Upload
ilene-carroll
View
213
Download
1
Embed Size (px)
Citation preview
ISO/TC37/SC4/TDG6ISO/TC37/SC4/TDG6Language Resource Language Resource
OntologiesOntologies2008-05-25, Marrakech2008-05-25, Marrakech
HASIDA KoitiHASIDA Koiti
[email protected]@aist.go.jp
CfSR, AIST, JapanCfSR, AIST, Japan
TDG6 IssuesTDG6 Issuesontologizationontologization
DC, LAF, LMF, FS, MAF, SemAF, SynAF, DC, LAF, LMF, FS, MAF, SemAF, SynAF, TDG3, etc.TDG3, etc.
Cf. the Pisa group’s work on LMFCf. the Pisa group’s work on LMFextension of RDF (and ontology extension of RDF (and ontology
framework) to more straightforwardly framework) to more straightforwardly address linguistic informationaddress linguistic informationextended RDF instead of XMLextended RDF instead of XML
nodes embedding nodes … rdf:Container?nodes embedding nodes … rdf:Container?
publish TRspublish TRslaunch ISslaunch ISs
2
OntologizationOntologizationontology-based reformulationontology-based reformulation
Most current standards are based on Most current standards are based on XML and lack standard framework for XML and lack standard framework for semantic interpretation.semantic interpretation.
not XML but RDF as base not XML but RDF as base description and modeling tooldescription and modeling toolSemantic interpretation is Semantic interpretation is
standardized not for XML but for RDF.standardized not for XML but for RDF.ontology as schemaontology as schema
not DTD, XML Schema, RELAXNG, not DTD, XML Schema, RELAXNG, etc.etc. 3
Motivations of Motivations of OntologizationOntologization
Lack of formal tool by which to Lack of formal tool by which to write schemas fully addressing the write schemas fully addressing the specifications in ISs.specifications in ISs.
DCR model lacks descriptive DCR model lacks descriptive power.power.
4
Weaknesses of DCR Weaknesses of DCR MetamodelMetamodel
DCR metamodel cannot addressDCR metamodel cannot addresssorts of DCs: such as unary predicate, sorts of DCs: such as unary predicate,
binary relation, symmetric binary binary relation, symmetric binary relation, etc.relation, etc.
types of the domain (1types of the domain (1stst arg.) and the arg.) and the range (2range (2ndnd arg.) of binary relations arg.) of binary relations (properties)(properties)
5
Semantic Mess of XMLSemantic Mess of XML
Semantic interpretation of XML is not Semantic interpretation of XML is not standardized but rather arbitrary.standardized but rather arbitrary.
Many inconsistent `standards’ on Many inconsistent `standards’ on overlapping issues.overlapping issues.
Huge standards containing many Huge standards containing many different semantic interpretation different semantic interpretation manners.manners.e.g., MPEG-7 > 2000 pagese.g., MPEG-7 > 2000 pages
6
RDFRDFResource Description FrameworkResource Description FrameworkW3C recommendation W3C recommendation
http://www.w3.org/RDF/basis of ontology standards such as basis of ontology standards such as
RDFS, OWL, and SKOS.RDFS, OWL, and SKOS.graph data modelgraph data modeltextual representationtextual representation
XMLXMLN3N3
7
RDF GraphRDF Graph
http://www.example.org/people#fredhttp://www.example.org/people#fredhttp://www.example.org/people#fredhttp://www.example.org/people#fred
http://meetings.example.com/m1/hphttp://meetings.example.com/m1/hphttp://meetings.example.com/m1/hphttp://meetings.example.com/m1/hp
m:homePagem:homePagem:homePagem:homePage
m:attendingm:attendingm:attendingm:attendinghttp://meetings.example.com/cal#m1http://meetings.example.com/cal#m1http://meetings.example.com/cal#m1http://meetings.example.com/cal#m1
m:givenNamem:givenNamem:givenNamem:givenName FredFredFredFred
m:hasEmailm:hasEmailm:hasEmailm:hasEmailmailto:[email protected]:[email protected]:[email protected]:[email protected]
8
Cf. RDF in TextCf. RDF in Text
XML
N3
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:m="http://www.example.org/meeting_organization#" xmlns="http://www.example.org/people#" xmlns:p="http://www.example.org/personal_details#"> <rdf:Description about="http://meetings.example.com/cal#m1"> <m:homePage resource="http://meetings.example.com/m1/hp"/> </rdf:Description> <rdf:Description about="http://www.example.org/people#fred"> <m:attending resource="http://meetings.example.com/cal#m1"/> <p:GivenName>Fred</p:GivenName> <p:hasEmail resource="mailto:[email protected]"/> </rdf:Description></rdf:RDF>
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:m="http://www.example.org/meeting_organization#" xmlns="http://www.example.org/people#" xmlns:p="http://www.example.org/personal_details#"> <rdf:Description about="http://meetings.example.com/cal#m1"> <m:homePage resource="http://meetings.example.com/m1/hp"/> </rdf:Description> <rdf:Description about="http://www.example.org/people#fred"> <m:attending resource="http://meetings.example.com/cal#m1"/> <p:GivenName>Fred</p:GivenName> <p:hasEmail resource="mailto:[email protected]"/> </rdf:Description></rdf:RDF>
@prefix p: <http://www.example.org/personal_details#> .@prefix m: <http://www.example.org/meeting_organization#> .<http://meetings.example.com/cal#m1> m:homePage <http://meetings.example.com/m1/hp> .<http://www.example.org/people#fred> p:GivenName "Fred"; p:hasEmail <mailto:[email protected]>; m:attending <http://meetings.example.com/cal#m1> .
