ApplicationService Issues

1
Application/Service Issues

• Sean Bechhofer

2
Ontologies in Applications

• If we want to make use of ontologies within
  applications, there are a number of things that
  we need to consider.
• How are we to make the information in the
  ontology available to the application?
• There are no right answers to these questions.

3
Ontologies in Applications

• In the past, ontologies have been embedded in the
  application
• This introduces problems with
• Maintenance
• Implementation
• Exchange of terms
• Considering the ontology as a separate resource
  provides greater flexibility

4
What is an Ontology?

• Its a collection of statements about concepts
• Definitions, relationships etc.
• In the Semantic Web, ontologies are considered as
  resources in their own right
• However, they are often implicitly thought of as
  some kind of static resource.
• Ontologies are OWL/RDF files.

5
What is an Ontology?

• Recall that one of the key aspects of OWL is the
  provision of semantics that ensure that everyone
  interprets descriptions in a consistent manner.
• Considering ontologies as static resources
  requires applications to be able to provide
  implementations of the language semantics
  (potentially non-trivial in the case of OWL).
• Instead we can think of them as dynamic objects
  providing services.
• Pushes the onus for implementation on to the
  service provider.

6
Grid Aside Ontologies as Service
7
Ontology Services

• Terminological
• Classification
• Validity and consistency checking
• Intentional Concept Descriptions
• Query
• Linguistic
• Mappings to/from terms
• Extrinsic/Annotation
• Application specific information
• Annotations that are not intensional or part of
  the intrinsic definition of the concepts.

8
Implementations of OWL

• What does it mean to provide implementations of
  ontologies?
• What do you want to do with the ontology?
• Read it
• Manipulate it
• Reason about it
• Where does the reasoner go?
• Inside or outside?

9
Ontology interactions

• Modelling
• Provide data structures that represent OWL
  ontologies/documents
• Parsing
• Taking some syntactic presentation, e.g. OWL-RDF
  and converting it to some useful internal data
  structure
• Serialization
• Producing a syntactic presentation, e.g. OWL-XML
  from a local data structure
• Manipulation
• Being able to manipulate the underlying objects
• Inference
• Providing a representation that
  implements/understands the formal semantics of
  the language

10
Implementation Aspects
Parsing
Manipulation
Modelling
Serializing
Inference
11
Files/Libraries/Services

• Access to Ontologies via URIs
• Basic representations e.g. RDF/XML
• Code Libraries/APIs
• Code libraries that expose the ontology and
  associated functionality via some API.
• Services (Web or otherwise).
• Shared resources.

12
Layering

• OWL can be layered on RDF.
• This layering provides us with a number of
  different options in terms of accessing the
  ontology.
• Direct access to the RDF triple structures
• E.g. Jena, Sesame, 3store
• May require client applications to understand
  or implement aspects of the language, e.g.
  inference.
• Access at a higher level using some API.
• OWL-API, Protégé API, Jena Ontology API

13
Modularisation

• Ontologies are likely to be distributed and
  modular
• Modularisation is poorly supported in the OWL
  specification
• owlimports
• Simply grabs the RDF graph and includes it.
• Ontology naming is problematic
• Theres no consistent mechanism for referring to
  context or collections of axioms.

14
Inside or Outside?
Application
Reasoner
Reasoner sits outside the Ontology Service.
Requests are answered w.r.t. the basic facts
asserted in the ontology.
Service API
Ontology Service
Ontology
Ontology
Ontology
15
Inside or Outside?
Application
Reasoner sits inside the Ontology Service.
Requests are answered w.r.t. the semantics.
Service API
Ontology Service
Reasoner
Ontology
Ontology
Ontology
16
Do You Really Need Reasoning?
Ontology Service answers requests w.r.t. the
basic facts asserted in the ontology.
Application
Reasoner used during design of the ontology, to
ensure a consistent and coherent hierarchy.
Service API
Ontology Service
Ontology
Reasoner
Ontology
Ontology
Taxonomy
17
COHSE

• We saw COHSE demonstrated yesterday.
• Its a system that uses an ontology to drive
  dynamic linking of documents
• An Ontology Service provides access to the
  ontologies.
• Ontologies are represented in OWL
• Third party ontologies (and documents) can be
  used

18
COHSE
Ontologies Lexicons
Semantics and world knowledge
Unlinked Document in
Ontology Service
Reasoner
COHSE Agent
Annotation Descriptions
Resource Service
Linked Document out
Resource Discovery Metadata Management
Search Engines
19
COHSE

• Ontologies accessed via URLs
• Third party ontologies
• Loaded into an Ontology Service
• Handled using OWL-API library code.
• This then exposes the ontology through a number
  of API calls
• Web Service like.
• May involve the use of a reasoner to, e.g.
  calculate the taxonomy.
• COHSE doesnt currently use much reasoning at
  deployment time, other than traversing the
  hierarchy.
• Ontology service reused in other applications