USAGE AND CHARACTERISTICS OF ONTOLOGY MODELS IN NETWORK ENABLED CAPABILITY OPERATIONS

1
USAGE AND CHARACTERISTICS OF ONTOLOGY MODELS IN
NETWORK ENABLED CAPABILITY OPERATIONS

• Capt. Mariusz Chmielewski
• mchmielewski_at_wat.edu.pl
• 1st Lt. Rafal Kasprzyk
• rkasprzyk_at_wat.edu.pl
• Computer Science Department,
• Cybernetics Faculty ,
• Military University of Technology

2
Presentation plan

• Modelling domain of NEC
• Introduction to ontologies
• Formal model of ontology and semantic netowork
• Triples Graph representation details
• Languages of semantic networks
• Resoning abilities
• Method for relational and object oriented model
  to ontology transformation
• Modeling Battlespace Ontology UBOM
• Overcoming ontology differences
• Ontology mediation
• Advantages of using ontologies
• Conclusions

3
Network Enabled Capabilities - Summary

• Network Enabled Capabilities - military doctrine
  or theory of war
• NEC supports fallowing ideas
• Efficient networked mechanisms - improving
  information flow and information sharing between
  all battlespace entities
• Information sharing improvement of transferred
  information quality and in result manufacturing
  battlespace situational awareness
• Shared situational awareness - enables
  collaboration between sensors, actors and command
  centers improving synchronization, and reduces
  the communication delays which in result speed
  up decision process and increase mission
  effectiveness.

4
The need of ontology

• Why should the study of ontology be important in
  case of NEC operations?
• This environment has many different systems (C4I
  Systems) migration rules, semantic mappings,
  automated migration
• Each systems employing their own data model
  (Relational or Object Oriented Model)
• Systems must interoperate
• Information interoperability is crutial

5
Semantic Models, Ontologies

• Origin of technology - Next stage of the Internet
  evolution, Semantic Web.
• Semantic description can improve the way
  information is presented.
• Semantic data representation - based on graph
  model - RDF concept of triples representing
  resources and data describing them.
• An ontology is used as a tool for describing and
  representing selected knowledge branch that is
  medicine, finances, battlefield etc.

6
Model of ontology
7

Model of semantic network

8
Languages of semantic networks

• dedicated languages
• RDF, RDF-S, SHOE, DAMLOIL, OWL (Ontology Web
  Language)
• share the same theoretical framework extended
  towards Description Logics to provide knowledge
  representation and reasoning mechanisms.
• In terms of the expressivity, these languages can
  be arranged in an order RDF, RDF-S, SHOE,
  DAMLOIL and OWL.
• OWL provides sublanguages
• OWL Lite - expressiveness is limited (restricted
  primitives list), but in this case the efficiency
  of reasoning is preferable,
• OWL DL (Description Logics) - is as expressive as
  possible on the premise of preserving
  completeness and decidability of reasoning
• OWL Full - is required for modeling domains using
  full spectrum expressiveness with no
  computational guarantees of reasoning.

9
The Concept
10
Modeling Battlespace Ontology

• (UBOM) Unified Battlespace Ontology Model.
• Development based on ontology modeling
  metodologies, contains
• Definition of researched domain and boundaries of
  modeling
• Existing ontology (domain models) overview and
  decision of application
• Definitions of elementary abstractions within the
  domain creation of important concept list
• Concept taxonomy modeling - class definition and
  their hierarchy
• Class property modeling identifying properties
  (slots) for classes (definition domain and range
  definition)
• Property restriction definition object type or
  datatype specification
• Identification of instances and their
  classification within the class taxonomy

11
Elements of UBOM

• Unified Battlespace Ontology Model (UBOM) has
  been divided into two parts
• MIP JC3 model based ontology describing wide
  range of military operations and the whole domain
  of military units and equipment Relational Model
• Decision support ontology containing concepts
  of decision, variant and the rest of decision
  process realized on the battlefield Object
  Oriented Model

12
Model transformations

• Definition of transformation rules for relational
  and object oriented models
• Revisioned other works to confront used
  transformation algorithms
• Because ontology is accepted as a formal,
  explicit specification of a shared
  conceptualization, we can naturally link
  ontologies with object models, which represent a
  system-oriented map of related objects, described
  as Abstract Data Types (ADTs).
• Implementation
• Sybase Power Designer 15 VBS extensions
  robust modeling environment extended with OWL DL
  abilities