@prefix p: <http://www.example.org/personal_details#> .@prefix m: <http://www.example.org/meeting_organization#> .<http://meetings.example.com/cal#m1> m:homePage <http://meetings.example.com/m1/hp> .<http://www.example.org/people#fred> p:GivenName "Fred"; p:hasEmail <mailto:[email protected]>; m:attending <http://meetings.example.com/cal#m1> .
9Let’s
forg
et these
texts
Let’s fo
rget t
hese te
xts
and use gra
phs!
and use gra
phs!
ISO 24610: Feature ISO 24610: Feature StructureStructure
typed feature structure as in HPSG, etc.typed feature structure as in HPSG, etc.ISO 24610-1: Feature Structure ISO 24610-1: Feature Structure
RepresentationRepresentationISO 24610-2: Feature System ISO 24610-2: Feature System
DeclarationDeclarationgraph modelgraph modelAVM (attribute-value matrix)AVM (attribute-value matrix)textual encoding by XMLtextual encoding by XML
10
FS GraphFS Graph
determinerdeterminerdeterminerdeterminerPOSPOSPOSPOS
SPECIFIERSPECIFIERSPECIFIERSPECIFIER
ORTHORTHORTHORTH lalalala
HEAHEADD
HEAHEADD
AGRAGRAGRAGR
AGAGRR
AGAGRR
nounnounnounnounPOSPOSPOSPOS
ORTHORTHORTHORTH pommepommepommepomme
singularsingularsingularsingularNUMBERNUMBERNUMBERNUMBER
11
FS in AVMFS in AVM
SPECIFIER
HEAD
POS determinerORTH `la’AGR [1][NUMBER singular]
POS nounORTH `pomme’AGR [1]
12
FS in XMLFS in XML<fs> <f name="specifier"> <fs> <f name="pos"><symbol value="determiner"/></f> <f name="orth"><string>la</string></f> <f name="agr"> <var label="n1"> <fs><f name="number"><symbol value="singular"/></f></fs> </var> </f> </fs> </f> <f name="head"> <fs> <f name="pos"><symbol value="noun"/></f> <f name="orth"><string>pomme</string></f> <f name="agr"><var label="n1"/></f> </fs> </f></fs>
<fs> <f name="specifier"> <fs> <f name="pos"><symbol value="determiner"/></f> <f name="orth"><string>la</string></f> <f name="agr"> <var label="n1"> <fs><f name="number"><symbol value="singular"/></f></fs> </var> </f> </fs> </f> <f name="head"> <fs> <f name="pos"><symbol value="noun"/></f> <f name="orth"><string>pomme</string></f> <f name="agr"><var label="n1"/></f> </fs> </f></fs>
13Let’s fo
rget t
his, to
o!
Let’s fo
rget t
his, to
o!
FS in RDF Graph (= FS Graph)FS in RDF Graph (= FS Graph)
determinerdeterminerdeterminerdeterminerPOSPOSPOSPOS
SPECIFIERSPECIFIERSPECIFIERSPECIFIER
ORTHORTHORTHORTH lalalala
HEAHEADD
HEAHEADD
AGRAGRAGRAGR
AGAGRR
AGAGRR
nounnounnounnounPOSPOSPOSPOS
ORTHORTHORTHORTH pommepommepommepomme
singularsingularsingularsingularNUMBERNUMBERNUMBERNUMBER
14
Ontologies Subsume Feature Ontologies Subsume Feature SystemsSystems
Features are partial functions, Features are partial functions, whereas RDF properties are relations whereas RDF properties are relations in general (possibly partial functions).in general (possibly partial functions).
