Upload
hannah-gill
View
214
Download
0
Tags:
Embed Size (px)
Citation preview
1
DAML PI Meeting Status BriefingDAML PI Meeting Status Briefing
Dynamics Research Corporation
Marti Hall
Lee Lacy
February 12-14, 2002
2
AgendaAgenda
Overview of DRC’s DAML Work DAML Military Ontologies
Light Weight Reusable Ontologies
Military Knowledge Representation Methodology
Lessons Learned
Quality Assurance Process (Mist)
DRC’s Upcoming Work
Deliverables
Metrics
3
Overview of DRC’s DAML WorkOverview of DRC’s DAML Work
Focused on Military Applications
Developed Methodology to Solve Complex Military Problems by Building Ontologies/Artifacts from Light Weight (Primitive/Basic) Ontologies
Developed Methodology for Our Quality Assurance Process for Ontologies and Artifacts
Investigated utility of DAML to provide information to the explosive ordnance disposal (EOD) specialist
Investigated utility of DAML to solve USAF Air Mobility Command problem (Foreign Clearance Guide)
4
DAML Military OntologiesDAML Military Ontologies
Task list
Explosive Ordnance Disposal (EOD) scenario/vignette
Event chronology
Fugitive/terrorist description
Military land platform taxonomy
Commercial shipping
Hazardous materials
Foreign Clearance Guide
Equipment Characteristics and Performance (C&P)
IMO intelligence report
FBI Most Wanted Terrorist
Center for Army Lessons Learned Thesaurus
5
Task List OntologyTask List Ontology
Supports representation of military task lists (e.g., UJTL, NTL, AUTL)
Populating sample instance file with Universal Joint Task List (UJTL)
OPERATIONALAccomplish Objectives ofAccomplish Objectives of Subordinate Campaigns and Major OperationsSubordinate Campaigns and Major Operations
STRATEGIC NATIONALAccomplish Objectives ofAccomplish Objectives of National Military StrategyNational Military Strategy
OP 1CONDUCT
OPERATIONALMOVEMENT & MANEUVER
OP 2 DEVELOP OPERATIONAL
INTELLIGENCE, SURVEILLANCE, AND RECONNAISSANCE
OP 3EMPLOY
OPERATIONALFIREPOWER
OP 4PROVIDE
OPERATIONALSUPPORT
OP 5EXERCISE
OPERATIONALCOMMAND & CONTROL
OP 6PROVIDE
OPERATIONALPROTECTION
SN 1 CONDUCT STRATEGIC
DEPLOYMENT &REDEPLOYMENT
SN 2 DEVELOP STRATEGIC INTELLIGENCE,
SURVEILLANCE,AND RECONNAISSANCE
SN 3EMPLOYFORCES
SN 4PROVIDE
SUSTAINMENT
SN 5 PROVIDESTRATEGIC DIRECTION
& INTEGRATION
SN 6CONDUCT
MOBILIZATION
SN 7CONDUCT
FORCEDEVELOPMENT
TA 1DEPLOY/CONDUCT
MANEUVER
TA 2DEVELOP
INTELLIGENCE
TA 3EMPLOY
FIREPOWER
TA 4PERFORM LOGISTICS AND
COMBAT SERVICE SUPPORT
TA 5EXERCISE COMMAND
& CONTROL
TA 6PROTECT THE
FORCE
ST 5 PROVIDE THEATER
STRATEGIC COMMAND
AND CONTROL
ST 6PROVIDE THEATERPROTECTION
ST 7 ESTABLISH THEATER
FORCE REQUIREMENTS AND READINESS
ST 8 DEVELOP AND MAINTAIN
ALLIANCEAND REGIONAL
RELATIONS
ST 1 DEPLOY, CONCENTRATE,AND MANEUVER
THEATER FORCES
ST 2 DEVELOP THEATER STRATEGIC
INTELLIGENCE, SURVEILLANCE, AND
RECONNAISSANCE
ST 3EMPLOY THEATER
STRATEGICFIREPOWER
ST 4SUSTAINTHEATER FORCES
STRATEGIC THEATERAccomplish Objectives ofAccomplish Objectives of Theater and Campaign StrategyTheater and Campaign Strategy
TACTICALAccomplish Objectives ofAccomplish Objectives of Battles and EngagementsBattles and Engagements
SN 8 FOSTERMULTINATIONAL
AND INTERAGENCY RELATIONS
6
EOD Scenario/VignetteEOD Scenario/Vignette
Developed specific task list for scenario of providing EOD support to clear mines from Straits of Hormuz and open lanes from Saudi Arabia to the oilfields
References tasks from UJTL instance file (shown on previous slide)
Potential applications include support for doctrine development, training, and operations (e.g., Joint EOD Mission Support Center and Decision Support Tools)
7
Fugitive / Terrorist Description OntologyFugitive / Terrorist Description Ontology
Based on FBI website information
Potential applications for “watch list” matching
Description properties include: Place of birth
FBI caution
Physical description
Languages used
8
Light Weight Reusable OntologiesLight Weight Reusable Ontologies
Locator
Point of contact
Versioning Element Set (VES)
Dublin Core (DC)
Person
Bibliographic information
Thesaurus – ANSI NISO Z39.19
9
Military Knowledge Representation MethodologyMilitary Knowledge Representation Methodology
Solving complex military problems requires knowledge representations of tasks, conditions, behaviors, units, and equipment
Our knowledge representation methodology: Develop a limited but realistic scenario
Build up knowledge representations by combining lightweight, reusable, inter-connectable ontologies (e.g., bibliographic references, military equipment)
Develop sample instance data
Develop prototypes of applications that employ the representations
10
Lessons LearnedLessons Learned
Common repository of instance data (artifacts) needed.
