1 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Ontologies

2 OntologiesOutline WordNet (Miller et al. 90, Fellbaum 98) WordNet (Miller et al. 90, Fellbaum 98) EuroWordNet (Vossen et al. 98) EuroWordNet (Vossen et al. 98) Spanish WordNet Spanish WordNet Combining Methods (Atserias et al. 97) Combining Methods (Atserias et al. 97) Mapping hierarchies (Daudé et al. 01) Mapping hierarchies (Daudé et al. 01) Mikrokosmos (Viegas et al. 96) Mikrokosmos (Viegas et al. 96) Cyc (Malesh et al. 96) Cyc (Malesh et al. 96) WordNet 2 (Harabagiu 98) WordNet 2 (Harabagiu 98) MindNet (Richardson et al. 97) MindNet (Richardson et al. 97) ThoughtTreasure (Mueller 00) ThoughtTreasure (Mueller 00) Meaning... Meaning...

3 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet & EuroWordNet

4 WordNet & EuroWordNet WordNet Universidad de Princeton (Miller et al. 1990) Universidad de Princeton (Miller et al. 1990) Conceptos lexicalizados (parabras, lexíes) Conceptos lexicalizados (parabras, lexíes) Relacionados entre sí por relaciones semánticas Relacionados entre sí por relaciones semánticas sinonimia sinonimia antonimia antonimia hiperonimia-hiponimia hiperonimia-hiponimia meronimia meronimia implicación implicación causa causa......

5 WordNet & EuroWordNet Relaciones Semánticas de WN1.5 SinonimiaSinonimia Conceptos Lexicalizados (SYNSETS)Conceptos Lexicalizados (SYNSETS) Noción débil de sinonimia: Sinonimia en contextoNoción débil de sinonimia: Sinonimia en contexto Synset: Conjunto de palabras o lexías que en un contexto dado expresan un conceptoSynset: Conjunto de palabras o lexías que en un contexto dado expresan un concepto Hiperonimia / HiponimiaHiperonimia / Hiponimia Relación de clase a subclaseRelación de clase a subclase

6 MeronimiasMeronimias Parte componenteParte componente {mano}  {brazo} Elemento de colectividadElemento de colectividad {persona}  {gente} SustanciaSustancia {periódico}  {papel} WordNet & EuroWordNet Relacions Semàntiques de WN1.5

7 Antonimia Antonimia {grande}  {pequeño} Causa Causa {matar}  {morir} Implicación Implicación {divorciarse}  {casarse} Derivación Derivación {presidencial}  {presidente} Similitud Similitud {bueno}  {positivo} WordNet & EuroWordNet Relaciones Semánticas de WN1.5

