Upload
frederick-hardy
View
217
Download
3
Tags:
Embed Size (px)
Citation preview
Semantic Web and Valued Information at the Right Time
(VIRT)
Curtis BlaisResearch Associate
MOVES InstituteNaval Postgraduate School
RichSemantic
Track
Semantic Interoperability
Joint C3 InformationExchange
Data Model
TacticalAssessment
MarkupLanguage
USW-XML
MIP
MobilityCOP
C-BML
AutonomousVehicle Command
Language
SavageModeling and
AnalysisLanguage
AT/FP
AVCL/AUVW
FCS
GIG
Plansand
Orders
JBML
VIRT
Research Agenda
C2 Ontology-Concepts-Relationships-Rules/Constraints
CEC
JTM
CMA
Semantic Web“An extension of the current Web in which
information is given well-defined meaning, better enabling computers and people to work in
cooperation.” – Berners-Lee, et. al., 2001
• Transforming documents to information (data in context)• Enabling automated reasoning• Equally accessible to human and software agents
The Evolving WebWeb of
Knowledge
HTML/HTTP
Resource Description FrameworkExtensible Markup Language
Self-Describing Documents
Foundation of the Current Web
Proof, Logic andOntology Languages
Shared terminologyMachine-Machine Communication
1990
2000
2010
J. Hendler presentation, W3C , 2001
Explicit semantics through standard languages
Semantic Web Stack
Source: I. Herman: “Introduction to the Semantic Web,” 12 November 2003. http://www.w3.org/2003/Talks/0624-BrusselsSW-IH/26.html
C2 Ontology
Too large to be built all at onceWould take too longWould be too hard to gain widespread acceptance
Evolutionary development Start small (e.g., tracks) Show community how it is done (methodology) Demonstrate benefits (and limitations) Create mechanisms for extensionDissertation: Ontological Foundation for Obtaining Valued Information
at the Right Time in Semantically Rich Network-Centric Architectures
Topics
Applying Semantic Web Technologies to the Tactical Assessment Markup Language (TAML) – ENS Candace Childers CS Thesis, June 2006Sponsor: DoN Chief Information Officer, USW-XML Working Group
Rich Semantic Track and Valued Information at the Right Time – Rick Hayes-Roth, NPS Information SystemsSponsor: NAVSEA PEO IWS6(CEC)
Technology Goal• Demonstrate application of existing and
emerging Semantic Web standards to Network-Centric Architecture
• Explore benefits and limitations of Semantic Web technologies
• Examples:• Conceptual querying across multiple
representations• Combining multiple source data and
drawing inferences across the data
From Feasibility to Practice
• ENS Childers thesis demonstrated feasibility and breadth of application of Semantic Web standards• Tool-based demonstration (Protégé, Twinkle)• Single representation of contacts (TAML)• Single domain (USW)• Static ontology
• Unified Application (Blais dissertation)• Formalized abstraction of Track concept for general
software implementation• Mediation of multiple track representations across
multiple domains• Operator and problem-driven evolution of the ontology
Rich Semantic Track
• Abstract model of battlespace perceptions
• Logical theory enabling reasoning over collected data
• Common semantics underlying multiple systems and across multiple domains• Global Command and Control System, Maritime
Domain Awareness, Cooperative Engagement Capability, TAML, Joint Track Management, etc.
Rich Semantic Track (RST)
Rich Semantic Track- conceptual hub for
interchange and automatedreasoning
CECTrack
Messages
JointTrack
ManagementData Model
CMAMaritime
Information Exchange Model
OtherTrack
Data Models
Rich Semantic TrackTrack
• Beliefs• Identity and Characteristics• Dynamic State at Time T• History of States (past “track”)• Predicted States (future “track”)
• Meta-Information (applicable to each element of belief)• Evidence• Inferences• Error and Uncertainty Estimates• Temporal Qualifications• Spatial Qualifications
*Hayes-Roth. Towards a rich semanticmodel of Track: Essential Foundation forInformation Sharing. NPS Research Paper.Monterey, CA. February 25, 2005.
