Poster Session: Methods for Storing & Displaying Large Ontologies for Network Applications: A Demonstration of a Solution Architecture
Darren Govoni

This paper presents an architecture approach to efficiently retrieve, visualize and navigate very large OWL/RDF [4] based ontologies. The problem exists for certain classes of applications how to provide ontological information, which may be vast and distributed, across network locations. Because the potential for ontology information is large, it is necessary to provide a visualization metaphor that helps the user mitigate the informational complexity.


Tackling the Semantic Interoperability of Modern Manufacturing Systems
Steve Ray

The need for increased precision in information standards coupled with the desire to automate parts of the system integration process has led us to the use of a formal semantic approach to systems integration. This paper provides a brief overview of some of the relevant semantics-based work underway within the Manufacturing Systems Integration Division at NIST.


A Prototype Effort using XML and Semantic Web Technologies for Counter-Terrorism
Mike Personick, Brad Bebee, Bryan Thompson, Bijan  Parsia

In the domain of Counter-Terrorism, understanding terrorist groups and actors along with their links and
relationships is critical to being able to discern and reveal terrorist group structure. In work being performed
for a government customer, SAIC/ASC is developing and participating in an experimental network to enable research technologies to be tested by real users with real data. 

Each new information resource brought online within the experimental network has it own schema and unique conventions for how the information is stored. In bringing repositories onto the network for sharing and analysis with research tools, the team encountered three distinct challenges in the areas of lack of description, tight coupling between tools and schemas, and robustness to the addition of new information. To enable a solution to this challenge of sharing information and providing a rich set of analytical tools, the SAIC/ASC team has initiated prototyping efforts using the XML specification of XPointer and the Semantic Web languages of the Resource Description Framework (RDF) and the Ontology Web Language (OWL) . The outcome of the efforts described in this paper will be applied on an experimental network with real users using Semantic Web technologies with real data to analyze and better understand terrorist organizations and their relationships. The team believes that the application of Semantic Web technologies will enable the describing and mediation of semantic for information sharing where other efforts have not succeeded to date.


The Business Value of Semantic Technologies
Mills Davis

TopQuadrant has begun a program of research to define and better understand emerging markets and practical applications for semantic technologies. This presentation highlights some preliminary findings about the business value of semantic technologies for government and commercial enterprises, based on primary and secondary fact gathering from over 300 sources.

Mr. Davis' talk will address the following topics:

o Appreciating the value of "meaning" - How do semantic technologies drive business value?
Sources of value and lifecycle economic return.

o Show me the ROI - In what areas are industry and government enterprises generating significant ROI today from semantic technology solutions? Case examples.

o Crossing the Chasm - What is the outlook for semantic technology markets? The road ahead.


Poster Session: The LinkSpace™ Vision
Thomas L. Bascom

Semantic Web visionaries hope that one day all the information published on the Web will be so well defined that intelligent software agents will be able to assemble bits of information from across the Web into complete authoritative responses to ad hoc queries.  This is easily achieved for well characterized data (e.g. What is the outside temperature in the area of zip code 14710?)  However, credible results are impossible for complex topics where the response is multifaceted and dependent on experience and opinion, (e.g. when will the Semantic Web be available?) 

At LinkSpace™ we envision people creating an organic (human-generated) layer of organization on top of the web by semantically connecting related information.

Imagine distributed collections across the Web being intelligently integrated into one virtual inter-related
collection. Imagine content as it is added to the web being placed in context of the entire collection.

This is our vision of the Organic Semantic Web™.

Justifying Semantic-Web-based Resource Registration and Discovery
Eric Peterson

A non-gratuitous use of semantic web technology should be grounded in functional requirements and technical justification. Such a justification must not only motivate the use of an ontology but also must motivate the need for that ontology to be web-deployed. We (i) outline the conditions under which a semantic-web-based framework for the registration and discovery of web resources is justified, (ii) put forth design philosophy for the registration ontology such as including an instance model of key portions of the real world, and (iii) offer ontology evaluation metrics and design goals for maximizing precision and recall, while minimizing the cost of registration and discovery.


Poster Session: Leveraging Communities to Collaboratively Develop & Operationalize Semantic Technologies
Peter P. Yim

Semantic Technologies, applied in the presence of the ever-increasing computing power and connectivity,could well bode the advent of a paradigm shift (in the sense of Thomas Kuhn's "Scientific Revolution") and usher the real "knowledge economy". What is going to be different this time, the authorpurports, is that it is more complex, and is going to happen much faster than people may think.