8 WordNet & EuroWordNet Ejemplo WordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_8.jpg", "name": "WordNet & EuroWordNet Ejemplo WordNet ", "description": "WordNet & EuroWordNet Ejemplo WordNet ", "width": "800" } 9 Proyecto LE-2 4003 Proyecto LE-2 4003 Telematics Application Programme de la UETelematics Application Programme de la UE Redes semánticas de diversas lenguas Redes semánticas de diversas lenguas Integradas e interconectadas Integradas e interconectadas InglésUniversidad de SheffieldInglésUniversidad de Sheffield HolandésUniv. de AmsterdamHolandésUniv. de Amsterdam ItalianoI.L.C. de PisaItalianoI.L.C. de Pisa EspañolUB, UPC, UNED.EspañolUB, UPC, UNED. Computers and the Humanities Computers and the Humanities (Vol.monográfico,1998) (Vol.monográfico,1998) http://www.hum.uva.nl/~ewn/ http://www.hum.uva.nl/~ewn/ WordNet & EuroWordNet EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_9.jpg", "name": "Proyecto LE-2 4003 Proyecto LE-2 4003 Telematics Application Programme de la UETelematics Application Programme de la UE Redes semánticas de diversas lenguas Redes semánticas de diversas lenguas Integradas e interconectadas Integradas e interconectadas InglésUniversidad de SheffieldInglésUniversidad de Sheffield HolandésUniv.", "description": "de AmsterdamHolandésUniv. de Amsterdam ItalianoI.L.C. de PisaItalianoI.L.C. de Pisa EspañolUB, UPC, UNED.EspañolUB, UPC, UNED. Computers and the Humanities Computers and the Humanities (Vol.monográfico,1998) (Vol.monográfico,1998) http://www.hum.uva.nl/~ewn/ http://www.hum.uva.nl/~ewn/ WordNet & EuroWordNet EuroWordNet.", "width": "800" } 10 EWN2EWN2 Alemán, Francés, Checo, Sueco, Estonio Proyecto ITEMProyecto ITEM Castellano, Catalán, Vasco CREL (Centre de Referència d’Enginyeria Lingüística)CREL (Centre de Referència d’Enginyeria Lingüística) Catalán (UB, UPC) WordNet & EuroWordNet Extensiones EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_10.jpg", "name": "EWN2EWN2 Alemán, Francés, Checo, Sueco, Estonio Proyecto ITEMProyecto ITEM Castellano, Catalán, Vasco CREL (Centre de Referència d’Enginyeria Lingüística)CREL (Centre de Referència d’Enginyeria Lingüística) Catalán (UB, UPC) WordNet & EuroWordNet Extensiones EuroWordNet", "description": "EWN2EWN2 Alemán, Francés, Checo, Sueco, Estonio Proyecto ITEMProyecto ITEM Castellano, Catalán, Vasco CREL (Centre de Referència d’Enginyeria Lingüística)CREL (Centre de Referència d’Enginyeria Lingüística) Catalán (UB, UPC) WordNet & EuroWordNet Extensiones EuroWordNet", "width": "800" } 11 Desarrollo de recursos BásicosDesarrollo de recursos Básicos Tratamiento interlingüístico de la informaciónTratamiento interlingüístico de la información - Sistemas multilingües de recuperación de información (p.e., Internet) - Módulo léxico-semántico de los sistemas de ingeniería lingüística  Extracción de información  Traducción automática WordNet & EuroWordNet Aplicaciones { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_11.jpg", "name": "Desarrollo de recursos BásicosDesarrollo de recursos Básicos Tratamiento interlingüístico de la informaciónTratamiento interlingüístico de la información - Sistemas multilingües de recuperación de información (p.e., Internet) - Módulo léxico-semántico de los sistemas de ingeniería lingüística  Extracción de información  Traducción automática WordNet & EuroWordNet Aplicaciones", "description": "Desarrollo de recursos BásicosDesarrollo de recursos Básicos Tratamiento interlingüístico de la informaciónTratamiento interlingüístico de la información - Sistemas multilingües de recuperación de información (p.e., Internet) - Módulo léxico-semántico de los sistemas de ingeniería lingüística  Extracción de información  Traducción automática WordNet & EuroWordNet Aplicaciones", "width": "800" } 12 Preservación de las relaciones semánticas específicas de cada lenguaPreservación de las relaciones semánticas específicas de cada lengua Máxima compatibilidad entre los diferentes recursosMáxima compatibilidad entre los diferentes recursos Relativa independencia de los WordNetsRelativa independencia de los WordNets en el proceso de construcciónen el proceso de construcción en el resultado finalen el resultado final WordNet & EuroWordNet Requisitos de Diseño { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_12.jpg", "name": "Preservación de las relaciones semánticas específicas de cada lenguaPreservación de las relaciones semánticas específicas de cada lengua Máxima compatibilidad entre los diferentes recursosMáxima compatibilidad entre los diferentes recursos Relativa independencia de los WordNetsRelativa independencia de los WordNets en el proceso de construcciónen el proceso de construcción en el resultado finalen el resultado final WordNet & EuroWordNet Requisitos de Diseño", "description": "Preservación de las relaciones semánticas específicas de cada lenguaPreservación de las relaciones semánticas específicas de cada lengua Máxima compatibilidad entre los diferentes recursosMáxima compatibilidad entre los diferentes recursos Relativa independencia de los WordNetsRelativa independencia de los WordNets en el proceso de construcciónen el proceso de construcción en el resultado finalen el resultado final WordNet & EuroWordNet Requisitos de Diseño", "width": "800" } 13 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_13.jpg", "name": "", "description": "", "width": "800" } 14 NúcleoNúcleo El ILIEl ILI La Top Concept Ontology (TCO)La Top Concept Ontology (TCO) Ontología de dominios (DO)Ontología de dominios (DO) PeriferiaPeriferia WordNets específicosWordNets específicos WordNet & EuroWordNet Componentes de EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_14.jpg", "name": "NúcleoNúcleo El ILIEl ILI La Top Concept Ontology (TCO)La Top Concept Ontology (TCO) Ontología de dominios (DO)Ontología de dominios (DO) PeriferiaPeriferia WordNets específicosWordNets específicos WordNet & EuroWordNet Componentes de EuroWordNet", "description": "NúcleoNúcleo El ILIEl ILI La Top Concept Ontology (TCO)La Top Concept Ontology (TCO) Ontología de dominios (DO)Ontología de dominios (DO) PeriferiaPeriferia WordNets específicosWordNets específicos WordNet & EuroWordNet Componentes de EuroWordNet", "width": "800" } 15 Colección no estructurada de elementosColección no estructurada de elementos Ligados conLigados con al menos, un synset de un EWNal menos, un synset de un EWN un elemento de la TCO o DOun elemento de la TCO o DO Asociados a synsets de WN 1.5Asociados a synsets de WN 1.5 WordNet & EuroWordNet Interlingual Index of EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_15.jpg", "name": "Colección no estructurada de elementosColección no estructurada de elementos Ligados conLigados con al menos, un synset de un EWNal menos, un synset de un EWN un elemento de la TCO o DOun elemento de la TCO o DO Asociados a synsets de WN 1.5Asociados a synsets de WN 1.5 WordNet & EuroWordNet Interlingual Index of EuroWordNet", "description": "Colección no estructurada de elementosColección no estructurada de elementos Ligados conLigados con al menos, un synset de un EWNal menos, un synset de un EWN un elemento de la TCO o DOun elemento de la TCO o DO Asociados a synsets de WN 1.5Asociados a synsets de WN 1.5 WordNet & EuroWordNet Interlingual Index of EuroWordNet", "width": "800" } 16 Jerarquía de conceptos independientes de la lenguaJerarquía de conceptos independientes de la lengua distinciones semánticas: objeto, lugar, dinámico, …distinciones semánticas: objeto, lugar, dinámico, … abstracta (no léxica)abstracta (no léxica) Superpuesta al ILISuperpuesta al ILI Tres tipos de entidades:Tres tipos de entidades: Primer orden: entidades concretasPrimer orden: entidades concretas Segundo orden: situaciones estáticas o dinámicasSegundo orden: situaciones estáticas o dinámicas Tercer orden: proposiciones abstractasTercer orden: proposiciones abstractas WordNet & EuroWordNet Top Concept Ontology of EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_16.jpg", "name": "Jerarquía de conceptos independientes de la lenguaJerarquía de conceptos independientes de la lengua distinciones semánticas: objeto, lugar, dinámico, …distinciones semánticas: objeto, lugar, dinámico, … abstracta (no léxica)abstracta (no léxica) Superpuesta al ILISuperpuesta al ILI Tres tipos de entidades:Tres tipos de entidades: Primer orden: entidades concretasPrimer orden: entidades concretas Segundo orden: situaciones estáticas o dinámicasSegundo orden: situaciones estáticas o dinámicas Tercer orden: proposiciones abstractasTercer orden: proposiciones abstractas WordNet & EuroWordNet Top Concept Ontology of EuroWordNet", "description": "Jerarquía de conceptos independientes de la lenguaJerarquía de conceptos independientes de la lengua distinciones semánticas: objeto, lugar, dinámico, …distinciones semánticas: objeto, lugar, dinámico, … abstracta (no léxica)abstracta (no léxica) Superpuesta al ILISuperpuesta al ILI Tres tipos de entidades:Tres tipos de entidades: Primer orden: entidades concretasPrimer orden: entidades concretas Segundo orden: situaciones estáticas o dinámicasSegundo orden: situaciones estáticas o dinámicas Tercer orden: proposiciones abstractasTercer orden: proposiciones abstractas WordNet & EuroWordNet Top Concept Ontology of EuroWordNet", "width": "800" } 17 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_17.jpg", "name": "", "description": "", "width": "800" } 18 Jerarquía de etiquetas de dominioJerarquía de etiquetas de dominio Reducción de la polisemiaReducción de la polisemia Dominios:Dominios: Tráfico:Tráfico: Tráfico rodado, tráfico aéreoTráfico rodado, tráfico aéreo Información InternacionalInformación Internacional MicologíaMicología MedicinaMedicina WordNet & EuroWordNet Domain Ontology of EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_18.jpg", "name": "Jerarquía de etiquetas de dominioJerarquía de etiquetas de dominio Reducción de la polisemiaReducción de la polisemia Dominios:Dominios: Tráfico:Tráfico: Tráfico rodado, tráfico aéreoTráfico rodado, tráfico aéreo Información InternacionalInformación Internacional MicologíaMicología MedicinaMedicina WordNet & EuroWordNet Domain Ontology of EuroWordNet", "description": "Jerarquía de etiquetas de dominioJerarquía de etiquetas de dominio Reducción de la polisemiaReducción de la polisemia Dominios:Dominios: Tráfico:Tráfico: Tráfico rodado, tráfico aéreoTráfico rodado, tráfico aéreo Información InternacionalInformación Internacional MicologíaMicología MedicinaMedicina WordNet & EuroWordNet Domain Ontology of EuroWordNet", "width": "800" } 19 Riqueza superior a WNRiqueza superior a WN Entre:Entre: synsets (módulos monolingües)synsets (módulos monolingües) registros ILI (multilingües):registros ILI (multilingües): {actuar-1} EQ-SYNONYM {‘behave in a certain manner’} registros ILI y TCO o ODregistros ILI y TCO o OD WordNet & EuroWordNet Relaciones de EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_19.jpg", "name": "Riqueza superior a WNRiqueza superior a WN Entre:Entre: synsets (módulos monolingües)synsets (módulos monolingües) registros ILI (multilingües):registros ILI (multilingües): {actuar-1} EQ-SYNONYM {‘behave in a certain manner’} registros ILI y TCO o ODregistros ILI y TCO o OD WordNet & EuroWordNet Relaciones de EuroWordNet", "description": "Riqueza superior a WNRiqueza superior a WN Entre:Entre: synsets (módulos monolingües)synsets (módulos monolingües) registros ILI (multilingües):registros ILI (multilingües): {actuar-1} EQ-SYNONYM {‘behave in a certain manner’} registros ILI y TCO o ODregistros ILI y TCO o OD WordNet & EuroWordNet Relaciones de EuroWordNet", "width": "800" } 20 WordNet & EuroWordNet Relaciones Interlingüísticas de EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_20.jpg", "name": "WordNet & EuroWordNet Relaciones Interlingüísticas de EuroWordNet", "description": "WordNet & EuroWordNet Relaciones Interlingüísticas de EuroWordNet", "width": "800" } 21 WordNet & EuroWordNet Relaciones de EuroWordNet { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_21.jpg", "name": "WordNet & EuroWordNet Relaciones de EuroWordNet", "description": "WordNet & EuroWordNet Relaciones de EuroWordNet", "width": "800" } 22 Spanish WordNet: Building Process German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_22.jpg", "name": "Spanish WordNet: Building Process German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya", "description": "Spanish WordNet: Building Process German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya", "width": "800" } 23 Spanish WordNet General Methodology 1) Mapping to WN1.5 n manual work n automatic derivation of equivalents, using bi- lingual dictionaries 2) Manual correction 3) Re-structuring { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_23.jpg", "name": "Spanish WordNet General Methodology 1) Mapping to WN1.5 n manual work n automatic derivation of equivalents, using bi- lingual dictionaries 2) Manual correction 3) Re-structuring", "description": "Spanish WordNet General Methodology 1) Mapping to WN1.5 n manual work n automatic derivation of equivalents, using bi- lingual dictionaries 2) Manual correction 3) Re-structuring", "width": "800" } 24 Spanish WordNet Main Steps: First Core (Manual Translation) –Nouns: n A) WN1.5’s Tops File plus first level of hyponyms (about 800 synsets). n B) The rest of EWN’s Common Base Concepts (which were not in our set). C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  – Verbs: n Manual translation of EWN’s Base Concepts (about 150 synsets) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_24.jpg", "name": "Spanish WordNet Main Steps: First Core (Manual Translation) –Nouns: n A) WN1.5’s Tops File plus first level of hyponyms (about 800 synsets).", "description": "n B) The rest of EWN’s Common Base Concepts (which were not in our set). C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  – Verbs: n Manual translation of EWN’s Base Concepts (about 150 synsets).", "width": "800" } 25 Spanish WordNet Main Steps: Subset 1 (Semi-automatic)  ouns:  ouns: –Applying authomatic methods using bi-lingual dictionaries –Manual validation of several subsets to check if the link is correct –Deriving a Confidence Score (CS) for every authomatic method (heuristic) –Selecting pairs synset-word above 85% CS –Some manual correction of this Subset 1 (mainly, filling gaps) n Verbs: –3600 English verbs connected to WN1.5 senses and ambiguously translated to Spanish are manually inspected and disambiguated { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_25.jpg", "name": "Spanish WordNet Main Steps: Subset 1 (Semi-automatic)  ouns:  ouns: –Applying authomatic methods using bi-lingual dictionaries –Manual validation of several subsets to check if the link is correct –Deriving a Confidence Score (CS) for every authomatic method (heuristic) –Selecting pairs synset-word above 85% CS –Some manual correction of this Subset 1 (mainly, filling gaps) n Verbs: –3600 English verbs connected to WN1.5 senses and ambiguously translated to Spanish are manually inspected and disambiguated", "description": "Spanish WordNet Main Steps: Subset 1 (Semi-automatic)  ouns:  ouns: –Applying authomatic methods using bi-lingual dictionaries –Manual validation of several subsets to check if the link is correct –Deriving a Confidence Score (CS) for every authomatic method (heuristic) –Selecting pairs synset-word above 85% CS –Some manual correction of this Subset 1 (mainly, filling gaps) n Verbs: –3600 English verbs connected to WN1.5 senses and ambiguously translated to Spanish are manually inspected and disambiguated", "width": "800" } 26 Spanish WordNet Main Steps: Subset 1 (Results 1) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_26.jpg", "name": "Spanish WordNet Main Steps: Subset 1 (Results 1)", "description": "Spanish WordNet Main Steps: Subset 1 (Results 1)", "width": "800" } 27 Spanish WordNet Main Steps: Subset 1 (Results 2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_27.jpg", "name": "Spanish WordNet Main Steps: Subset 1 (Results 2)", "description": "Spanish WordNet Main Steps: Subset 1 (Results 2)", "width": "800" } 28 Spanish WordNet Main Steps: Subset 2 Main goals n enhance the quality of the Subset 1 by manual revision n extend it by manual building of synsets n 4 Sub-tasks { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_28.jpg", "name": "Spanish WordNet Main Steps: Subset 2 Main goals n enhance the quality of the Subset 1 by manual revision n extend it by manual building of synsets n 4 Sub-tasks", "description": "Spanish WordNet Main Steps: Subset 2 Main goals n enhance the quality of the Subset 1 by manual revision n extend it by manual building of synsets n 4 Sub-tasks", "width": "800" } 29 Spanish WordNet Main Steps: Subset 2 1) Covering manually those gaps in the hyponymy chains covered by other languages 2) Manual cleaning of some automatically-generated variants. –(a) pairs of synsets which are adjacent in the hyponymy chain and share at least one variant. n deleting redundant variants n re-locating to either pre-existant or newly created synsets –(b) multi-word expressions present in synsets. n Deleting non-lexicalized { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_29.jpg", "name": "Spanish WordNet Main Steps: Subset 2 1) Covering manually those gaps in the hyponymy chains covered by other languages 2) Manual cleaning of some automatically-generated variants.", "description": "–(a) pairs of synsets which are adjacent in the hyponymy chain and share at least one variant. n deleting redundant variants n re-locating to either pre-existant or newly created synsets –(b) multi-word expressions present in synsets. n Deleting non-lexicalized.", "width": "800" } 30 Spanish WordNet Main Steps: Subset 2 3) Manual addition of new vocabulary which has been considered relevant. –It mainly comes from the Catalan WordNet: since we are building both wordnets in parallell, we detected those synsets which were built for Catalan and not for Spanish 4) Manual addition of cross-part of speech relations between nominal and verbal synsets. –This work has been based mainly on noun-verb pairs obtained by means of morphological criteria. (Work carried out by UNED –Madrid-) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_30.jpg", "name": "Spanish WordNet Main Steps: Subset 2 3) Manual addition of new vocabulary which has been considered relevant.", "description": "–It mainly comes from the Catalan WordNet: since we are building both wordnets in parallell, we detected those synsets which were built for Catalan and not for Spanish 4) Manual addition of cross-part of speech relations between nominal and verbal synsets. –This work has been based mainly on noun-verb pairs obtained by means of morphological criteria. (Work carried out by UNED –Madrid-).", "width": "800" } 31 Spanish WordNet Main Steps: Subset 2 (Results) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_31.jpg", "name": "Spanish WordNet Main Steps: Subset 2 (Results)", "description": "Spanish WordNet Main Steps: Subset 2 (Results)", "width": "800" } 32 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_32.jpg", "name": "", "description": "", "width": "800" } 33 Spanish WordNet Main Steps: Beyond Subset 2 n Massive Manual Checking (from Nov’98) –Using WEI –Variants automatically generated –Filling gaps in the hierachy –New vocabulary –New Adjectives { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_33.jpg", "name": "Spanish WordNet Main Steps: Beyond Subset 2 n Massive Manual Checking (from Nov’98) –Using WEI –Variants automatically generated –Filling gaps in the hierachy –New vocabulary –New Adjectives", "description": "Spanish WordNet Main Steps: Beyond Subset 2 n Massive Manual Checking (from Nov’98) –Using WEI –Variants automatically generated –Filling gaps in the hierachy –New vocabulary –New Adjectives", "width": "800" } 34 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_34.jpg", "name": "", "description": "", "width": "800" } 35 Spanish WordNet Main Steps: Beyond Subset 2 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_35.jpg", "name": "Spanish WordNet Main Steps: Beyond Subset 2", "description": "Spanish WordNet Main Steps: Beyond Subset 2", "width": "800" } 36 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_36.jpg", "name": "", "description": "", "width": "800" } 37 Spanish WordNet Main Steps: Parole Coverage { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_37.jpg", "name": "Spanish WordNet Main Steps: Parole Coverage", "description": "Spanish WordNet Main Steps: Parole Coverage", "width": "800" } 38 Spanish WordNet Current Figures –Spanish, Catalan, Basque, (English) –http://nipadio.lsi.upc.es/wei2.html { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_38.jpg", "name": "Spanish WordNet Current Figures –Spanish, Catalan, Basque, (English) –http://nipadio.lsi.upc.es/wei2.html", "description": "Spanish WordNet Current Figures –Spanish, Catalan, Basque, (English) –http://nipadio.lsi.upc.es/wei2.html", "width": "800" } 39 Combining Multiple Methods for the Automatic Construction of Multilingual WordNets German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_39.jpg", "name": "Combining Multiple Methods for the Automatic Construction of Multilingual WordNets German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya", "description": "Combining Multiple Methods for the Automatic Construction of Multilingual WordNets German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya", "width": "800" } 40 n Ten class methods –Four monosemic criteria –Four polysemic criteria –two hybrid criteria n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods... Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_40.jpg", "name": "n Ten class methods –Four monosemic criteria –Four polysemic criteria –two hybrid criteria n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods...", "description": "Outline.", "width": "800" } 41 –Four Classes Combining Multiple Methods... Ten class methods SWEW SWEWEW SWEWEWSW SWEWSW { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_41.jpg", "name": "–Four Classes Combining Multiple Methods... Ten class methods SWEW SWEWEW SWEWEWSW SWEWSW", "description": "–Four Classes Combining Multiple Methods... Ten class methods SWEW SWEWEW SWEWEWSW SWEWSW", "width": "800" } 42 –Four monosemic criteria SWEWSWEWEW Synset Synset Synset SynsetSWEWEWSWSynsetSynset SWEW SW Combining Multiple Methods... Ten class methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_42.jpg", "name": "–Four monosemic criteria SWEWSWEWEW Synset Synset Synset SynsetSWEWEWSWSynsetSynset SWEW SW Combining Multiple Methods...", "description": "Ten class methods.", "width": "800" } 43 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ Synset+ Synset+ Synset+ Synset+ Synset+SWEWSW Combining Multiple Methods... Ten class methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_43.jpg", "name": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ Synset+ Synset+ Synset+ Synset+ Synset+SWEWSW Combining Multiple Methods...", "description": "Ten class methods.", "width": "800" } 44 –Variant criterion –Field criterion SW SW Combining Multiple Methods... Ten class methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_44.jpg", "name": "–Variant criterion –Field criterion SW SW Combining Multiple Methods... Ten class methods", "description": "–Variant criterion –Field criterion SW SW Combining Multiple Methods... Ten class methods", "width": "800" } 45 n Results Combining Multiple Methods... Ten class methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_45.jpg", "name": "n Results Combining Multiple Methods... Ten class methods", "description": "n Results Combining Multiple Methods... Ten class methods", "width": "800" } 46 n Conceptual Distance (Agirre et al. 94) –length of the shortest path –specificity of the concepts n using WordNet n Bilingual dictionary Combining Multiple Methods... Conceptual Distance methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_46.jpg", "name": "n Conceptual Distance (Agirre et al.", "description": "94) –length of the shortest path –specificity of the concepts n using WordNet n Bilingual dictionary Combining Multiple Methods... Conceptual Distance methods.", "width": "800" } 47 n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods... Conceptual Distance methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_47.jpg", "name": "n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods...", "description": "Conceptual Distance methods.", "width": "800" } 48 abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods... Conceptual Distance methods (Example CD2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_48.jpg", "name": " abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods...", "description": "Conceptual Distance methods (Example CD2).", "width": "800" } 49 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods... Conceptual Distance methods (Example CD2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_49.jpg", "name": " 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods...", "description": "Conceptual Distance methods (Example CD2).", "width": "800" } 50 n Results Combining Multiple Methods... Three CD methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_50.jpg", "name": "n Results Combining Multiple Methods... Three CD methods", "description": "n Results Combining Multiple Methods... Three CD methods", "width": "800" } 51 n Results Combining Multiple Methods... Combining methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_51.jpg", "name": "n Results Combining Multiple Methods... Combining methods", "description": "n Results Combining Multiple Methods... Combining methods", "width": "800" } 52 Combining Multiple Methods... Resulting Spanish WordNets { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_52.jpg", "name": "Combining Multiple Methods... Resulting Spanish WordNets", "description": "Combining Multiple Methods... Resulting Spanish WordNets", "width": "800" } 53 Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_53.jpg", "name": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "description": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "width": "800" } 54 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_54.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 55 Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_55.jpg", "name": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "description": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "width": "800" } 56 C1 C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_56.jpg", "name": "C1 C2 C3 C5 C6 C4", "description": "C1 C2 C3 C5 C6 C4", "width": "800" } 57 Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al. 98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_57.jpg", "name": "Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al.", "description": "98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting.", "width": "800" } 58 animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_58.jpg", "name": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "description": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "width": "800" } 59 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_59.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 60 –Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_60.jpg", "name": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "description": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "width": "800" } 61 1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels. 4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_61.jpg", "name": "1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels.", "description": "4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm.", "width": "800" } 62 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_62.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 63 –Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++ { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_63.jpg", "name": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "description": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "width": "800" } 64 –II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_64.jpg", "name": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "description": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "width": "800" } 65 –AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_65.jpg", "name": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "description": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "width": "800" } 66 –IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_66.jpg", "name": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "description": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "width": "800" } 67 –AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_67.jpg", "name": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "description": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "width": "800" } 68 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_68.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 69 –Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec. Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_69.jpg", "name": "–Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec.", "description": "Cov..", "width": "800" } 70 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_70.jpg", "name": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "description": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "width": "800" } 71 PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_71.jpg", "name": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "description": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "width": "800" } 72 piel visón marta (substance ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_72.jpg", "name": "piel visón marta (substance )", "description": "piel visón marta (substance )", "width": "800" } 73 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_73.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 74 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_74.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR.", "width": "800" } 75 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_75.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs).", "width": "800" } 76 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_76.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs.", "width": "800" } 77 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_77.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F.", "width": "800" } 78 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_78.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G.", "width": "800" } 79 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_79.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G.", "width": "800" } 80 A Complete... ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_80.jpg", "name": "A Complete...", "description": "ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6.", "width": "800" } 81 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_81.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym.", "width": "800" } 82 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_82.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy.", "width": "800" } 83 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_83.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall.", "width": "800" } 84 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_84.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall.", "width": "800" } 85 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_85.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall.", "width": "800" } 86 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_86.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall.", "width": "800" } 87 –First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_87.jpg", "name": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "description": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "width": "800" } 88 –mapping other structures n WN-EDR, WN-LDOCE, etc. n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_88.jpg", "name": "–mapping other structures n WN-EDR, WN-LDOCE, etc.", "description": "n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work.", "width": "800" } 89 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_89.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "width": "800" } 90 MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_90.jpg", "name": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "description": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "width": "800" } 91 MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,... comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_91.jpg", "name": "MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,...", "description": "comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation.", "width": "800" } 92 Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_92.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "description": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "width": "800" } 93 Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_93.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "description": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "width": "800" } 94 Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_94.jpg", "name": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "description": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "width": "800" } 95 Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION “The transfer of possession event where the agent transfers an object to its possession” IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_95.jpg", "name": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "description": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "width": "800" } 96 Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc. Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_96.jpg", "name": "Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc.", "description": "Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser.", "width": "800" } 97 Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_97.jpg", "name": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "description": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "width": "800" } 98 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_98.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "width": "800" } 99 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_99.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "width": "800" } 100 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_100.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "width": "800" } 101 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_101.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "width": "800" } 102 Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_102.jpg", "name": "Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu.", "description": "El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España.", "width": "800" } 103 Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_103.jpg", "name": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "description": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "width": "800" } 104 Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_104.jpg", "name": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "description": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "width": "800" } 105 Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_105.jpg", "name": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "description": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "width": "800" } 106 Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_106.jpg", "name": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "description": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "width": "800" } 107 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_107.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "width": "800" } 108 WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_108.jpg", "name": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "description": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "width": "800" } 109 WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_109.jpg", "name": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "description": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "width": "800" } 110 WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_110.jpg", "name": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "description": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "width": "800" } 111 WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE--> { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_111.jpg", "name": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "description": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "width": "800" } 112 WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_112.jpg", "name": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "description": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "width": "800" } 113 WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_113.jpg", "name": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "description": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "width": "800" } 114 WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_114.jpg", "name": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "description": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "width": "800" } 115 WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_115.jpg", "name": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "description": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "width": "800" } 116 WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_116.jpg", "name": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "description": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "width": "800" } 117 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_117.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "width": "800" } 118 ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,... A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_118.jpg", "name": "ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,...", "description": "A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager.", "width": "800" } 119 ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_119.jpg", "name": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "description": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "width": "800" } 120 ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_120.jpg", "name": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "description": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "width": "800" } 121 ThoughtTreasureExample Who created Bugs Bunny? Who created Bugs Bunny? 1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_121.jpg", "name": "ThoughtTreasureExample Who created Bugs Bunny. Who created Bugs Bunny.", "description": "1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)).", "width": "800" } 122 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_122.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "width": "800" } 123 n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al. 98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_123.jpg", "name": "n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al.", "description": "98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview.", "width": "800" } 124 Meaning Exploiting EWN Semantic Relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_124.jpg", "name": "Meaning Exploiting EWN Semantic Relations ", "description": "Meaning Exploiting EWN Semantic Relations ", "width": "800" } 125 Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3. La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, “como partido”, han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_125.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3.", "description": "La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, como partido , han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos..", "width": "800" } 126 Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan. Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_126.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan.", "description": "Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo..", "width": "800" } 127 Multilingual Central Repository ItalianEWN BasqueEWNSpanishEWN EnglishEWN Basque Web Corpus Italian Web Corpus English Web Corpus CatalanEWN Spanish Web Corpus Catalan Web Corpus ACQ ACQACQ ACQ UPLOADUPLOAD UPLOADUPLOAD PORT PORT PORT PORT WSD WSD WSD WSDMeaningArquitecture

