29
Representing Translations on the Semantic Web Elena Montiel-Ponsoda, Jorge Gracia, Guadalupe Aguado-de-Cea, Asunción Gómez-Pérez Ontology Engineering Group (OEG) Facultad de Informática Universidad Politécnica de Madrid http://www.oeg-upm.net { ti l j i l }@fi {emontiel, jgracia, lupe, asun}@fi.upm.es

Representing Translations on the Semantic Web

Embed Size (px)

DESCRIPTION

Representing Translations on the Semantic Web. Elena Montiel-Ponsoda. ISWC2011

Citation preview

Page 1: Representing Translations on the Semantic Web

Representing Translations on the Semantic Web

Elena Montiel-Ponsoda, Jorge Gracia, Guadalupe Aguado-de-Cea, Asunción Gómez-Pérez

Ontology Engineering Group (OEG)Facultad de Informática

Universidad Politécnica de Madrid

http://www.oeg-upm.net

{ ti l j i l }@fi{emontiel, jgracia, lupe, asun}@fi.upm.es

Page 2: Representing Translations on the Semantic Web

The (Multilingual) Web of Data

• We know that the Web is multilingual….

• Is the Web of Data also multilingual?• Is the Web of Data also multilingual?• Ell, B., Vrandecic, D., and Simperl, E. (2011). Labels in the Web

of Data

1 language specified: 2.2%N l ifi d 0 7%

English: 44.72%German: 5.22 %F h 11%

Most usedlanguage tags:

N languages specified: 0.7%

2

French: 5.11%

Page 3: Representing Translations on the Semantic Web

The (Multilingual) Web of Data

data.bnf.fr – Bibliothèque national de FranceGeoLinkedData.es – Spanish geospatial datap g pRechtspraak.nl – Netherlands Council of the JudiciaryFAO geopolitical ontology – with labels in en, fr, es, ar, zh, ru, it

3

AGROVOC Linked Open Data – AGROVOC agricultural thesaurus

Page 4: Representing Translations on the Semantic Web

The problem

4

Page 5: Representing Translations on the Semantic Web

Our proposal

• To propose a representation mechanism for explicitp p p p

translation relations between natural language

descriptions associated to ontology elements and data.

• To implement it as a metamodel in OWL offered as a

module of the lemon model, lexicon-ontology model to

account for the linguistic descriptions associated tog p

ontologies and linked data.

5

Page 6: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

6

Page 7: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

7

Page 8: Representing Translations on the Semantic Web

RDFS, SKOS

RDF(S), OWLifrs:FinancialAssets rdfs:label “financial assets”@en

RDF(S), OWL

rdfs:SubPropertyOf

SKOSifrs:FinancialAssets “financial assets”@en

skos:prefLabel

SKOS labels:  prefLabel, altLabel & hiddenLabel. 

8

p ,

Page 9: Representing Translations on the Semantic Web

SKOS

SKOS enables a simple form of multilingual labeling:

ifrs:FinancialAssets “financial assets”@enskos:prefLabelp

“activos financieros”@esskos:prefLabel

What happens when we have more than one label perWhat happens when we have more than one label per language? Food and Agriculture Organization and FAO?

How can we create explicit links between labels?How can we create explicit links between labels?

Say that one is translation, acronym of the other?

9

Page 10: Representing Translations on the Semantic Web

SKOS-XL

skosxl:LabelSKOS XLclass

skosxl:LabelSKOS‐XL

rdf:type

ifrs:FinancialAssets ifrs:FinancialAssetsLabelskosxl:prefLabel

skosxl:literalForm

“financial assets”@en

10

Page 11: Representing Translations on the Semantic Web

SKOS-XL

rdf:type

skosxl:LabelSKOS‐XLifrs:FinancialAssets

skosxl:literalForm

ifrs:FinancialAssetsLabel1skosxl:prefLabel

“financial assets”@enskosxl:labelRelation

ex:isTranslationOfrdfs:subPropertyOf

ex:isTranslationOf

“activos financieros”@es

skosxl:literalForm

ifrs:FinancialAssetsLabel2

rdf:typeskosxl:prefLabel

11

skosxl:Label

Page 12: Representing Translations on the Semantic Web

LIR

RiverENFR

12

Page 13: Representing Translations on the Semantic Web

Limitations

Th l ti k! These solutions work! ……but with some limitations

Rigid models Simple translation relation insufficient for:p original vs. target label type of translation relation source of the translation adequacy or reliability of translations

