Web Chat with Government Computer News Brand Niemann

Web Chat with Government Computer News Brand Niemann (US EPA), Chair, Semantic Interoperability Community of Practice (SICo. P) Best Practices Committee (BPC), Federal CIO Council January 11, 2006 http: //web-services. gov/ http: //colab. cim 3. net/cgi-bin/wiki. pl? SICo. P http: //colab. cim 3. net/cgibin/wiki. pl? DRMImplementation. Through. Iterationand. Testing. Pilot. Projects 1

Questions • Q 1: We know the DRM was created to help agencies share information. Now that version 2 of the DRM has been released, what steps can agencies take to incorporate the DRM into their systems? • Q 2: What can we expect at the upcoming 4 th Semantic Interoperability for E-Government Conference, coming up February 9 -10? • Q 3: What sorts of advanced tasks can you execute using documents with well-formed metadata? • Q 4: What did you find significant about that recent paper on using Wikis for government information? • Q 5: Where is all of this leading us? 2

Question 1 • We know the DRM was created to help agencies share information. Now that version 2 of the DRM has been released, what steps can agencies take to incorporate the DRM into their systems? – First, they need to read the new E-Gov Report and comply with the new FEA’s Enterprise Architecture Assessment Framework Version 2. 0, especially Section 1. 3. 3 Data Architecture (Information Management). (See slides 4 -5). – Second, I suggest they review the DRM Implementation Guidance Paper the DRM ITIT* Team prepared that summarized about six months of very extensive discussions and piloting so far (See slides 6 -9). *DRM ITIT: DRM Implementation Through Iteration and Testing. See slide 1 for URL. 3

Question 1 Source: Expanding E-Government, Improved Service Delivery for the American People Using Information Technology, December 2005, pages 2 -3. http: //www. whitehouse. gov/omb/budintegration/expanding_egov_2005. pdf 4

Question 1 • EAAF 2. 0 Section 1. 3. 3 Data Architecture (Information Management): – Example for Level 3* (out of 5) Practices: • Activities: The agency has created a high-level target data architecture that identifies opportunities for information sharing and consolidation. • When applicable and required by law and policy, the agency has prepared and published inventories of the agency's major information holdings and dissemination products, and otherwise made them available for use by all interested and authorized parties including other agencies and as appropriate, the general public, industry, academia, and other specific user groups. • Artifacts: Target Data Architecture *Level 3 Maturity: Utilized – EA processes and products are documented, under 5 stood, and are being used in at least some agency decision-making activities.

Question 1 Metamodel by Andreas Tolk, 2005. • • See next slide for explanation. What is it? Taxonomies and Ontologies for describing information relationships and associations in a way that can be accessed and searched. What am I expected to do? Use the DRM Abstract Model to guide both your agency data architecture and your interagency data sharing activities. What are some best practices for doing it? See Ontology and Taxonomy Coordinating Work Group, etc. How do I work both locally in my Agency and more globally with other agencies on this? Participate in the Collaborative Workshops, the DRM ITIT Team, etc. 6

Question 1 • Explanation of Metamodel by Andreas Tolk (2005): There are four rectangular boxes on top of one another (labeled from bottom to top: data, metadata, model, and metamodel, respectively) and each box contains 2 -4 circular colored dots, and these colored dots are connected with lines, meant to show that there are relationships, or need to be relationships, between say data and metadata, between metadata and models, and between models and metamodels. The purpose is to show that we need to describe information model relationships and associations in a way that can be accessed and searched. Note: This is also provided for Section 508 Compliance of the graphics. 7

Question 1 Mapping DRM Abstract Model to OMB Section 207 d / DRM Guidance DRM Abstract Model Best Practice Example (Joe Chiusano, Nov. 18, 2005) (preliminary – subject to SICo. P ONTAC WG Review) Data models Data Description Numerous Data dictionaries Data Description UBL 1. 0 Thesauri Data Context CSA/NBII Taxonomies Data Context JPL Topic maps Data Context IRS Ontologies Data Context FEA-RMO & UMLS-SN Controlled vocabularies Data Context World Bank Exchange packages Data Sharing NIEM 0. 1 See Footnote 6 (Dec. 16, 2005) 8

Question 1 Use DRM Version 2. 0 itself as a pilot project for education and FEA information sharing! See http: //web-services. gov and Dynamic Knowledge Repositories 9