9 Proyecto LE-2 4003 Proyecto LE-2 4003 Telematics Application Programme de la UETelematics Application Programme de la UE Redes semánticas de diversas lenguas Redes semánticas de diversas lenguas Integradas e interconectadas Integradas e interconectadas InglésUniversidad de SheffieldInglésUniversidad de Sheffield HolandésUniv. de AmsterdamHolandésUniv. de Amsterdam ItalianoI.L.C. de PisaItalianoI.L.C. de Pisa EspañolUB, UPC, UNED.EspañolUB, UPC, UNED. Computers and the Humanities Computers and the Humanities (Vol.monográfico,1998) (Vol.monográfico,1998) http://www.hum.uva.nl/~ewn/ http://www.hum.uva.nl/~ewn/ WordNet & EuroWordNet EuroWordNet

10 EWN2EWN2 Alemán, Francés, Checo, Sueco, Estonio Proyecto ITEMProyecto ITEM Castellano, Catalán, Vasco CREL (Centre de Referència d’Enginyeria Lingüística)CREL (Centre de Referència d’Enginyeria Lingüística) Catalán (UB, UPC) WordNet & EuroWordNet Extensiones EuroWordNet

11 Desarrollo de recursos BásicosDesarrollo de recursos Básicos Tratamiento interlingüístico de la informaciónTratamiento interlingüístico de la información - Sistemas multilingües de recuperación de información (p.e., Internet) - Módulo léxico-semántico de los sistemas de ingeniería lingüística  Extracción de información  Traducción automática WordNet & EuroWordNet Aplicaciones

12 Preservación de las relaciones semánticas específicas de cada lenguaPreservación de las relaciones semánticas específicas de cada lengua Máxima compatibilidad entre los diferentes recursosMáxima compatibilidad entre los diferentes recursos Relativa independencia de los WordNetsRelativa independencia de los WordNets en el proceso de construcciónen el proceso de construcción en el resultado finalen el resultado final WordNet & EuroWordNet Requisitos de Diseño

13

14 NúcleoNúcleo El ILIEl ILI La Top Concept Ontology (TCO)La Top Concept Ontology (TCO) Ontología de dominios (DO)Ontología de dominios (DO) PeriferiaPeriferia WordNets específicosWordNets específicos WordNet & EuroWordNet Componentes de EuroWordNet

15 Colección no estructurada de elementosColección no estructurada de elementos Ligados conLigados con al menos, un synset de un EWNal menos, un synset de un EWN un elemento de la TCO o DOun elemento de la TCO o DO Asociados a synsets de WN 1.5Asociados a synsets de WN 1.5 WordNet & EuroWordNet Interlingual Index of EuroWordNet

16 Jerarquía de conceptos independientes de la lenguaJerarquía de conceptos independientes de la lengua distinciones semánticas: objeto, lugar, dinámico, …distinciones semánticas: objeto, lugar, dinámico, … abstracta (no léxica)abstracta (no léxica) Superpuesta al ILISuperpuesta al ILI Tres tipos de entidades:Tres tipos de entidades: Primer orden: entidades concretasPrimer orden: entidades concretas Segundo orden: situaciones estáticas o dinámicasSegundo orden: situaciones estáticas o dinámicas Tercer orden: proposiciones abstractasTercer orden: proposiciones abstractas WordNet & EuroWordNet Top Concept Ontology of EuroWordNet

17

18 Jerarquía de etiquetas de dominioJerarquía de etiquetas de dominio Reducción de la polisemiaReducción de la polisemia Dominios:Dominios: Tráfico:Tráfico: Tráfico rodado, tráfico aéreoTráfico rodado, tráfico aéreo Información InternacionalInformación Internacional MicologíaMicología MedicinaMedicina WordNet & EuroWordNet Domain Ontology of EuroWordNet

19 Riqueza superior a WNRiqueza superior a WN Entre:Entre: synsets (módulos monolingües)synsets (módulos monolingües) registros ILI (multilingües):registros ILI (multilingües): {actuar-1} EQ-SYNONYM {‘behave in a certain manner’} registros ILI y TCO o ODregistros ILI y TCO o OD WordNet & EuroWordNet Relaciones de EuroWordNet

20 WordNet & EuroWordNet Relaciones Interlingüísticas de EuroWordNet

21 WordNet & EuroWordNet Relaciones de EuroWordNet

22 Spanish WordNet: Building Process German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya

23 Spanish WordNet General Methodology 1) Mapping to WN1.5 n manual work n automatic derivation of equivalents, using bi- lingual dictionaries 2) Manual correction 3) Re-structuring

24 Spanish WordNet Main Steps: First Core (Manual Translation) –Nouns: n A) WN1.5’s Tops File plus first level of hyponyms (about 800 synsets). n B) The rest of EWN’s Common Base Concepts (which were not in our set). C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  C) Manual translation of synsets intermediate between (A) and (B) following WN1.5 hyerarchy  thus building a compact taxonomy equivalent to WN1.5 without gaps  – Verbs: n Manual translation of EWN’s Base Concepts (about 150 synsets)

25 Spanish WordNet Main Steps: Subset 1 (Semi-automatic)  ouns:  ouns: –Applying authomatic methods using bi-lingual dictionaries –Manual validation of several subsets to check if the link is correct –Deriving a Confidence Score (CS) for every authomatic method (heuristic) –Selecting pairs synset-word above 85% CS –Some manual correction of this Subset 1 (mainly, filling gaps) n Verbs: –3600 English verbs connected to WN1.5 senses and ambiguously translated to Spanish are manually inspected and disambiguated

