Objective Stable

Slide6: 

Data Evolution Timeline Age of Programs Age of Proprietary Data Age of Open Data Age of Open Metadata Age of Semantic Models Program-Data GIGO/minis/micros www / Netscape Web services OWL Text, Office Docs Databases (proprietary schema) HTML, XML (open schema) Namespaces, Taxonomies, RDF Ontologies & Inference 1945 -1970 2000 - 2003 1994 - 2000 1970 - 1994 2003 - Procedural Programming Object-Oriented Programming Model-Driven Programming “Data is less important than code” “Data is as important as code” “Data is more important than code”

Slide7: 

Entity Based XML Lessons Learned: Separate entities from transactions Lesson: Use RDF data model to design your XML schemas. Examples: DDMS Global Justice XML Data Model (GJXDM)

Entity-Based XML(2) : 

Entity-Based XML(2) Biometrics Automated Toolset Army Fielded System (i.e. Centcom) Functions Biometrics: Fingerprints, Iris, Facial Recognition GUID “Dossiers” XML Driven. Below is “entities.xml” <ENTITIES> <ENTITY name="Person" tablename="PersonalData" XMLBaseName="FORMS" representativeFields="LNAME;FNAME;MNAME;SVCIDNO;ALERT"> <RELATIONSHIP withentity="Account" relationshipname="Related Account" tablename="Pers_Acct_Junct" entity1="Person" entity1key="PERS_GUID" entity2="Account" entity2key="ACCT_GUID"></RELATIONSHIP> … </ENTITY><!-- other entities Account, Facility, Equipment, Location, etc. … --> … </ENTITIES>

Slide9: 

FEA DRM Structure

Slide10: 

Revised DRM Structure

Slide11: 

XML Profile of the DRM Subject Context Security Context Service Context Information Access Data Element Description Resource Description Who What When Where Why Context Information Exchange Sharing Description Structured Semi-Structured Unstructured XML Topic Map (XTM) Web Service Entry (UDDI) IC Information Security Marking “Document” XML Schema Federated Query & RSS Dublin Core/DDMS XML Schema Types XLINK XLINK/XPointer XLINK

Slide12: 

The Road to Revolution Done with Graphics Fidelity … shifting to Data Fidelity Smart Data Continuum “Smart” = Application Independent, Contextual and enables Inference Relevance Reuse © Microsoft

Slide13: 

Revolution (2): Copernican Shift Apps Data We had it all wrong… In the beginning… Put the “smarts” in the data… watch what happens… Apps Apps Apps

Slide14: 

Three R’s of Semantics Relevance The intersection of what the user wants and what is available. Reuse The repurposing of the smallest semantic chunk of content. Relation First-class Associations between things.

Relevance: Right Information to the Right Person at the Right Time: 

Relevance: Right Information to the Right Person at the Right Time - 5W Search Axes - Taxonomy - Query - Visualization - Question/Answer - Profiles - Feedback Loop - Entity/Associations - Class/Property - Class/Instance - Predicate/Inferences - System and Domain Metamodels Structured - Semi-Structured - Unstructured - Web-Services - RSS Bottom Line: You have to model what you mean by “Right”

Slide16: 

Personal Observations on Relevance Missing the Mark Traveling to Las Cruces, New Mexico Off by one in a known email address What should have happened Relevance should enable “What is best for me…” Context should enable “What I mean versus What I say…” Tucson Arizona Denver Colorado Albuquerque New Mexico Las Cruces New Mexico Literal path Optimal path

3R’s: Reuse: 

Solution Set: XML Taxonomy Wordnet Entity Extraction Multiple Search Axes: Time, Geography, People, Subject… 3R’s: Reuse Problem: Reusable, Modular Content Semantic Chunks Doctrine, Video, Entities Digital production Workflow / Transactions Linchpin Granularity people

3R’s: Relation: 

3R’s: Relation Standard Associations Example: PartOf versus SubclassOf

Targeted Semantics: 

Targeted Semantics Precedent: general natural language recognition is far-off, domain-specific is here today (e.g. airline reservations) Focus: 5W Search Axes: Who, What, When, Where, Why Entities, Associations, Events Community of Interest modeling Go with the Stable Subset Precedent: Programming Language Features Stick with Standards Sequence is Important Semantic Bootstrapping

Semantic Bootstrapping: 

Semantic Bootstrapping

Slide21: 

Current DHS Efforts Introduction to Metadata Flash DHS Taxonomy Enterprise Metadata Registry DHS Metadata Working Group Core XML Schema Types Standards Adoption Process DHS Data Strategy

Slide22: 

Semantic Web Central http://www.semwebcentral.org DARPA sponsored Tools! Tutorials! Resources!

Slide23: 

Your Challenge Get Positive! The only security is action! Get Involved! Semantic Interoperability Community of Practice (SiCOP) http://www.web-services.gov Get Going! After the Semantic Web is the Pragmatic Web