1 Rogelio Nazar & Maarten Janssen IULA, Universitat Pompeu Fabra, Barcelona

2  Dictionaries good source for information  Long tradition of taxonomy extraction  Calzolari (1977), Amsler (1981), Chodorow et al (1985), Fox et al. (1988), Alshawi (1989), Boguraev (1991), Barrière & Popowich (1996), Chang (1998), Renau & Battaner (2008)  Exploiting Machine Readable Dictionaries  Parsing definitional phrases  Pattern extraction, Shallow parsing  Full treatment of a single dictionary

3  There is a lot of information available  Hand crafted, high-qualify resources  Combining yields new data  Taxonomy from multiple dictionaries  Language-independent shallow method  Combining definitions of the same word  Various dictionaries, online versions  DRAE, DGLE, Clave, DEM  Frequency Based

4  Dictionaries differ ◦ Different lexicon and definitions ◦ Even if only for legal reasons  Hyperonym should be the same ◦ A cat is an animal ◦ Unless there is uncertainty in the hyperonym  Most dictionaries should use same genus ◦ Statistically relevant

5 3x ablandabrevas persona 2x com. inútil 1x substantivo común fig.

6  Directly from harvested text ◦ With begin/end tags  No textual analysis  More than definitions ◦ Examples, multiple senses, etc.  Sense matching impossible ◦ Entries unsystematic ◦ Dictionaries do not match in senses

7  Minimum number of dictionaries  Raw frequency count ◦ Hyperonym tends to be repeated  Candidates have to be words ◦ Of the same word-class  Use of a stop-list ◦ Dictionary generated ◦ Words that occur in more than 10% entries

8 # deconstrucción (3 dictionaries) teoría21 EWN: 0.desconstrucción; 0.deconstrucción; 1.teoría filosófica; 1.doctrina filosófica; 2.filosofía; 3.creencia; 4.contenido mental; 5.conocimiento; 5.cognición; 6.rasgo psicológico; # descubrimiento (5 dictionaries) acción33 cosa35 efecto2- EWN: 0.descubrimiento; 1.logro; 1.presentación; 1.revelación; 2.realización; 2.información; 2.exposición; 3.acción; 3.hecho; 3.acto de habla; 3.comunicación visual; 4.acto; 4.actividad humana; 4.comunicación; 5.relación social; 6.relación; 7.abstracción; # cumbia (5 dictionaries) danza2- EWN: 0.cumbiamba; 0.cumbia; 1.baile regional; 1.danza popular; 2.baile social; 3.baile; 4.recreación; 4.diversión; 5.actividad; 6.acto; 6.actividad humana; # asta (5 dictionaries) mar6- lanza6- media5- toro5- cuerno5- bandera4- EWN: 0.cuerno; 0.asta; 1.tomadero; 1.materia animal; 1.cogedero; 1.bastón; 1.agarradera; 1.asimiento; 1.asidero; 1.asa; 2.materia; 2.apéndice; 2.vara; 2.palo; 3.porción; 3.sustancia; 3.parte; 3.herramienta; 4.utillaje; 5.artefacto; 6.objeto físico; 6.cosa; 6.objeto; 6.objeto inanimado; 7.competente; 7.respirar; 7.capaz; 7.entidad;

9  WordNet (still) best available taxonomy ◦ Not the best resource for evaluation  Automatic Verification ◦ 100 Random nouns ◦ Best 5 hyperonymy candidates ◦ Match when candidate in chain  Only about 50% accurracy

10

11  WordNet ◦ Many intermediate/artificial levels ◦ Compulsory hyperonym ◦ Contains proper names  Dictonaries ◦ More word-senses ◦ Alternative definitions (synonymy, paraphrasis, …)  Differences ◦ Different choice of hyperonym ◦ Different lexicon

12

13

14  Question?