Usual feature systems have no Usual feature systems have no taxonomy of features, whereas usual taxonomy of features, whereas usual ontologies have taxonomies of ontologies have taxonomies of properties (e.g., due to properties (e.g., due to rdfs:subPropertyOf).rdfs:subPropertyOf).
16
wordwordwordword
<fsDecl type="word" baseTypes="sign"> <fsDescr>The fundamental type for individual words</fsDescr> <fDecl name="orth"> <fDescr>The orthographic representation for this word</fDescr> <vRange><string/></vRange> </fDecl></fsDecl>
<fsDecl type="word" baseTypes="sign"> <fsDescr>The fundamental type for individual words</fsDescr> <fDecl name="orth"> <fDescr>The orthographic representation for this word</fDescr> <vRange><string/></vRange> </fDecl></fsDecl>
orthorthorthorth
Feature Structure Feature Structure DeclarationDeclaration
17
signsignsignsign
rdfs:domainrdfs:domainrdfs:domainrdfs:domainstringstringstringstringrdfs:rangerdfs:rangerdfs:rangerdfs:range
rdfs:subClassOfrdfs:subClassOfrdfs:subClassOfrdfs:subClassOf
The fundamental type for individual wordsThe fundamental type for individual words
rdfs:commentrdfs:commentrdfs:commentrdfs:comment
The orthographic representation for this wordThe orthographic representation for this word
rdfs:commentrdfs:commentrdfs:commentrdfs:commentowl:FunctionalPropertyowl:FunctionalPropertyowl:FunctionalPropertyowl:FunctionalProperty
rdf:typerdf:typerdf:typerdf:type
Constraint (Conditional)Constraint (Conditional)
18
XXXX
invinvinvinv
truetruetruetrue
finfinfinfin
auxauxauxaux
vformvformvformvform
<cond> <fs> <f name="inv"> <binary value="true"/> </f> </fs> <then/> <fs> <f name="aux"> <binary value="true"/> </f> <f name="vform"> <symbol value="fin"/> </f> </fs></cond>
<cond> <fs> <f name="inv"> <binary value="true"/> </f> </fs> <then/> <fs> <f name="aux"> <binary value="true"/> </f> <f name="vform"> <symbol value="fin"/> </f> </fs></cond>
XXXX truetruetruetrue
condcondcondcond
named graph
FS Ontologization FS Ontologization (Summary)(Summary)
RDF ⊃ FSRDF ⊃ FSUse ontologies for feature-system Use ontologies for feature-system
declarations.declarations.We need RDF-based notations to We need RDF-based notations to
encode constraints.encode constraints.Defaults are outside of ontology.Defaults are outside of ontology.
19
ISO 24612: Linguistic ISO 24612: Linguistic Annotation FrameworkAnnotation Framework
20
RDF Extended for EmbeddingRDF Extended for Embedding
● ●● ●● ●● ●
NUMBERNUMBERNUMBERNUMBER
a node embedding nodes 21
rdfs:typerdfs:typerdfs:typerdfs:type NPNPNPNP
TheTheTheThe
clockclockclockclock
SINGSINGSINGSING
rdfs:typerdfs:typerdfs:typerdfs:type TOKENTOKENTOKENTOKEN
POSPOSPOSPOS
BASEBASEBASEBASETHETHETHETHE
DETDETDETDET
rdfs:typerdfs:typerdfs:typerdfs:type
POSPOSPOSPOS NNNNNNNN
BASEBASEBASEBASECLOCKCLOCKCLOCKCLOCK
possibly stand-off annotation
ProspectsProspectsRDF as basic data structureRDF as basic data structure
Graph modelGraph model is essential.is essential.Forget about textual encoding such as Forget about textual encoding such as
XMLXMLthough W3C insists on plain-test encoding.though W3C insists on plain-test encoding.
ontology to address FSDontology to address FSDstraightforward to basically declare straightforward to basically declare
features and feature structuresfeatures and feature structuresneed some inventions for constraintsneed some inventions for constraints
extension of RDFextension of RDFembeddings (of strings)embeddings (of strings)collections (sets, bags, lists)collections (sets, bags, lists)
lots more to dolots more to do 22