CO-ODE / HyOntUse: User Oriented Collaborative Ontology Development

1
CO-ODE / HyOntUse User Oriented Collaborative
Ontology Development

• Alan Rector, Hai Wang, Nick Drummond, Matthew
  HorridgeInformation Management Group / Bio
  Health Informatics ForumDepartment of Computer
  Science, University of ManchesterHolger
  Knublauch, Mark MusenStanford Medical
  Informatics, Stanford University
• rector_at_cs.man.ac.uk co-ode-admin_at_cs.man.ac.uk
• www.co-ode.orgprotege.stanford.orgwww.opengalen.
  org

2
Ontologies with EverythingFoundation stone of
the Semantic Web

• But how do we build them
• How do we let domain expert users build them
• Ontologies work best when owned and managed by
  their community
• OWL a major step towards standardisation
• But a major jump in understanding required from
  users

3
A Collaboration

• OWL DLs ( OilEd - Manchester) Frames and
  Plug Play (Protégé - Stanford)
• with
• Knowledge Elicitation from users Epistemics/AKT
• Loosely coupled collaborative development
  OpenGALEN
• Terminology communities
• Medicine
• Bioinformatics
• Libraries
• and funders
• JISC, EPSRC, US National Cancer Institute

4
A Collaboration of Users

• Protégé OilEd User Communities
• E-Science community
• Semantic Web Community
• JISC and Library community
• Industrial collaborators

An invitation Join the Forum Download the
toolsContribute your views www.co-ode.org
5
Goal Ontologies built, owned, and used by domain
experts

• User optimised interfaces for OWL
• Multiple interfaces for different users tasks
• Focus on what users find hard
• Multiple explanation and debugging aids
• Hybrid reasoning
• For what cant be done with DLs
• For what we need from Frames
• And other specialised reasoners
• Plug and Play Collborative Environments

6
Computer Science IssuesFrames DLs more
Different than they Look

• Primitive concepts - in a hierarchy
• Described but not defined
• Properties - relations between concepts
• Also in a hierarchy
• Descriptors - property-concept pairs

• qualified by some, only, at least, at
  most
• Defined concepts
• Made from primitive concepts and descriptors
• Axioms
• disjointness, further description of defined
  concepts
• A Reasoner
• to organise it for you

• Meta data
• Prototypical Knowledge
• Defaults Exceptions
• Reflective queries
• Individuals
• Hybrid reasoning

OWL / DLs
Frames
7
OWL-Frame Mapping

• Restrictions map to a new kind of class
• NOT to filled slots and facets
• A slot filled with a value is not a first class
  object in Protégé
• Therefore cannot annotate
• Lossy inverse-mapping
• Prototypical knowledge and defaults
• No OWL equivalent
• Mapping to OWL annotation properties under
  discussion
• Metadata
• Want metadata in a standard form but also use of
  reasoner - alternatives
• Express cleanly in OWL-Full but ignore for
  classification
• Express messily with annotation properties but
  stick to OWL-DL standard

8
Ontology Anchored Knowledge Resources

• To Build Knowledge Intensive applications
• Knowledge bases anchored on ontologies supporting
  information resources a challenge and a means
  to interworking
• Meta data with everything

9
HCI Issues

• Understanding new users confusions
• Workshops and teaching
• Understanding routine users task models
• Observation, meetings,
• Visualisation
• Text, graphics,
• Adaptation and modularisation
• Protégé plug and play Web services interfaces

10
Progress So Far

• The Logicians Interface Proof of underlying
  concepts

11
(No Transcript)
12
Visualisation (before classification)
13
After classification
14
Wizards for common tasks
15
Help Common Operations

• New Expression Editor
• Drag and drop from multiple ontologies coming RSN
• Transforming primitive to defined classes
• Requires moving restrictions to axioms
• Now all on the same pane
• Cut and past/ drag and drop
• Support good practice
• Primitive skeleton and normalisation supported

16
Debugging Tools Coming very soon

• Warnings for common errors
• Domain/Range constraint effects
• Domain / range constraints in OWL are Axioms
• Trivially satisfiable universal restrictions
• Unsatisfiability tracing
• Where to look when it all turns red
• Explanations for classification
• For the easy cases

17
Dublin Core and Other Annotations

• Provenance and metadata with everything
• Flexible environment for experimentation
• Mapping to OWL annotation properties
• But a much richer structure on top

18
Summary

• New tools
• Easier to use
• Cut time Pizza example from 1 hour to 10
  minutes
• Easier to understand
• More object oriented more information on basic
  pane
• Better visualisation
• Debugging
• Easier to teach
• The easy way is the right way

19
Collaboration for User Oriented Ontology
Development Environments

• Partners
• UK JISC EPSRC
• Manchester Information Management Group
• Southampton University / AKT
• Epistemics Software house specialising in
  Knowledge Acquisiton
• US National Cancer Institute
• Stanford Medical Informatics Protege
• Technologies
• Protégé OilEd
• AKT and Epistemics
• Material from OpenGALEN, GONG, GOAT, myGrid,
  Geodise, WonderWeb, OntoGrid, Semantic Mining,

20
Join UsIf you need Ontology Building

• Joint the forum
• Download the tools
• Get in touch with problems
• Propose collaborations, workshops, feasibility
  studies
• What do YOU need?

www.co-ode.orgprotégé.stanford.edu