Problems with sites changing content without changing version number and problems with sites changing URIs or dropping out of existence.
Improved tools such as mark-up tools, ontology development tools, validation, complete set of test cases
Problems representing procedural concepts
Needed quality assurance process for test environment (answered it with a methodology we call, “The Mist”)
11
Quality Assurance Process (Mist)Quality Assurance Process (Mist)
The Mist is an implementation of a quality assurance methodology.
It consists of a testing environment for newly developed or new versions of both ontologies and artifacts.
It was developed to take advantage of the DAML Validator and the RDF Validator web-based tools.
The solution is a testing directory on DRC DAML website in which Ontologists have read/write privileges. The directory was dubbed the “Mist.”
The Mist also serves as the inbox for publish-ready ontologies and artifacts.
The Mist preserves the integrity of published DAML files, while empowering Ontologists to validate new/revised DAML files via web-based tools.
12
Continue EOD Work Support Joint EOD technology demonstration by providing
ontologies, DAML artifacts
Support IPT meetings
Participate in the DAML Experiment Provide linkage from Afghanistan scenario to UJTL
Represent the national goals and objectives
Provide various supporting ontologies (e.g., weather ontology)
DRC’s Upcoming WorkDRC’s Upcoming Work
13
DeliverablesDeliverables
EOD Requirements analysis documents for EOD knowledge representation
for EOD DSS (Word document)
Data model diagrams (IDEF1X or UML tool files)
DAML ontologies and sample artifacts (DAML files)
DAML Experiment Artifact tying Afghanistan scenario back to UJTL (DAML file)
Related ontologies, e.g., weather ontology and UJTL conditions (DAML files)
Artifact representing the national goals and objectives (DAML files)
14
MetricsMetrics
Developed 26 ontologies
Developed 26+ artifacts
Supported 4 military customers (NWDC, CALL, EOD, AMC)
Transitioning 2 programs to other funding sources (EOD, FCG)
15
Questions?Questions?
http://orlando.drc.com/DAML/
16
BACKUP MATERIALSBACKUP MATERIALS
17
What is the Mist?What is the Mist?
The Mist is an implementation of a quality assurance methodology.
It consists of a testing environment for newly developed or new versions of both ontologies and artifacts.
It was developed to take advantage of the DAML Validator and the RDF Validator web-based tools. The web tools require the file under test to have a valid URL.
However, for a variety of reasons, a webmaster should be the only one allowed make changes to website content.
While under test, a DAML file is likely to change too frequently to be officially published by a webmaster.
This conflict created a need for a location wherein an Ontologist can “publish” developing DAML files for testing purposes.
18
The Mist SolutionThe Mist Solution
The solution is a testing directory on DRC DAML website in which Ontologists have read/write privileges. The directory was dubbed the “Mist.” One constraint on use of the Mist is that the resulting URI for a DAML
file under test, must be virtually identical to the URI it should have once officially published.
Thus the only difference between the a Mist URI and an officially published URI is the inclusion of the .../mist/… directory, i.e.:
http://orlando.drc.com/daml/ontology/Locator/G3/Locator-ont-g3r1.daml
http://orlando.drc.com/daml/mist/ontology/Locator/G3/Locator-ont-g3r1.daml
19
Publishing from the MistPublishing from the Mist
The Mist also serves as the inbox for publish-ready ontologies and artifacts. The Ontologist simply needs to inform the Webmaster that a
particular DAML file in the Mist is ready for publication.
The Webmaster can then remove references to the Mist directory from the validated DAML file and move it to its official location.
The Mist preserves the integrity of published DAML files, while empowering Ontologists to validate new/revised DAML files via web-based tools.
20
Event Chronology OntologyEvent Chronology Ontology
Potential application for intelligence report representation (event-centric)
Populated sample file with events from 9/11 based on CNN chronology
Populated another sample with FBI chronology of Atta activities prior to attacks
21
Military Land Platform OntologyMilitary Land Platform Ontology
Ontology focused on a taxonomy of equipment types
Ontology is modeled after Distributed Interactive Simulation (DIS) Entity Enumerations document
22
Ontology Development
Commercial Shipping OntologyCommercial Shipping Ontology
• Define Commercial Ship• Classes• Subclasses• Attributes• Instance Data
“Authoritative” Data Source
Data Analysis &
Decomposition
23
Foreign Clearance GuideForeign Clearance Guide
AFRL conducting research to reduce cost and time required to obtain clearances from foreign governments
Focused on lead times associated with diplomatic (DIP) clearances
Migrated IDEF1X data model to DAML+OIL ontology
Knowledge acquisition at Scott AFB in joint project with BBN
24
Equipment Characteristics and PerformanceEquipment Characteristics and Performance
Equipment descriptions used by simulation applications for accurately representing platforms
Leveraged DRC work for Army Modeling and Simulation Office (AMSO)
Sample artifact developed based on Universal Threat System for Simulators (UTSS) sample data for AH-64A
25
DAML Development for UTSS ExperimentDAML Development for UTSS Experiment
Created an Entity DAML ontology with Platform and Munitions (based on DIS Entity Enumeration taxonomy)
Created UTSS-specific DAML ontology tied to Platform ontology (i.e., A64A subclass of A64A class)
Translated UTSS data into DAML artifact / instance file