Question 2 • What can we expect at the upcoming 4 th Semantic Interoperability for E-Government Conference, coming up February 9 -10? – Another in a series of very successful conferences of a growing Community of Practice where social learning occurs because the common interest in Semantic Interoperability problem has led to collaboration over an extended period to share ideas, find solutions, and build innovations. – The agenda is posted at the SICo. P Wiki: • It includes Keynote Presentations and Demonstrations, White Paper Presentation and Panel Discussion, Work Group Reports, Featured Presentations and Demonstrations, Posters and Vendors Exhibits, and Tutorials. – See http: //colab. cim 3. net/cgibin/wiki. pl? Fourth. Semantic. Interoperabilityfor. EGovernment. Conference_ 2006_2_0910 – The February 10, 2006, Opening Session will feature a report by the DRM 2. 0 Implementation Through Iteration and Testing Team and Open Discussion! 10

Question 2 • Some Highlights: – Opening Keynote: Introduction to the Semantic Web for Bioinformatics, Ken Baclawski Northeastern University. • See K. Baclawski and T. Niu, Ontologies for Bioinformatics, MIT Press, October, 2005. – Opening Demonstration: Boston Children's Hospital "smart search" and Semantic UMLS Ontology-based Professional Language Processing Pub. Med Search, Michael Belanger, Seman. Tx. Life Sciences. • See Ontology Graph-Assisted Search at http: //gw. jarg. com/jsp/index. jsp – Some Featured Presentations: • Semantic Wave 2006: Executive Guide to the Business Value of Semantic Technologies, SICo. P White Paper Series Module 2, Mills Davis, Project 10 X. • HL 7 RIM, Professor Barry Smith, State University of New York at Buffalo and Co-Director of the National Center for Ontological Research. – Closing Keynote and Challenge: Professor Jim Hendler, University of Maryland. • See “From Atoms to OWLs: The new ecology of the WWW, ” Opening Keynote for the XML 2005 Conference, November 15, 2005. 11

Question 2 http: //gw. jarg. com/jsp/index. jsp 12

Question 3 • What sorts of advanced tasks can you execute using documents with well-formed metadata? – First, understand the Evolution of Metadata – see next slide. – Second, consider the excellent Semantic Continuum Diagram by Mills Davis* – see slides 15 -16. • This diagram depicts a spectrum of information sharing capabilities. Moving from lower right to upper left of the diagram, we see that more expressive forms of metadata and semantic modeling encompass the simpler forms, and extend their capabilities. From bottom-to-top, the amount, kinds, and complexity of metadata and modeling increases. From left-toright, capabilities for intelligent search, retrieval, discovery of relevant information, question answering, and autonomous action advance. *Source: SICo. P White Paper “Semantic Wave 2006: Executive Guide to the Business Value of Semantic Technologies, ” 55 pages, Version 1. 0, 1/9/2006. 13

Question 3 • The Evolution of Metadata: – In the beginning there was data, and hopefully its documentation – but it was not accessible so we resorted to: – “Metadata for Discovery” – but we still wanted to see the actual data – now both are on the Web. – “Metadata for Integration” – but that is really hard. – And now the new paradigm is “Executable Metadata” – the data (XML), metadata (RDF), models (RDF/S) and metamodels (OWL) are all integrated to support knowledge computing, statistical computing, and stochastic inference under conditions of uncertainty referred to as the “Bayesian Web: • See "Ontologies for Bioinformatics“, Ken Baclawski and Tianhua Niu, MIT Press, October 2005: http: //ontobio. org/ • And see the National Center for Ontological Research (NCOR): http: //ncor. us 14

Question 3 See next slide for explanation. 15

Question 3 • The role of semantic metadata in increasing search capability: – In this XY graph, the X axis is labeled "Increasing Search Capability" (with sub-labels of Recovery, Discovery, Intelligence, Question Answering, and Reasoning) and the Y-Axis is labeled "Increasing Metadata" (with sub-labels from Weak Semantics to Strong Semantics). A straight line from the origin to the upper right has labels of Syntactic Interoperability (sub-label "Many Federal applications do not enable data sharing"), Structural Interoperability (DRM 2. 0 sets the bar here), and Semantic Interoperability (Some Intelligence, Defense, Security, Health, Science & Business applications share information at these levels) from bottom to top. The point of this XY graph is that Increasing Metadata (from glossaries to ontologies) is highly correlated with Increasing Search Capability (from discovery to reasoning). Note: This is also provided for Section 508 Compliance of the graphics. 16