In this paper, the author will share his insight on how we can collaboratively develop & operationalize semantic technologies, through developing and engaging virtual open communities. The author will explore the kind of "shared understanding" that needs to be developed; and the stakeholders, expertise and resources that have to be engaged, before we stand a chance to tackle the type of problem (which Horst Rittle calls a "wicked problem") at hand. The author will proceed to describe the CIM3 methodology, which involves the holistic application of collaboration in tools, process and people, and the “Fishnet” organizational form which is quite ideally suited for the purpose.

To demonstrate his point, the author will be citing examples and learning from the Ontolog-Forum, an open international community of practice on business domain ontologies, and their ebXML Core Component Type formalization project work.

The author concludes that open virtual enterprising over such Fishnet organizational form will afford us the agility we need to be competitive and innovative in tomorrow's knowledge economy. However, he also warns
that the key to true "open" collaboration will lie in our own "attitude towards sharing."


Panel Presentation and Poster Session: A Case Study in Progress: Semantic Data Integration at the Tennessee Valley Authority (TVA) John Morgan

this paper discusses the current state of data and application integration at TVA; the difference between data integration and interfacing applications; how TVA plans to use Semantic Information Management as an architectural approach to manage data based on a shared understanding of its business meaning; the development of an enterprise data architecture based on a common or semantic information model; and the technical architecture being implemented to support data integration on an enterprise wide level. The reader must understand and remember that this architectural vision is very much a work in progress.

TVA is America's largest public power company, with 31,658 megawatts of dependable generating capacity. TVA’s power facilities include 11 fossil plants, 29 hydroelectric dams, three nuclear plants, six combustion turbine plants, a pumped-storage facility, and 17,000 miles of transmission lines. TVA provides power to nearly 8.5 million residents in the Tennessee Valley.


Clustering Semantically Enhanced Web Search Results
Anantha Bangalore, Arun Sood, Noorullah Moghul

In this paper, we describe an approach to significantly improve the precision of web search results. The web
search result set is passed through a filter that semantically enhances the result set using biomedical concepts and semantic types to which these concepts have been assigned by the Unified Medical Language System (UMLS). The UMLS developed at the National Library of Medicine (NLM) is a widely used biomedical resource, designed to facilitate retrieval and integration of biomedical information sources. It has been previously shown that clustering web search results helps users to find relevant information, more easily and quickly. The goal of our approach is to show that, by semantically enhancing the web pages with domain knowledge from the UMLS, and clustering the results, users can find relevant biomedical information very quickly. Our approach yields much lower false alarm rates. Even though this paper focuses on the biomedical domain, we believe that this approach can be extended to other domains.


Panel Presentation and Poster Session: Towards a Semantic Web Solution for the Child Support Enforcement (CSE) Program
Nagaraj Garimalla

The standardization of data combined with the somewhat limited success in improving interstate functionality with diverse state CSE systems is paving the way for a new era in child support communications: state to state, state to federal government and stakeholder to the state and federal CSE agencies. Both the Federal Office of Child Support Enforcement (OCSE) and state child support agencies have a need for continuous improvement in the delivery of child support services to meet the program’s objectives. The need for improved services, which must be both more timely and more cost effective, is leading the child support business community to employ new and improved technologies that can support immediate case activities and the delivery of the benefits of child support services electronically, expediting both collections and disbursement of funds.

The ongoing need for new technologies is clear to meet the goal of the federal agency is to create a nationwide Child Support Enterprise comprised of disparate systems that can communicate on a “nearly real time” basis while providing a long term adaptable solution with the least impact on state and federal funds as well as performance.


Poster Session: Semantic Indexing Delivers Powerful Dynamic Classification
Claude Vogel, Joshua Powers, Andrew Podolsky

This paper pursues two related arguments: first, that “dynamic classification” provides a better model than
keyword search for end-users to retrieve relevant information and second, that semantic indexing provides an
optimal methodology for implementing such dynamic document classification.

Basic search solutions identify and retrieve the artifacts containing information that fulfill the requirements
of the query. In contrast, dynamic classification returns the “big picture.” It provides information that
encourages an individual to investigate and drill down into areas in which he hadn’t realized there was an
association; in other words, supporting a highly informed, real-time, analytic problem-solving course of action.
The browse operation becomes contextual to the query operation’s results.

