Using Concept Maps to facilitate communal ontology building Karim Nashar University of Manchester

1
Using Concept Maps to facilitate communal
ontology buildingKarim NasharUniversity of
Manchester
2
Background

• - Ontology to guide the annotation of
  environmental transcriptomics metadata
• - Experiments are characteristically difficult
  to represent in ontology
• diversity
• e.g. gene expression changes in marine
  organisms in response to tidal cycle
• or plant-bacterial ecosystem - expression
  changes in response to biotic stressor (insect
  herbivore)
• - Describe biosource conditions species
  geographical locations habitat name
  description controlled versus uncontrolled
  conditions etc
• - Guiding annotation need expressive formalism
  OWL
• e.g. Class (Two_Sample_design
• Experiment_design
• restriction (includes exactly 2 Sample))
• - Research in communal bio-ontology building
  methods does not address how to mediate between
  the ontologist domain expert

3

• Problems
• - Direct OWL ontology building inefficient
• - OWL creates a communication barrier
• - OWL makes the novice uneasy
• - Visualisation tools only display taxonomy
• - Growth Problem (Gomez-Gauchia, 2004)

Knowledge Acquisition Bottleneck
4
Concept Maps (Novak Gowin 1984)
Conceptual knowledge constitutes one of the most
significant forms of knowledge in design.
Concepts are fundamental to design thinking,
since they operate on an ideational level
(Oxman, 2004)
Semantic Unit
5
CM in PC Pack epistemics.co.uk
6
CM in Cmap Tools cmap.ihmc.us
7
Knowledge Differentiation
Stakeholders e.g. RSBI, NuGO-WG, MGED-WG
Domain General Ontology
Domain Experts (bench biologists)
Env
Nut
Other Domain Specific
8
Knowledge Differentiation

• Stakeholder Knowledge
• Breadth
• Meta-knowledge

• Expert Knowledge
• Depth
• Domain-specific content

9
Advanced Knowledge Elicitation

• Continuous stakeholder-expert feedback
  dynamically evolves the Cmap representation of
  knowledge
• Cognitive-flexibility theory knowledge is
  revised in different contexts for advanced KA
• Concepts semantic units trigger the
  externalisation of tacit knowledge

10
Summary

• Cmaps reduce the Knowledge Acquisition Bottleneck
  
• Enhancing the efficiency of KE
• Reducing requirement for highly trained knowledge
  engineers
• Cmaps reduce ontology Growth Problem
• Profile the community
• Identify Stakeholders and Domain Experts
• Identify their roles in the KE process
• Use Cmaps to enable active participation of the
  community in KR
• Revise knowledge in different contexts use
  precursors to externalise tacit knowledge

11
Acknowledgements
Manchester Microarray Users Users from the
Institute for Food Research Susanna Sansone
(EBI) Helen Parkinson (EBI) Philippe Rocca Serra
(EBI) Norman Morrison (Manchester) Alex Garcia
(EBI) Jennifer Fostel (NIH/NIEHS) Chris Taylor
(EBI)
References Gómez-Gauchía H, Diaz-Agudo B,
González-Calero P Two-layered approach to
knowledge representation using conceptual maps
and description logics. In Proceedings of the
first international conference on concept
mapping 2004 Pamplona, Spain 2004. Oxman R
Think-maps teaching design thinking in design
education. Design Studies 2004, 2563-91. Garcia
Castro A, Sansone S-A, Rocca-Serra P, Tailor C,
Nashar K, Brass A, Stevens RD, Ragan MA The use
of concept maps during knowledge elicitation in
ontology development processes - the
nutrigenomics use-case. BMC Bioinformatics 2006,
7(267).