13
ER model transformation

• Entity element ontology concept represented by
  OWLClass.
• All classes in OWL are identified by URI. Naming
  nomenclature use the URL of research unit or
  source C4I system followed by entity name.
• Relationship element OWL property represented
  by OWLObjectProperty element.
• An object property in OWL relates an individual
  to other individuals. An object property is
  defined as an instance of the predefined OWL
  class owlObjectProperty
• Relationships represent connections, links, or
  associations between two or more entities.
• Naming nomenclature for properties assumes
  concatenation of the URL of research unit or
  source C4I system followed by from entity name,
  identified relationship name and to entity
  name.
• Attribute OWLDataTypeProperty.
• An attribute represents a common characteristic
  of some entity instances.
• OWL data type properties are used to link
  individuals to data values.
• A data type property is defined as an instance of
  the predefined OWL class owlDatatypeProperty.
• Attribute naming assumes to concatenate URL of
  research unit or source C4I system, parent entity
  name and attribute name.
• Domain OWLDataRange element.
• A data range in OWL can be either a literal type
  or an enumeration of literals.
• Atomic domains in relational models restrict, the
  value space of the datatype identified using the
  baseType attribute.
• Domain naming assumes to use URL of research unit
  or source C4I system, and identified domain name.

14
OOM Transformation

• Object oriented model enables to extract more
  transition rules due to the object-oriented
  methodology expressiveness.
• In case of relationships OOM can provide several
  enhancements
• specialisation-generalisation relation
  (inheritance), association, aggregation,
  composition,dependency
• Extensions of ER model transformation
• extended representation of concept taxonomy based
  on the inheritance relationship
• representation of concepts aggregation and
  composition
• additional mappings for representing relationship
  restrictions
• OWLRestriction,
• HasValueRestriction,
• AllValuesFromRestriction,
• SomeValuesFromRestriction,
• CardinalityRestriction,
• MaxCardinalityRestriction,
• MinCardinalityRestriction,

15
Overcoming ontology differences

• Two kinds of ontology mediation
• ontology mapping
• Correspondences are stored separately from the
  ontologies
• Mappings are not part of the ontologies
  themselves.
• Correspondences can be used for, querying
  heterogeneous knowledge bases using a common
  interface or transforming data between different
  representations.
• ontology merging
• Produces a new ontology which is the union of the
  source ontologies
• Merged ontology encapsulates all the knowledge
  from the original sources
• Described process must ensure that all
  correspondences and differences between the
  ontologies are reflected in the merged ontology.

16
Overcoming ontology differences

• Stages of semantic mapping
• Importing the content of the ontologies to chosen
  ontology language
• Normalization of defined vocabularies through
  elimination of lexical and syntactical
  differences
• Similarity evaluation of ontology entities using
  defined set of parameters
• Ontology quantity analysis - predefined sets of
  indicators
• Establishing correspondences between similar
  entities (concept thesaurus), in the form of
  semantic bridges linking similar concepts
• Utilizing developed mappings for instance
  transformation
• Revision of prepared mappings for improvements

17
Ontology modeling and processing

• Protege JENA Semantic Framework
• process semantic models stored in RDF, RDFS, DAML
  and OWL
• Automated modeling
• supporting SPARQL query language,
• SWRL implementation - rule language for semantic
  data
• Inference mechanisms
• Graph persistence
• Graph visualisations

18
Modeling environment
19
Main advantages of using ontologies

• Considering the availability of tools for
  designing, developing ontologies and inference
  engines they offer greater support for automation
  of data processing specially in
• process of data migration from heterogeneous
  sources
• integration of variety of domain models
• reusability of existing domain models
• rule based reasoning
• query processing based on semantic data
  representation (RDQL and RQL family languages)

20
Future work

• Development of transformation rules in native
  Protégé environment
• Research on improvement of Unified Battlespace
  Ontology Model
• Considering wide spectrum of ontology model
  applications authors want to develop methodology
  for creation and evolution of ontology models
  applied to military decision support tools
• Ontology based tools for decision support and
  training (simulation tools)

21
Thank You

• Authors
• Capt. Mariusz Chmielewski
• email mchmielewski_at_wat.edu.pl
• 1st Lt. Rafal Kasprzyk
• email rkasprzyk_at_wat.edu.pl
• Computer Science Department,
• Cybernetics Faculty ,
• Military University of Technology

ACKNOWLEDGEMENTS This work was partially
supported by grant Research Project No
PBZ-MNiSW-DBO-02/I/2007.