26 Spanish WordNet Main Steps: Subset 1 (Results 1)

27 Spanish WordNet Main Steps: Subset 1 (Results 2)

28 Spanish WordNet Main Steps: Subset 2 Main goals n enhance the quality of the Subset 1 by manual revision n extend it by manual building of synsets n 4 Sub-tasks

29 Spanish WordNet Main Steps: Subset 2 1) Covering manually those gaps in the hyponymy chains covered by other languages 2) Manual cleaning of some automatically-generated variants. –(a) pairs of synsets which are adjacent in the hyponymy chain and share at least one variant. n deleting redundant variants n re-locating to either pre-existant or newly created synsets –(b) multi-word expressions present in synsets. n Deleting non-lexicalized

30 Spanish WordNet Main Steps: Subset 2 3) Manual addition of new vocabulary which has been considered relevant. –It mainly comes from the Catalan WordNet: since we are building both wordnets in parallell, we detected those synsets which were built for Catalan and not for Spanish 4) Manual addition of cross-part of speech relations between nominal and verbal synsets. –This work has been based mainly on noun-verb pairs obtained by means of morphological criteria. (Work carried out by UNED –Madrid-)

31 Spanish WordNet Main Steps: Subset 2 (Results)

32

33 Spanish WordNet Main Steps: Beyond Subset 2 n Massive Manual Checking (from Nov’98) –Using WEI –Variants automatically generated –Filling gaps in the hierachy –New vocabulary –New Adjectives

34

35 Spanish WordNet Main Steps: Beyond Subset 2

36

37 Spanish WordNet Main Steps: Parole Coverage

38 Spanish WordNet Current Figures –Spanish, Catalan, Basque, (English) –http://nipadio.lsi.upc.es/wei2.html

39 Combining Multiple Methods for the Automatic Construction of Multilingual WordNets German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya

40 n Ten class methods –Four monosemic criteria –Four polysemic criteria –two hybrid criteria n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods... Outline

41 –Four Classes Combining Multiple Methods... Ten class methods SWEW SWEWEW SWEWEWSW SWEWSW

42 –Four monosemic criteria SWEWSWEWEW Synset Synset Synset SynsetSWEWEWSWSynsetSynset SWEW SW Combining Multiple Methods... Ten class methods

43 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ Synset+ Synset+ Synset+ Synset+ Synset+SWEWSW Combining Multiple Methods... Ten class methods

44 –Variant criterion –Field criterion SW SW Combining Multiple Methods... Ten class methods

45 n Results Combining Multiple Methods... Ten class methods

46 n Conceptual Distance (Agirre et al. 94) –length of the shortest path –specificity of the concepts n using WordNet n Bilingual dictionary Combining Multiple Methods... Conceptual Distance methods

47 n Three conceptual distance methods –CD1: using pairwise word coocurrences –CD2: using headword and genus –CD3: using bilingual Spanish entries with multiple translations Combining Multiple Methods... Conceptual Distance methods

48 abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods... Conceptual Distance methods (Example CD2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_48.jpg", "name": " abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods...", "description": "Conceptual Distance methods (Example CD2).", "width": "800" } 49 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods... Conceptual Distance methods (Example CD2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_49.jpg", "name": " 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods...", "description": "Conceptual Distance methods (Example CD2).", "width": "800" } 50 n Results Combining Multiple Methods... Three CD methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_50.jpg", "name": "n Results Combining Multiple Methods... Three CD methods", "description": "n Results Combining Multiple Methods... Three CD methods", "width": "800" } 51 n Results Combining Multiple Methods... Combining methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_51.jpg", "name": "n Results Combining Multiple Methods... Combining methods", "description": "n Results Combining Multiple Methods... Combining methods", "width": "800" } 52 Combining Multiple Methods... Resulting Spanish WordNets { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_52.jpg", "name": "Combining Multiple Methods... Resulting Spanish WordNets", "description": "Combining Multiple Methods... Resulting Spanish WordNets", "width": "800" } 53 Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_53.jpg", "name": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "description": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "width": "800" } 54 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_54.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 55 Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_55.jpg", "name": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "description": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "width": "800" } 56 C1 C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_56.jpg", "name": "C1 C2 C3 C5 C6 C4", "description": "C1 C2 C3 C5 C6 C4", "width": "800" } 57 Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al. 98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_57.jpg", "name": "Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al.", "description": "98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting.", "width": "800" } 58 animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_58.jpg", "name": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "description": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "width": "800" } 59 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_59.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 60 –Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_60.jpg", "name": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "description": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "width": "800" } 61 1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels. 4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_61.jpg", "name": "1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels.", "description": "4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm.", "width": "800" } 62 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_62.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 63 –Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++ { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_63.jpg", "name": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "description": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "width": "800" } 64 –II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_64.jpg", "name": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "description": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "width": "800" } 65 –AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_65.jpg", "name": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "description": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "width": "800" } 66 –IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_66.jpg", "name": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "description": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "width": "800" } 67 –AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_67.jpg", "name": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "description": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "width": "800" } 68 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_68.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 69 –Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec. Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_69.jpg", "name": "–Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec.", "description": "Cov..", "width": "800" } 70 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_70.jpg", "name": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "description": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "width": "800" } 71 PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_71.jpg", "name": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "description": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "width": "800" } 72 piel visón marta (substance ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_72.jpg", "name": "piel visón marta (substance )", "description": "piel visón marta (substance )", "width": "800" } 73 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_73.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 74 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_74.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR.", "width": "800" } 75 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_75.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs).", "width": "800" } 76 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_76.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs.", "width": "800" } 77 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_77.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F.", "width": "800" } 78 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_78.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G.", "width": "800" } 79 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_79.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G.", "width": "800" } 80 A Complete... ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_80.jpg", "name": "A Complete...", "description": "ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6.", "width": "800" } 81 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_81.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym.", "width": "800" } 82 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_82.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy.", "width": "800" } 83 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_83.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall.", "width": "800" } 84 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_84.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall.", "width": "800" } 85 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_85.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall.", "width": "800" } 86 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_86.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall.", "width": "800" } 87 –First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_87.jpg", "name": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "description": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "width": "800" } 88 –mapping other structures n WN-EDR, WN-LDOCE, etc. n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_88.jpg", "name": "–mapping other structures n WN-EDR, WN-LDOCE, etc.", "description": "n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work.", "width": "800" } 89 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_89.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "width": "800" } 90 MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_90.jpg", "name": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "description": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "width": "800" } 91 MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,... comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_91.jpg", "name": "MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,...", "description": "comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation.", "width": "800" } 92 Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_92.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "description": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "width": "800" } 93 Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_93.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "description": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "width": "800" } 94 Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_94.jpg", "name": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "description": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "width": "800" } 95 Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION “The transfer of possession event where the agent transfers an object to its possession” IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_95.jpg", "name": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "description": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "width": "800" } 96 Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc. Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_96.jpg", "name": "Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc.", "description": "Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser.", "width": "800" } 97 Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_97.jpg", "name": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "description": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "width": "800" } 98 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_98.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "width": "800" } 99 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_99.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "width": "800" } 100 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_100.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "width": "800" } 101 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_101.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "width": "800" } 102 Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_102.jpg", "name": "Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu.", "description": "El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España.", "width": "800" } 103 Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_103.jpg", "name": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "description": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "width": "800" } 104 Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_104.jpg", "name": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "description": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "width": "800" } 105 Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_105.jpg", "name": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "description": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "width": "800" } 106 Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_106.jpg", "name": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "description": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "width": "800" } 107 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_107.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "width": "800" } 108 WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_108.jpg", "name": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "description": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "width": "800" } 109 WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_109.jpg", "name": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "description": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "width": "800" } 110 WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_110.jpg", "name": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "description": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "width": "800" } 111 WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE--> { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_111.jpg", "name": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "description": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "width": "800" } 112 WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_112.jpg", "name": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "description": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "width": "800" } 113 WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_113.jpg", "name": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "description": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "width": "800" } 114 WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_114.jpg", "name": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "description": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "width": "800" } 115 WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_115.jpg", "name": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "description": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "width": "800" } 116 WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_116.jpg", "name": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "description": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "width": "800" } 117 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_117.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "width": "800" } 118 ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,... A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_118.jpg", "name": "ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,...", "description": "A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager.", "width": "800" } 119 ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_119.jpg", "name": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "description": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "width": "800" } 120 ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_120.jpg", "name": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "description": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "width": "800" } 121 ThoughtTreasureExample Who created Bugs Bunny? Who created Bugs Bunny? 1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_121.jpg", "name": "ThoughtTreasureExample Who created Bugs Bunny. Who created Bugs Bunny.", "description": "1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)).", "width": "800" } 122 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_122.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "width": "800" } 123 n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al. 98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_123.jpg", "name": "n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al.", "description": "98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview.", "width": "800" } 124 Meaning Exploiting EWN Semantic Relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_124.jpg", "name": "Meaning Exploiting EWN Semantic Relations ", "description": "Meaning Exploiting EWN Semantic Relations ", "width": "800" } 125 Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3. La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, “como partido”, han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_125.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3.", "description": "La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, como partido , han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos..", "width": "800" } 126 Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan. Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_126.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan.", "description": "Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo..", "width": "800" } 127 Multilingual Central Repository ItalianEWN BasqueEWNSpanishEWN EnglishEWN Basque Web Corpus Italian Web Corpus English Web Corpus CatalanEWN Spanish Web Corpus Catalan Web Corpus ACQ ACQACQ ACQ UPLOADUPLOAD UPLOADUPLOAD PORT PORT PORT PORT WSD WSD WSD WSDMeaningArquitecture