We posit that the only approach which can even attempt to cope with the complexity of human needs and behavior is a semantic one. We make this argument for two reasons. First, language is the medium of communication we all share. Any search and retrieval technology is itself a form of communication (albeit an automated one) between the human end-user and the content. Second, we are convinced that the variety of non-semantic approaches that have been tried over the years have failed to deliver.

By leveraging semantic networks and hierarchical taxonomies we can combine indexing and categorization within one comprehensive loop. Classification then enables users to dynamically view all possible categories of information that may be viewed, cross-correlated, mixed and matched at will. As a result, users gain the freedom and creativity to decide for themselves how they would like to organize their knowledge space.


Poster Session: Open Vocabulary English Business Rules for Semantic Integration of Networked Databases
Adrian Walker

We describe some issues that occur in requirements gathering and software engineering,
* in single systems
* in multiple systems that are manually integrated, and
* in future systems that will have to be self-integrating

We argue that a move away from current software engineering techniquesis needed to address the issues. We suggest that direct specification, in the form of open vocabulary business rules in
English, is a useful approach. The rules are run in an interpreter/compiler that is based on some theory papers showing how to assign a highly declarative meaning. The English part of the approach is lightweight, in that there is no need to maintain a dictionary or grammar, yet the open vocabulary semantics are strict.

The approach is described in the "Semantic Web Presentation" at www.reengineeringllc.com.

Examples illustrating the approach can be viewed and run, live, online, by pointing a browser to a system called Internet Business Logic (R), also at www.reengineeringllc.com .

The live online examples provided include
* semantics based mining of medical data
* adding semantic knowledge to CIA world factbook data
* semantic matching of a retailer's requirements to a manufacturer's capabilities
* English querying of RDF data, with English explanations of the results
* merging and inter-operation of ontologies

Visitors are welcome to use the system to write and run their own examples.

Ontologies and the Semantic Web for Semantic Interoperability
Leo Obrst

This briefing focuses on the use of ontologies for semantic interoperability and integration. Information technology has evolved into a world of largely loosely coupled, heterogeneous systems and as such, needs increasingly more explicit, machine-interpretable semantics. 

Ontologies in the form of logical domain theories and their knowledge bases offer the richest representations of machine-interpretable semantics for systems and databases in the loosely coupled world, thus ensuring greater semantic interoperability and integration. Recently, the Semantic Web has emerged as a vision for the next generation of the Web, that addresses the explicit representation of semantics for Web artifacts. Semantic Web languages such as the Resource Description Framework/Schema (RDF/S) and the Web Ontology Language (OWL), and rule languages such as RuleML and the Semantic Web Rule Language (SWRL) promise to enable that explicit representation of semantics for our documents, systems, and services. We discuss how ontologies and the Semantic Web support semantic interoperability in the real, commercial and governmental world.

Poster Session: The "A-semantic Platform": Solving Basic Semantic Web Problems in Security-related Fields
Z. Bjelogrlic and D. Norheim @semantics

The paper presents some of basic problems found building Semantic Web applications in security-related fields. First experiences are presented and our approach is introduced.

The problems range from mixing free, commercial and classified information, through management of dynamic data, work in federated environment, performances problems and internationalization. Similar problems can also be found in commercial applications.

Our approach starts from a basic architecture layer build on top of information sources. This layer is dedicated to RDF storage, data management including gathering and editing of data and it uses a Resource Registry component which is able to work also in a federated/distributed environment. We call this architectural layer the "A-semantic Platform" since no application specific semantics are being used here.


Poster Session: Semantic Interoperability - Convergence of AI Theory and Traditional Knowledge Practice.
Denise Bedford

While there is no silver bullet solution to achieving semantic interoperability, some practical advances are being realized through the integration of artificial intelligence theory, semantic systems and computational linquisitics, and traditional methods of knowledge organization. This poster session will provide a big picture view of how an integration of these tools and techniques can create a semantically interoperable information environment.
 

Semantics: Delivering One Language to the Enterprise
Zvi Schreiber

This session will focus on how semantics are being applied today at leading enterprises to manage data resources - a best practice known as Semantic Information Management (SIM). The talk will show how semantics takes the enterprise from thousands of incompatible data languages to one common business language, captured in an ontology model. This approach does not disrupt the underlying data sources thereby providing maximum impact for minimum cost and disruption. Semantic Information Management provides a foundation for quicker and more flexible process integration and business intelligence while allowing IT costs to be reduced by eliminating redundancy and increasing productivity. Using federated query technology, the common business language may be converted into a virtual enterprise database, a methodology known as Semantic Enterprise Information Integration (EII).