Technologies to provide Valued Information at the Right Time
Screens and marshals the “data storm”Assists in the filtering of information to
provide guidance based on situational priorities
Automatically adapts to the environment by inferring valued information
• Especially via modeling spatiotemporal “tracks”
• Other class models useful, too
Personalizes, using context, role, and state
What is VIRT?
The Basic Ideas
Synchronize groups by having them operate on semantically aligned and high-value information
Determine what concepts operators’ missions depend on and make those standard
Notice what beliefs underlie mission plans and Courses of Action (COAs)
Automatically inform operators when changes in data affect their beliefs and planning rationales
COIs
Overall Vision: Model-based Communication Networks, VIRT
and Rich Semantic Track
Past
GlobalInformation
Grid
PastPresent
Future
PresentFuture Present
Past
Present
Future
Past
Future
COIs
Shared World Models
Shared World ModelsCommon Track SemanticsState-full Network
COI = Condition of Interest
Overall Vision: Model-based Communication Networks, VIRT
and Rich Semantic Track
Past
PastPresent
Future
PresentFuture Present
Past
Present
Future
Past
Future
Shared World ModelsCommon Track SemanticsState-full Network
Valued Bits
Value
d Bits
GlobalInformation
Grid
Model-based Communication Network (MCN)
Instead of Stateless Networking, State-full• Maintain shared state among collaborators• State = current values of models, e.g.
• The route plan, position, velocity of an aircraft• The current and future position and behavior of a unit• The hypothesized position, status and intention of a system
A shared world model is the goal• Collectively, what the collaborators believe• Distributed, replicated for efficiency• Autonomously updated, through dead-reckoning
Like a distributed blackboard of hypotheses• Re-conceptualize Common Operational Picture• Obviate “communication” of non-news• Emphasize “information,” especially valuable information
Semantic Modeling and Condition Monitoring
• Strong semantic representations of Track data improve automated search, information filtering and reasoning
• Improved computer interpretation and processing of data provides better information products and reduces human processing load
• Semantics of military orders implies critical conditions of interest – automated derivation of COIs from orders creates valued information flows
CEC/VIRT Project
• Investigated mapping of CEC track data to RST conceptual model
• Constructed various RST representations for software and web-based implementations• Formal logic representation for automating
mappings across track data models• Semantic Web representations for web-based
data interchange and machine reasoning
• Constructed various web-based expressions of COIs
CEC/VIRT Conditions of Interest
• Significant change in expected motion of air tracks based on decision thresholds for the magnitude of change in expected position and expected velocity
• Significant change in Track ID information (e.g., from FRIEND to HOSTILE)
• Significant change in Track IFF information (e.g., change in IFF mode responses)
• Start and end of engagements• Assumption/Operational Decision: No need to
send CEC network-specific Cooperating Unit, Time, and Status messages
Air Target Tracking Scenario Results
CEP-to-Track User Messages*
Number of Messages Number of Bits
Without VIRT
With VIRT**
Without VIRT
With VIRT**
Valued Bit Ratio
Track Data 1,852 141 2,483,712
189,504 0.076
Track ID 18 6 9,216 3,072 0.333
Track IFF 121 0 42,592 0 0.000
Cooperating Unit
552 0 176,640 0 0.000
Engagement Status
0 0 0 0 N/A
Status 0 0 0 0 N/A
Time 0 0 0 0 N/A
Totals 2,543 147 2,712,160
192,576 0.071
* From RTTS XML message stream ** Dead Reckoning thresholds of 100m and 10m/sec
Valued bits made up 7% of the total bits transmitted – bit traffic can be reduced by 93% if only valued bits are sent!
CEC/VIRT Simulation Event Graph
Embedding a semantically rich knowledge base into a simulation framework for testing/experimentation with VIRT/RST concepts in a dynamic context
Embedding a semantically rich knowledge base into a simulation framework for testing/experimentation with VIRT/RST concepts in a dynamic context
Common Implementation Pattern
Contacts
Curtis [email protected]
MOVES Institute, Naval Postgraduate SchoolMonterey California 93943-5000 USA
1.831.656.3215 voice1.831.656.7599 fax