49 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods... Conceptual Distance methods (Example CD2) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_49.jpg", "name": " 06 ARTIFACT 06 ARTIFACT abadía_1_2 Iglesia o monasterio regido por un abad o abadesa (abbey, a church or a monastery ruled by an abbot or an abbess) Combining Multiple Methods...", "description": "Conceptual Distance methods (Example CD2).", "width": "800" } 50 n Results Combining Multiple Methods... Three CD methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_50.jpg", "name": "n Results Combining Multiple Methods... Three CD methods", "description": "n Results Combining Multiple Methods... Three CD methods", "width": "800" } 51 n Results Combining Multiple Methods... Combining methods { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_51.jpg", "name": "n Results Combining Multiple Methods... Combining methods", "description": "n Results Combining Multiple Methods... Combining methods", "width": "800" } 52 Combining Multiple Methods... Resulting Spanish WordNets { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_52.jpg", "name": "Combining Multiple Methods... Resulting Spanish WordNets", "description": "Combining Multiple Methods... Resulting Spanish WordNets", "width": "800" } 53 Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_53.jpg", "name": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "description": "Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC", "width": "800" } 54 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_54.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 55 Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_55.jpg", "name": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "description": "Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4", "width": "800" } 56 C1 C2 C3 C5 C6 C4 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_56.jpg", "name": "C1 C2 C3 C5 C6 C4", "description": "C1 C2 C3 C5 C6 C4", "width": "800" } 57 Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al. 98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_57.jpg", "name": "Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al.", "description": "98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting.", "width": "800" } 58 animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_58.jpg", "name": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "description": "animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )", "width": "800" } 59 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_59.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 60 –Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_60.jpg", "name": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "description": "–Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm", "width": "800" } 61 1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels. 4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_61.jpg", "name": "1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels.", "description": "4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm.", "width": "800" } 62 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_62.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 63 –Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++ { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_63.jpg", "name": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "description": "–Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++", "width": "800" } 64 –II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_64.jpg", "name": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "description": "–II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO", "width": "800" } 65 –AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_65.jpg", "name": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "description": "–AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +", "width": "800" } 66 –IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_66.jpg", "name": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "description": "–IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +", "width": "800" } 67 –AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + + { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_67.jpg", "name": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "description": "–AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +", "width": "800" } 68 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_68.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 69 –Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec. Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_69.jpg", "name": "–Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec.", "description": "Cov..", "width": "800" } 70 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_70.jpg", "name": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "description": "–Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.", "width": "800" } 71 PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_71.jpg", "name": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "description": "PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results", "width": "800" } 72 piel visón marta (substance ) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_72.jpg", "name": "piel visón marta (substance )", "description": "piel visón marta (substance )", "width": "800" } 73 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_73.jpg", "name": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "description": "–Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline", "width": "800" } 74 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_74.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR.", "width": "800" } 75 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_75.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs).", "width": "800" } 76 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_76.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs.", "width": "800" } 77 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_77.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F.", "width": "800" } 78 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_78.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G.", "width": "800" } 79 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_79.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G.", "width": "800" } 80 A Complete... ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_80.jpg", "name": "A Complete...", "description": "ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6.", "width": "800" } 81 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_81.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym.", "width": "800" } 82 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_82.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy.", "width": "800" } 83 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_83.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall.", "width": "800" } 84 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_84.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall.", "width": "800" } 85 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_85.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall.", "width": "800" } 86 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_86.jpg", "name": "A Complete WN1.5 to WN1.6 Mapping...", "description": "ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall.", "width": "800" } 87 –First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_87.jpg", "name": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "description": "–First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions", "width": "800" } 88 –mapping other structures n WN-EDR, WN-LDOCE, etc. n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_88.jpg", "name": "–mapping other structures n WN-EDR, WN-LDOCE, etc.", "description": "n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work.", "width": "800" } 89 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_89.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos", "width": "800" } 90 MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_90.jpg", "name": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "description": "MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action", "width": "800" } 91 MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,... comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_91.jpg", "name": "MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,...", "description": "comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation.", "width": "800" } 92 Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_92.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "description": "Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic", "width": "800" } 93 Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_93.jpg", "name": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "description": "Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION", "width": "800" } 94 Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_94.jpg", "name": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "description": "Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines", "width": "800" } 95 Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION “The transfer of possession event where the agent transfers an object to its possession” IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_95.jpg", "name": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "description": "Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION The transfer of possession event where the agent transfers an object to its possession IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN", "width": "800" } 96 Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc. Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_96.jpg", "name": "Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc.", "description": "Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser.", "width": "800" } 97 Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_97.jpg", "name": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "description": "Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections", "width": "800" } 98 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_98.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance", "width": "800" } 99 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_99.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four", "width": "800" } 100 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_100.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN", "width": "800" } 101 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_101.jpg", "name": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "description": "Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake", "width": "800" } 102 Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_102.jpg", "name": "Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu.", "description": "El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España.", "width": "800" } 103 Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_103.jpg", "name": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "description": "Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN", "width": "800" } 104 Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_104.jpg", "name": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "description": "Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts", "width": "800" } 105 Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_105.jpg", "name": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "description": "Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997", "width": "800" } 106 Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_106.jpg", "name": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "description": "Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797", "width": "800" } 107 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2 { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_107.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2", "width": "800" } 108 WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_108.jpg", "name": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "description": "WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do", "width": "800" } 109 WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_109.jpg", "name": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "description": "WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS", "width": "800" } 110 WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_110.jpg", "name": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "description": "WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)", "width": "800" } 111 WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE--> { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_111.jpg", "name": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "description": "WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->", "width": "800" } 112 WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_112.jpg", "name": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "description": "WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT", "width": "800" } 113 WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_113.jpg", "name": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "description": "WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles", "width": "800" } 114 WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_114.jpg", "name": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "description": "WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences", "width": "800" } 115 WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_115.jpg", "name": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "description": "WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator", "width": "800" } 116 WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_116.jpg", "name": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "description": "WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations", "width": "800" } 117 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_117.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure", "width": "800" } 118 ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,... A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_118.jpg", "name": "ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,...", "description": "A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager.", "width": "800" } 119 ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_119.jpg", "name": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "description": "ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts", "width": "800" } 120 ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_120.jpg", "name": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "description": "ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents", "width": "800" } 121 ThoughtTreasureExample Who created Bugs Bunny? Who created Bugs Bunny? 1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_121.jpg", "name": "ThoughtTreasureExample Who created Bugs Bunny. Who created Bugs Bunny.", "description": "1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny)).", "width": "800" } 122 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_122.jpg", "name": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "description": "German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning", "width": "800" } 123 n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al. 98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_123.jpg", "name": "n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al.", "description": "98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview.", "width": "800" } 124 Meaning Exploiting EWN Semantic Relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_124.jpg", "name": "Meaning Exploiting EWN Semantic Relations ", "description": "Meaning Exploiting EWN Semantic Relations ", "width": "800" } 125 Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3. La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, “como partido”, han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_125.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3.", "description": "La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, como partido , han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos..", "width": "800" } 126 Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan. Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_126.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan.", "description": "Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo..", "width": "800" } 127 Multilingual Central Repository ItalianEWN BasqueEWNSpanishEWN EnglishEWN Basque Web Corpus Italian Web Corpus English Web Corpus CatalanEWN Spanish Web Corpus Catalan Web Corpus ACQ ACQACQ ACQ UPLOADUPLOAD UPLOADUPLOAD PORT PORT PORT PORT WSD WSD WSD WSDMeaningArquitecture

50 n Results Combining Multiple Methods... Three CD methods

51 n Results Combining Multiple Methods... Combining methods

52 Combining Multiple Methods... Resulting Spanish WordNets

53 Mapping Conceptual Hierarchies Using Relaxation Labelling German Rigau i Claramunt TALP Research Center UPC

54 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline

55 Mapping Conceptual Hierarchies using Relaxation Labelling Setting C1C2 C3 C5 C6 C4

56 C1 C2 C3 C5 C6 C4

57 Connecting already existing Hierarchies –Relaxattion labelling Algorithn –Constraints Between –Spanish taxonomy automatically derived from an MRD (Rigau et al. 98) –WordNet n using a bilingual MRD Mapping Conceptual Hierarchies using Relaxation Labelling Setting

58 animal ave rapaz faisán (Tops ) (person ) (animal ) (artifact ) (food ) (person ) (animal ) (food ) (animal ) (artifact ) (food ) (person )

59 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline

60 –Iterative algorithm for function optimization based on local information –it can deal with any kind of constraints n variables (senses of the taxonomy) n labels (synsets) –Finds a weight assignment for each possible label for each variable n weights for the labels of the same variable add up to one n weigth assignation satisfies -to the maximum possible extent- the set of constraints Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm

61 1) Start with a random weight assigment 2) Compute the support value for each label of each variable (according to the constraints) 3) Increase the weights of the labels more compatible with context and decrease those and decrease those of the less compatible labels. 4) If a stopping/convergence is satisfied, stop, otherwiese go to step 2. Mapping Conceptual Hierarchies using Relaxation Labelling Relaxation Labelling Algorithm

62 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline

63 –Rely on the taxonomy structure –Coded with three characters n X: Spanish Taxonomy,I (immediate), n Y: English Taxonomy,A (ancestor) n X: Relation, E (hypernym), O (hyponym), B (both) –Examples: Mapping Conceptual Hierarchies using Relaxation Labelling Constraints IIEAAB ++ ++

64 –II Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints NAACL’2001 IIEIIBIIO

65 –AI Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AIEAIB + + NAACL’2001 AIO + +

66 –IA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints IAEIAB + + NAACL’2001 IAO + +

67 –AA Constraints Mapping Conceptual Hierarchies using Relaxation Labelling Hierarchical Constraints AAEAAB + + NAACL’2001 AAO + + + + + +

68 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline

69 –Four monosemic criteria SWEWSWEWEW Synset 92%5% Synset89%1% Synset Synset89%2% SWEWEWSW Synset85%4% Synset SWEW SW Combining Multiple Methods...RANLP’97 Eight class methods Prec. Cov.

