GraniteOntologies

1
GraniteOntologies
Gunnar Aastrand Grimnes, Stuart Schalmers, Pete
Edwards Alun Preece

• Acquiring and Applying Ontologies in GraniteNights

GraniteNights, Gunnar Aastrand Grimnes, CIA,
August 2003.
2
Introduction Talk Overview

• GraniteNights
• Initially made for Agentcities Competition.
• Demoed _at_ AC ID3 in Barcelona, Feb. 03.
• GraniteNights also presetned in a proper talk _at_
  CIA2003.
• Talk
• Overview
• Ontologies
• Tools
• Conclusions

3
Screenshots I
4
Screenshots II

• GraniteNights remembers users and learns their
  preferences over time.

5
Architecture Diagram
6
Messages and Data in GN

• Everything is RDF, use as little SL as possible.
• Not just data
• All messages are represented in RDF.
• Jade message envelopes are SL (but could be XML
  or RDF using the appropriate plugin).
• To support this several ontologies are needed.

7
Ontologies

• Re-use is the name of the game.
• Ontologies from Agentcities.RTD project
• Shows (cinema)
• Restaurants
• Utilities ( address, time/dates etc. )
• Home grown ontologies
• Beer Pubs ( based on Shoe example ontology )
• Evening plans User Profiles
• Query by Example

8
Tools

• Jena for handling RDF
• Supports N3, great for hand-writing RDF, convert
  to RDF/XML.
• Frodo RDFSViz
• Home-made
• Query By Example.
• Java ontology classes generator.
• Ontology driven Classfiller for generating
  instance data.
• XEmacs! The one and only text-editor!

9
Query By Example

• QBex is an RDF based query language.
• Variables constraints are supported.
• Internally converted to RDQL

ltqQuerygt ltqtemplategt ltaktAcademicgt
ltaktfamily-namegt Brown lt/aktfamily-namegt
lt/aktAcademicgt lt/qtemplategt
lt/qQuerygt
SELECT ?x WHERE (?x, ?y, ?z), ( ?x, ltrdf
typegt, ltakt Academicgt ), ( ?x, ltakt
family-namegt, "Brown" )
10
QBeX Variables

• Using an RDF based constraint interchange format,
  QBeX can represent expressions like

Query(X) type(X, Restaurant), serves(X,
Tandoori), openingTime(X,Y), Ygt1900.
11
Jena vocabulary classes

• Vocabulary Java files are generated directly from
  the XML files.
• Javadoc comments generated includes
• Labels Comments
• Sub and Super classes
• Range / Domain of Properties
• Support RDFS DAMLOIL.
• Combined with Frodo RDFSViz it makes a powerful
  tool for understanding new ontologies.

12
Ontology driven Web form for RDF generation

• Fully web-based instance creator.
• Generates forms for each possible class.
• Forms have fields for each property of a class.
• Supports sub-class sub-property inference.
• Supports DAMLOIL RDFS ontologies.
• Written in PHP, using RDF API for PHP (RAP).

13
Conclusions

• Lightweight ontologies are more than sufficient.
• The representation powers of DAMLOIL are too
  much to handle.
• Jade agents RDF/XML for message content work
  well!
• Using RDF makes interoperability and reusability
  a breeze.
• Several other projects in Aberdeen, re-using
  components of GraniteNights!

14
Wishlist for ontologies the future AC net

• Standardise on OWL-Lite for ontology
  representation.
• Ontology repository which is used...
• and which includes DOCUMENTATION )
• And btw, what about using RDF/XML for all
  messages?
• and a Jade plugin that will convert them to
  Java object(s)?

15
Questions?
16
QBeX CIF Example
ltqQuerygt ltqtemplategt ltrRestaurantgt
ltrserves rdfresourcerTandoori" /gt
ltropen-timegt ltcifVariable rdfID"x"gt
ltcifvarnamegtxlt/cifvarnamegt
lt/cifVariablegt lt/ropen-timegt
lt/rRestaurantgt lt/qtemplategt
ltqconstraintsgt ltcifComparisongt
ltcifcomparisonOperatorgtgtlt/cifcomparisonOperat
orgt ltcifcomparisonOp1gt
ltcifVariable rdfabout"x"/gt
lt/cifcomparisonOp1gt ltcifcomparisonOp2gt
ltcifIntegerconstgt
ltcifconstantValuegt1900lt/cifconstantValuegt .. .
.
17
The single most important thing to remember from
this talk

• VOTE FOR US IN THE COMPETITION!

Dr Pete Edwards
Dr Alun Preece
Gunnar Grimnes
Stuart Chalmers