13

Page 14: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

14

Page 15: Representing Translations on the Semantic Web

The lemon model

An RDF‐based ontology‐lexicon model for ontologiesAn RDF‐based ontology‐lexicon model for ontologies

Main features:

• Semantics by reference

• Rich lexical and terminological description of ontology elements

• Concise (i.e., trade off between complexity and expressivity)expressivity)

• Descriptive not prescriptive (i.e., uses data categories)

15

• Modular and extensible

Page 16: Representing Translations on the Semantic Web

The lemon model

But this is also quite complex, isn’t it?Not so much… remember its

modular naturemodular nature

16

Page 17: Representing Translations on the Semantic Web

The lemon model

17

Page 18: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

18

Page 19: Representing Translations on the Semantic Web

Typology of translation relations

Ontology Localization

Multilingual Ontology(an ontololgy in which labels are documented in multiple NLs)

b tbut…

Does a 1 to 1 correspondence between always exist?

19

Does a 1 to 1 correspondence between always exist?

Page 20: Representing Translations on the Semantic Web

Typology of translation relations

Types of domains

CInternationalizedor standardized

domains

Culturallyinfluenceddomains

Types of conceptualizations

Conceptualizationsshared among the

Conceptualizations thatrepresent mismatchesshared among the

languages representedin the ontology

represent mismatchesbetween cultures and

languages

20

Page 21: Representing Translations on the Semantic Web

Literal vs. Cultural equivalence Translation

Ontology A (German)

Ontology B (English)

Concept A Concept Bp p

Sparkasse German savings institution Savings bank

Literal translation Cultural equivalence

21

translation

Page 22: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

22

Page 23: Representing Translations on the Semantic Web

lemon module for translationsLexicon

language:String

LexicalEntry LexicalSense Ontology termreference

entryisSenseOf

sourceLexicalSense targetLexicalSenselexicalForm

translationOriginForm TranslationResource

representation:String confidenceLevel:double

CulturalEquivalenceTranslationLiteralTranslation q

23

Page 24: Representing Translations on the Semantic Web

Example of literal translation

LEXICONEN

LexicalEntry LexicalSenseONTOLOGY“payment method”

http://purl.org/goodrelations/v1#PaymentMethods

Translation

p y

Translation

LexicalEntry LexicalSense

“medio de pago”

LEXICON

medio de pago

LEXICONES

Page 25: Representing Translations on the Semantic Web

Example of literal translation

LEXICONEN

LexicalEntry LexicalSenseONTOLOGY“Cabinet of Spain”

http://dbpedia.org/page/Consejo_de_Ministros

LiteralTranslation

p

LiteralTranslation

LexicalEntry LexicalSense

“Consejo de Ministros”

LEXICON

Consejo de Ministros

LEXICONES

Page 26: Representing Translations on the Semantic Web

Example of cultural equivalence translation

LEXICONEN

LexicalEntry LexicalSensehttp://www.oegov.us/democracy/us/core/owl/usgov#CABINET“Cabinet”

ONTOLOGY

LexicalEntry

CulturalEquivalenceTranslationCulturalEquivalenceTranslation

LexicalEntry LexicalSense

“Consejo de Ministros”

http://dbpedia.org/page/Cabinet_of_Spain

LEXICON ONTOLOGY

Consejo de Ministros

LEXICONES

Page 27: Representing Translations on the Semantic Web

Outline

1. Current mechanisms for translation relations

2. lemon

3. Typology of translation relations

4 Proposed lemon module for translations4. Proposed lemon module for translations

Examples of use

5. Conclusions

27

Page 28: Representing Translations on the Semantic Web

Conclusions

Benefits of the approach: Direct explicit translations can be represented Direct, explicit translations can be represented Distinction between literal/culturally equivalent translation Translation metadata can be accounted for Moderate complexity Expressivity of lemon model Conceptual/lexical layers remain separate

Future work: Test this with more real examplesTest this with more real examples Algorithms to distinguish literal/culturally equivalent

translations

28

Page 29: Representing Translations on the Semantic Web

Thanks for your attention!

29