70 –Four polysemic criteria SWEWSWEWEW SWEW EWSW Synset+ 80% 8% Synset+ 75% 2% Synset+ Synset+ 58% 17% Synset+ 61% 60% Synset+SWEWSW Combining Multiple Methods...RANLP’97 Eight class methods Prec.Cov.

71 PolyTOK, FOKTOK, FNOKtotal animal279 (90%)30 (91%)209 (90%) food166 (94%)3 (100%)169 (94%) cognition198 (67%)27 (90%)225 (69%) communication533 (77%)40 (97%)573 (78%) allTOK, FOKTOK, FNOKtotal animal424 (93%)62 (95%)486 (90%) food166 (94%)83 (100%)249 (96%) cognition200 (67%)245 (90%)445 (82%) communication536 (77%)234 (97%)760 (81%) Combining Multiple Methods...RANLP’97 Experiments & Results

72 piel visón marta (substance )

73 –Setting –Relaxation Labelling Algorithm –Constraints –Experiments & Results I (multilingual) –Experiments & Results II (monolingual) –Further work Mapping Conceptual Hierarchies using Relaxation Labelling Outline

74 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n All Relationships –also-see, similar-to, attribute, antonym, etc. RR

75 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Generalized Constraints n Non-structural constraints –W: number of word coincidences –G: word coincidences in glosses –F: number of frame coincidences (verbs)

76 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 POS mapping depencences Nouns Adjectives Verbs Adverbs

77 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints for Verbs n Structural constraints –hyper/hyponymy –antonymy –also-see n Non-structural constraints –W, G and F

78 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adjectives n Structural constraints –Adj-to-Adj n antonymy, similar-to and also-see –Adj-to-Verb n participle-of –Adj-to-Noun n pertains and attribute n Non-structural constraints –W and G

79 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Constraints Adverbs n Structural constraints –Adv-to-Adv n antonymy –Adv-to-Adj n derived n Non-structural constraints –W and G

80 A Complete... ACL’00, NAACL’01 Example extra-POS 02025107a evangelical evangelistic 04237485n Gospel Gospels evangel 00843344a evangelical evangelistic 02025107aevangelical 04853575n Gospel Gospels evangel 00842521aenthusiastic pertainym Similar to WN1.5 WN1.6

81 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Example extra-POS 00057615r impossibly absurdly 01393725aimpossible 00294844rimpossibly derived fromWN1.5WN1.6 01752468aimpossible 00294658apossibly antonym

82 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints –Nouns: AA hyper/hyponym –Verbs: AA hyper/hyponym, II also-see –Adjectives: II antonymy, similar-to, also-see –Adverbs: II antonymy

83 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set: structural constraints CoverageAmbigousOverallN V A R80.8% 94.1% 96.9% 99.7% 94.9% - 99.6% 97.6% - 99.8% 82.8% - 98.9% 97.5% - 100% 93.5% - 99.2% 94.6% - 99.2% 89.5% - 99.4% 99.0% - 100% Precision - recall

84 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic constraint set + W, G and F for verbs CoverageAmbigousOverallN V A R99.5% 98.9% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 98.8% 97.5% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.3% 99.0% - 100% Precision - recall

85 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships CoverageAmbigousOverallN V A R88.0% 95.8% - --- 95.8% - 98.9% 69.2% - 94.2% -- 90.9% - 99.4% 97.9% - 98.1% Precision - recall

86 A Complete WN1.5 to WN1.6 Mapping... ACL’00, NAACL’01 Results n Basic + extra-POS relationships + WGF CoverageAmbigousOverallN V A R99.6% 99.0% 99.8% 99.9% 97.5% - 97.7% 98.8% - 98.9% 96.5% - 99.1% 98.3% - 100% 99.4% - 99.7% 99.3% - 99.6% 97.9% - 99.5% 99.3% - 100% Precision - recall

87 –First complete mapping between Wordnet versions –Combining structural and non-structural information –Robust approach based on local information, but with global effects –Incremental POS approach –http://www.lsi.upc.es/~nlp –90 downloads (since November 2000) Mapping Conceptual Hierarchies using Relaxation Labelling Conclusions

88 –mapping other structures n WN-EDR, WN-LDOCE, etc. n Other language taxonomies to EuroWordNet –SpanishEWN to WN1.6 –symmetrical philosophy rather than source- target Mapping Conceptual Hierarchies using Relaxation Labelling Further Work

89 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Mikrokosmos

90 MikrokosmosOutline Introduction Introduction Representational Issues Representational Issues The Lexicon The Lexicon The Ontology The Ontology Acquisition Process Acquisition Process Lexicon Acquisition Lexicon Acquisition Guidelines Guidelines Ontology/Lexicon Trade-off Ontology/Lexicon Trade-off Semantics in Action Semantics in Action

91 MikrokosmosIntroduction Knowledge Base Machine Translation (KBMT) Knowledge Base Machine Translation (KBMT) CRL, NMSU CRL, NMSU 5,000 concepts 5,000 concepts Events Events Objects Objects Properties Properties 7,000 Spanish word senses 7,000 Spanish word senses 40,000 word senses 40,000 word senses after expansion with productive Lexical Rules after expansion with productive Lexical Rules comprar -> comprador, comprable,... comprar -> comprador, comprable,... Text Meaning Representation Text Meaning Representation

92 Mikrokosmos Representational Issues: The Lexicon Typed Feature Structures (Pollard and Sag 87) Typed Feature Structures (Pollard and Sag 87) language-dependant language-dependant 10 zones 10 zones phonology phonology orthography orthography morphology morphology Syntactic (subcategorization) Syntactic (subcategorization) Semantic (Lexical Semantic Representation) Semantic (Lexical Semantic Representation) syntax-semantic linking syntax-semantic linking stylistics stylistics paradigmatic paradigmatic syntacmatic syntacmatic

93 Mikrokosmos Representational Issues: The Lexicon Adquirir-V1 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: OBJECT Adquirir-V2 syn:subj: cat: NP obj:cat:NP sem:acquire agent:HUMAN theme: INFORMATION

94 Mikrokosmos Representational Issues: The Ontology Taxonomic multi-hierarchical Taxonomic multi-hierarchical 14 local or inherited links in average 14 local or inherited links in average language-impartial language-impartial EVENTS, OBJECTS, PROPERTIES EVENTS, OBJECTS, PROPERTIES Methodology & Guidelines Methodology & Guidelines

95 Mikrokosmos Representational Issues: The Ontology ACQUIRE ACQUIRE DEFINITION “The transfer of possession event where the agent transfers an object to its possession” IS - A TRANSFER-POSSESSION SOURCEHUMAN PLACE THEMEOBJECT (NOT HUMAN) AGENTANIMAL (DEFAULT HUMAN) DESTINATIONANIMAL PLACE (DEFAULT HUMAN) INHERITED BENEFICIARYHUMAN

96 Mikrokosmos Acquisition Process: The Lexicon Multi-lingual Multi-lingual French, English, Japanese, Russian, Spanish, etc.French, English, Japanese, Russian, Spanish, etc. Multi-media Multi-media Multi-process Multi-process Analysis Analysis Generation (mono and multilingual) Generation (mono and multilingual) MT MT Summarization Summarization IE IE Speech Processing Speech Processing Tools Tools corpus-search, lookup dictionary, ontology browser corpus-search, lookup dictionary, ontology browser

97 Mikrokosmos Acquisition Process: The Ontology Guidelines Guidelines 1) Do not add instances as concepts Instances do not have their own instances Instances do not have their own instances Concepts do not have fixed position in space/time Concepts do not have fixed position in space/time 2) Do not decompose concepts further 3) Use close concepts 4) Do not add EVENTs with particular arguments 5) Do not add concepts with instance-specific aspects, temporal relations 6) Do not add language-specific concepts 7) Do not add ontologycal concepts for collections

98 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Daily negociations Daily negociations lexicon acquirers lexicon acquirers ontology acquirers ontology acquirers Possibilities Possibilities one-to-one mapping one-to-one mapping lexicon unspecification lexicon unspecification lexicon ontology balance lexicon ontology balance

99 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off one-to-one mapping one-to-one mapping Problems Problems Lexical: every word in a language is a concept Lexical: every word in a language is a concept conceptual: cuire in french is not ambiguous conceptual: cuire in french is not ambiguous PREPARE-FOOD INST: COOKING-EQUIPMENT COOK INST: STOVE BAKE INST: OVEN cook : cuire sur le feu bake : cuire ou four

100 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon Unspecification Lexicon Unspecification Problems Problems BAKE is not in the ontology BAKE is not in the ontology PREPARE-FOOD INST: COOKING-EQUIPMENT cook : cuire sur le feu bake : cuire ou four INST: OVEN

101 Mikrokosmos Acquisition Process: Ontology/Lexicon Trade-off Lexicon-Ontology Balance Lexicon-Ontology Balance PREPARE-FOOD INST: COOKING-EQUIPMENT FRY INST: STOVE INST: FRYING-PAN BAKE INST: OVEN cook : cuire bake

