03.11.2012.MeetingInRome

Agenda

Data: 03.11.2012, Meeting Venue: Rome, Italy.

Participants

• Claudio Baldassarre (FIPS)
• Julien Barde (IRD)
• Chryssoula Bekiari (FORTH)
• Pascal Cauquil
• Martin Doerr (FORTH)

Discussion Summary

The TLO is a core ontology of relationships (rdf properties) for schema integration and not a terminological system of general terms for querying. The TLO can be used to formulate global sparql queries in terms of TLO properties and the classes necessary to constrain the applicability of those properties. The terms will remain that of the individual knowledge bases accessed by that query. By “terms” we mean the names of species, ecosystem types, industrial activity types etc. Queries may take into account equivalence relationships between terms in the different knowledge bases, but there will not be a super terminology (ontology) of common terms.

Actions

• Presentation of proposed TLO (attached slides).
• Discussion about use cases. The following topics were mentioned:
  • Expanding search terms for documents by semantic relationships taken from different knowledge bases. This scenario was not supported by a specific example in the discussion.
  • Services for common access (replacing physical import of parts of one knowledge base into another, such as Worms into Ecoscope)
  • Data about species and occurrence points and occurrence date (is the date of the observation took place)
  • Define relationships between different Taxon lists:
    • Correspondence
    • Overlaps
    • Include
    • Not found
  • Fact sheet generator on species, collecting propositions from multiple knowledge bases:
    • FLOD and Ecoscope don’t describe the species.
    • They want to demonstrate a few data sets in the evaluation meeting and show Open data sets that the reviewers will have access.
    • Julien said that they have a group working on open data set of observations. They will make the raw observation data available.
    • We agreed to find data records that make sense and to find queries intersecting the sections.
    • An example is “give publications for Aquaculture published by FAO that have annotations”
    • There is a co reference management problem, and there are part whole relationships between identifiers.
    • They are interesting in exploiting relationships in other data source
    • It is needed to query for the language
    • The information about “preys” are given by Ecoscope
    • We will use Ecoscope URI as a reference
    • FAO has the authority for Latin Name, scientific name.
    • WoRMS has the authority for species taxonomy
    • GbiF has the authority of biodiversity data
    • Taxon ranks are important.
    • FLOD use pictures annotated by tag for resolving species ambiguity
    • Ecoscope uses depiction and the metadata of the picture for taxonomic purposes.
    • Fishing gears exploiting species
    • A user can make
      • a selection of properties
      • to find the originator of reference
      • to distinguish the provenance
      • to find the species name and genus
      • to find the source of the information

TLO demonstration

We agreed to modify the presented draft TLO for the needs of the fact sheet generator. Through the TLO the FAO, Ecoscope and WoRMS, SPARQL endpoints will be accessed for providing all the information needed for the fact sheet generator. Intellectually, the draft TLO by FORTH is regarded a reasonable approach. In case the complexity of mapping to current TLO is too high, TLO will be adjusted closer to the current knowledge bases to allow for an early demonstrator.

Next actions

FORTH will find the classes, metaclasses and properties of the TLO necessary to define the mappings from FLOD and Ecoscope needed for the fact sheet use case and will describe these concepts in RDFS or OWL. If possible, mappings will be described in OWL.