Question 4 • What did you find significant about that recent paper on using Wikis for government information? – DRM 2. 0 used the Collaborative Wiki and SICo. P and the DRM ITIT used the same two layer information model – see next slide – and the same idea about implementing it in Wiki’s – see slide 19. – Supports SICo. P’s White Papers and Pilots for the Knowledge Management Working Group of the Best Practices Committee – see slide 20. • Building Semantic Webs for Knowledge Management and ELearning in E-Government. * • Semantic Technologies for Knowledge Work Automation & Knowledge Worker Augmentation. * * SICo. P Papers Submitted for the 7 th Annual Knowledge Management Conference 17 Featuring the E-Learning Solutions Seminar, April 19 -21, 2006.

Question 4 Information Model* Two Connected Layers: Knowledge Map and the Information Resources Topic Maps Ontology Topics Concepts Associations Relationships (between Topics & Occurrences) (between Concepts and Instances) Occurrences Instances * SICo. P and DRM Implementation Through Iteration and Testing: Making It Real, Federal Metadata Management Consortium, December 13, 2005. http: //web-services. gov/scopefmmc 12132005. ppt Also used in “Building Semantic Webs for e-government with Wiki technology. ” 18 http: //colab. cim 3. net/file/work/SICo. P/2006 -02 -09/EGov%20 Wiki. pdf

Question 4 • Some New Initiatives: – “What Weblogs and RSS Feeds are about is that you make it very easy to write something that you can think of, like an email, but it goes up onto a Web site…and it doesn’t interfere with your normal inbox. ” • Bill Gates, Microsoft CEO Summit, May 20, 2004. – “It was suggested that recent self-organizing and information-sharing tools from the Internet, the Wiki and the Blog, be deployed on the SIPRNet. ” • The Wiki and the Blog: Toward a Complex Adaptive Intelligence Community, D. Calvin Andrus, Central Intelligence Agency, July 2, 2005. – Keynote Debate: Blog, Wiki, and RSS Technology - Are they Enterprise Ready? Applicable? Or a Passing Tempest in a Teacup? • The Gilbane Content Management Technologies Conference, November 29 December 1, 2005. – Quickly build light-weight tools like “Semantic Wiki’s” that let people casually start modeling things. • Fourth International Semantic Web Conference, November 6 -10, 2005, Galway, Ireland. Using Co. Ps To Simplify Processes and Unify Work Across Agencies: Cross-Industry Applications, KM Collaboration Conference, Orlando, Florida, December 5, 2005. http: //web-services. gov/sicopkmc 12052005. ppt 19

Question 4 • Three Steps in Work Flow: – Conceptualization – Markup – Execution (searching, reasoning, rules, etc. ) • Tools for Each Step (examples): – Suggested Upper Merged Ontology-Word. Net now and Common Semantic Model (COSMO) Agent in 2007. – Collaborative Ontology Development Infrastructure Services (CODS-multi-user Protege) now and Standardized Wiki Semantic Markup Language and Mapping Tools in 2007. – Oracle 10 g R 2 now and Oracle 11 g in 2007. 20

Question 5 • Where is all of this leading us? – A Knowledge Reference Model that: • Builds on DRM 2. 0 Concepts – see next slide. • Brings Knowledge Management and Enterprise Architecture Together with Semantic Models. • Really Addresses the Search/Metadata Issue. • Supports SICo. P White Paper Module 3 – Implementation of the Semantic Web for EGovernment. • Delivers Measurable Semantic Interoperability Benefits. 21

Question 5 • Data: Three Types – structured (20%), and unstructured and semistructured (80%). – Originally it was the Data and Information Reference Model. • Metadata: Three Roles – discovery, integration, and reasoning. – Recombine data and metadata for sharing and reuse and address Section 207 d requirements (recall slide 8). • Model: Three Functions – description, context, and sharing. – DRM XML Schema and DRM Abstract Model. • Reporting: Three Documents – reference, management strategy, and implementation guide. – Integrated in the DRM Education Pilot with Pilot Metrics and Co. P/Co. I Templates. • Metamodel: Three Implementation Levels – organizational, technical, and semantic interoperability or agency, Co. I, and cross. Co. I. – European Interoperability Framework, Tolk Metamodel, Enterprise Architecture Assessment 2. 0, Do. D Net-Centric Strategy, etc. 22