102 Mikrokosmos Semantics in Action El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche, a través de su compañía en España, adquirió Doctor Andreu. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. El grupo Roche adquirió Doctor Andreu a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. La adquisición de Doctor Andreu por el grupo Roche fue hecha a través de su compañía en España. ACQUIRE-1Agent: ORGANIZATION-1 Theme: ORGANIZATION-2 Instrument: ORGANIZATION-3 ORGANIZATION-1 Object-Name: Grupo Roche ORGANIZATION-2Object-Name: Doctor Andreu ORGANIZATION-3Location: España

103 Mikrokosmos Semantics in Action Onto-Search: Ontological search mechanism to check constraints Onto-Search: Ontological search mechanism to check constraints check-onto(ACQUIRE, EVENT) = 1 check-onto(ACQUIRE, EVENT) = 1 since ACQUIRE is a type of EVENT since ACQUIRE is a type of EVENT check-onto(ORGANIZATION, HUMAN) = 0.9 check-onto(ORGANIZATION, HUMAN) = 0.9 since ORGANIZATION HAS-MEMBER HUMAN since ORGANIZATION HAS-MEMBER HUMAN

104 Mikrokosmos Semantics in Action 1) a-través-de INSTRUMENT, LOCATION adquirir require PHYSICAL-OBJECT 2) en LOCATION, TEMPORAL España is not a TEMPORAL-OBJECT 3) adquirir ACQUIRE, LEARN Doctor Andreu is not an INFORMATION 4) Doctor Andreu ORGANIZATION, HUMAN the Theme of ACQUIRE is not HUMAN 5) compañía CORPORATION, SOCIAL-EVENT ORGANIZATIONs typically fill the INSTRUMENT slot of ACQUIRE acts

105 Mikrokosmos Experiment: WSD Text1234Mean words347385370353364 words/sentence16.524.026.420.821.4 open-class words183167177177176 ambiguous words5742573548 syntax2119201218 correct5141453443 %9799939997

106 Mikrokosmos Experiment: WSD TextMeanMean Unseen words364390 words/sentence21.426 open-class words176104 ambiguous words4826 syntax189 correct4323 %9797

107 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya WordNet2

108 WordNet2Outline Introduction Introduction Text Inferences Text Inferences Defining Features Defining Features Plausible inferences Plausible inferences Inference Rules Inference Rules Semantic Paths Semantic Paths What WordNet cannot do What WordNet cannot do

109 WordNet2Introduction (Harabagiu 98) (Harabagiu 98) Commonse reasoning requires extensive knowledge Commonse reasoning requires extensive knowledge ~ 100 millions of concepts and relations ~ 100 millions of concepts and relations WordNet WordNet represents almost all English words represents almost all English words 100.000 synsets 100.000 synsets linked by semantic relations linked by semantic relations WordNet2 WordNet2 each synset has a gloss that, when disambiguated may increase the number of relations each synset has a gloss that, when disambiguated may increase the number of relations WordNet glosses into semantic networks WordNet glosses into semantic networks NEW RELATIONS NEW RELATIONS

110 WordNet2 Text Inferences German was hungry He opened the refrigerator hungry (feeling a need or desire to eat) hungry (feeling a need or desire to eat) eat (take in solid food) eat (take in solid food) refrigerator (an appliance in which foods can be stored at low temperature) refrigerator (an appliance in which foods can be stored at low temperature)

111 WordNet2 Defining Features Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links Transform each concept’s gloss into a graph where concepts are nodes and lexical relations are links (all the knowledge shared by society) (all the knowledge shared by society) --AGENT--> --AGENT--> (licensed medical practitioner) (licensed medical practitioner) --ATRIBUTTE--> --ATRIBUTTE-->

112 WordNet2 Defining Features pilotperson qualified guide water difficult GLOSS ATTRIBUTE PURPOSELOCATION ATTRIBUTE ship OBJECT

113 WordNet2 Inference Rules Rule 1Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2IS-AVC3VC2ENTAILVC3 -------------------------------------------------- VC1IS-AVC3VC1ENTAILVC3 Rule 3Rule 2 VC1IS-AVC2VC1IS-AVC2 VC2R_IS-AVC3VC2R_ENTAIL VC3 -------------------------------------------------- VC1PLAUSIBLE (not VC3)VC1EXPLAINS VC3 16 + 1 regles 16 + 1 regles

114 WordNet2 Semantic Paths 0) Create and load the KB 1) Place markers on KB concepts 2) Propagate markers The algorithm avoids cycles 3) Detect collisions To each marker collision it corresponds a path 4) Extract Inferences

115 WordNet2 Semantic Paths Inference sequence German was hungry German was hungry German felt a desire to eat German felt a desire to eat German felt a desire to take in food German felt a desire to take in food COLLISION: German=he felt a desire to take food, stored in an appliance, which he opened He opened an appliance where food is stored He opened an appliance where food is stored He opened the refrigerator He opened the refrigerator

116 WordNet2 What WordNet cannot do Major WordNet limitations: 1) The lack of compound concepts 2) The small number of causation and entailment relations 3) the lack of preconditions for verbs 4) the absence of case relations

117 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya ThoughtTreasure

118 ThoughtTreasureOverview a comprehensive platform for a comprehensive platform for NLP English, French NLP English, French commonsense reasoning commonsense reasoning A hotel room has a bed, night table,... A hotel room has a bed, night table,... People has fingernails People has fingernails soda is a drink soda is a drink one hangs up at the end of a phone call one hangs up at the end of a phone call the sky is blue the sky is blue dogs bark dogs bark someone who is 16 years old is a teenager someone who is 16 years old is a teenager

119 ThoughtTreasureOverview 25,000 concepts organized into a hierarchy 25,000 concepts organized into a hierarchy EVIAN -> FLAT-WATER -> DRINKING-WATER 55,000 words (English, French)55,000 words (English, French) food aliment FOOD 50,000 asertions about concepts50,000 asertions about concepts green-pea is green 100 scripts100 scripts

120 ThoughtTreasureOverview Text Agents for recognizing names, phones, etc Text Agents for recognizing names, phones, etc mechanisms for learning new words mechanisms for learning new words X-phile is someone who likes XX-phile is someone who likes X a syntactic parser a syntactic parser a NL generator a NL generator a semantic parser a semantic parser an anaphoric parser an anaphoric parser planning agents for achieving goals planning agents for achieving goals understanding agents understanding agents

121 ThoughtTreasureExample Who created Bugs Bunny? Who created Bugs Bunny? 1.0 (create human-interrogative-pronoun Bugs-Bunny) 1.0 (create human-interrogative-pronoun Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 0.9 (create rock-group-the-Who Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 1.0 (create Tex-Avery Bugs-Bunny) 0.1 (not (create rock-group-the-Who Bugs-Bunny)) 0.1 (not (create rock-group-the-Who Bugs-Bunny))

122 German Rigau i Claramunt http://www.lsi.upc.es/~rigau TALP Research Center Departament de Llenguatges i Sistemes Informàtics Universitat Politècnica de Catalunya Meaning

123 n Bases de Conocimiento –Enriquecimiento automático de EWN (modelos verbales, etc.) –Aproximación mixta (KB + ML) –Q/A n Problema –ambigüedad estructural y léxica n Aproximación –localizar automáticamente ejemplos de sentidos (Leacock et al. 98, Mihalcea y Moldovan 99) –WSD a gran escala (Boosting, SVM, transductivos …) –Acquisición Conocimiento (Ribas 95, McCarthy 01) MeaningOverview

124 Meaning Exploiting EWN Semantic Relations { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_124.jpg", "name": "Meaning Exploiting EWN Semantic Relations ", "description": "Meaning Exploiting EWN Semantic Relations ", "width": "800" } 125 Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3. La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, “como partido”, han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_125.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3.", "description": "La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, como partido , han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos..", "width": "800" } 126 Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan. Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo. { "@context": "http://schema.org", "@type": "ImageObject", "contentUrl": "http://images.slideplayer.es/12/3561503/slides/slide_126.jpg", "name": "Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan.", "description": "Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo..", "width": "800" } 127 Multilingual Central Repository ItalianEWN BasqueEWNSpanishEWN EnglishEWN Basque Web Corpus Italian Web Corpus English Web Corpus CatalanEWN Spanish Web Corpus Catalan Web Corpus ACQ ACQACQ ACQ UPLOADUPLOAD UPLOADUPLOAD PORT PORT PORT PORT WSD WSD WSD WSDMeaningArquitecture

125 Meaning Exploiting EWN Semantic Relations partido 1 Todos los partidos piden reformas legales para TV3. La derecha planea agruparse en un partido. El diputado reiteró que ni él ni UDC, “como partido”, han recibido dinero de Pellerols. partido 2 Pero España puso al partido intensidad, ritmo y coraje. El seleccionador cree que el partido de hoy contra Italia dará la medida de España El Racing no gana en su campo desde hace seis partidos.

126 Meaning Exploiting EWN Semantic Relations partido 1 No negociaremos nunca com un partido político que sea partidario de la independencia de Taiwan. Una vez más es noticia la desviación de fondos destinadoss a la formación ocupacional hacia la financiación de un partido político. Estas lleyess fueron votadas gracias a un consenso general de los partidos políticos. partido 2 Rivera pide el suporte de la afición para encarrilar las semifinales. Sólo el equipo de Valero Ribera puede sentenciar una semifinal como lo hizo ayer en un Palau Blaugrana completamente entregado. El Racing ganó los cuartos de final en su campo.

127 Multilingual Central Repository ItalianEWN BasqueEWNSpanishEWN EnglishEWN Basque Web Corpus Italian Web Corpus English Web Corpus CatalanEWN Spanish Web Corpus Catalan Web Corpus ACQ ACQACQ ACQ UPLOADUPLOAD UPLOADUPLOAD PORT PORT PORT PORT WSD WSD WSD WSDMeaningArquitecture