Question 5 Paradigm Shifts • FEA Reference Model Taxonomies Ontology • FEA “Common • FEA Semantic Model Language” • DRM 1. 0 by • DRM 2. 0 by open, committee collaborative process – Implementation after development. – Implementation though iteration and testing during development. 23

Question 5 • Imagine we did DRM 2. 0 the “concept ontology” way and started by listing all the concepts we wanted to include (beyond those in DRM 1. 0) and then got an “ontological commitment” to their definitions and instances. – Ontological commitment is a social contract between those that need ontologies and those that know how to develop them for the purpose of information sharing (Gruber, 2004) • I believe we would have ended up with an even better result in less time! – So lets try an experiment! – see slides 25 -26. 24

Question 5 • Introduce a concept in the form of a question. • Answer that question with a definition and an instance that illustrates the relationship we mean between the concept and the instance. • Provide a flow of concepts and instances that supports logic and reasoning. • This illustrates the Knowledge Reference Model we are working towards! 25

Question 5 • The Suggested First Experiment – The Search/Metadata Issue: – Enter the GSA/RFI Report and Other Search/Metadata Resources Into the Collaborative Wiki to Create An Initial Content Layer with Some Structure and Sub-document Link References (“purple numbers”). – Convene a Collaborative Workshop to develop a list of concepts and make assignments to develop their definitions and instances in the Collaborative Wiki after the workshop. – Convene a Collaborative Workshop(s) to vet the list of concepts and their suggested definitions and instances. – Convene a Collaborative Workshop(s) to reconcile any remaining differences and issue an authoritative “living document” in the Collaborative Wiki. 26

Question 5 Where is SICo. P DRM Implementation Going? Stage in Evolution Best Practice Example(s) Knowledge Computing Visual. Knowledge – ”Bio. CAD” NLM – “UMLS-SN” and Siber. Logic Digital Harbor – “Business Ontology” Metallect – “Google for IT” Smart Content Composite Applications Operational Enterprise Architecture Source: Mills Davis, http: //web-services. gov/Net. Centric. Semantics 051110. pdf 27

Question 5 • Pilot presentations and demonstrations at our DRM Collaborative Expedition Workshops: – Metallect – “Google for IT”: • July 19, 2005. – Digital Harbor – “Business Ontology”: • June 13, 2005, and September 14, 2005. – NLM – “UMLS-SN” and Siber. Logic: • October 5, 2005. • September 14, 2005. – Visual. Knowledge –”Bio. CAD”: • December 6, 2005 28

Question 5 Where is SICo. P DRM Implementation Going? Super Pilot: Address as Many Boxes as Possible! Scale of Activity / Metadata Function Discovery Agency e. g. , EAAF 2. 0 Co. P/Lo. B Cross-Co. P/Lo. B e. g. , FHA/DAWG e. g. , Indicators Integration Reasoning ? ? Yes Co. P: Community of Practice Lo. B: Line of Business FHA/DAWG: Federal Health Architecture – Data Architecture Work Group See FHA Data Architecture Working Group: SICo. P DRM 2. 0 Pilot, December 28, 2005. http: //web-services. gov/scopefhadawg. ppt 29

Question 5 Where is SICo. P DRM Implementation Going? • Five Key Activities Over the Next Year: – (1) Education and Training in DRM Version 2. 0 and use in FEA – DRM-based Information Sharing Pilots (started June 13 th). – (2) Testing of XML Schemas and OWL Ontologies by NIST and the National Center for Ontological Research, respectively, among others (began October 27 th). – (3) Inventory/Repository of Semantic Interoperability Assets and Development of a Common Semantic Model (COSMO) by the new Ontology and Taxonomy Coordinating Work Group (ONTACWG) (started October 5 th). – (4) Continued early implementation of DRM 2. 0 concepts and artifacts by industry in “open collaboration with open standards” pilot projects and workshops (started July 19 th). – (5) Fostering champions of DRM Best Practices to improve (1) agency data architectures within agencies and (2) cross-agency data sharing across agencies in funded projects (started June 13 th). 30