Jaouad.Mousser(at)uni-konstanz.de
University of Konstanz
Department of Linguistics
Arabic VerbNet Abstract
Levin's class based approache to verb semantic was very productive in classifying a large number of verbs in English. It was exploited to build a large coverage verb lexicon providing many levels of information (syntax, semantic etc.) (Verbnet). Verbnet was successfully used in many NLP applications. Levin's classes were tentatively used for other languages like German, Korean, Bangala etc., but no comprehensive large coverage verb classification was performed for these languages. Arabic Verbnet is one of the first verb lexica, which classify the most used verbs of Modern Standard Arabic (MSA) on the basis Levin's verb classes using the development procedure of Schuler. The verb lexicon in its current state has 334 classes which contain 7672 verbs and 1393 frames providing information about verb root, the deverbal form, the participle, thematic roles, subcategorization frames and syntactic and semantic descriptions of verbs.
Kreuzlingen, 12-05 2011: ARACON, A new, powerfull verb conjugator for Arabic
I just finished programming a verb conjugator for Arabic as part of a morphological analyser and generator. It is a java API implementing a root-pattern based algorithm for analysing and generating verbs forms.
Oxford, UK January 12–01 2011: International Conference on Computational Semantics
Valetta, Malta Mai 19-21 2010 LREC Conference
The conference startet for me on the 17th Mai with a workshop about verb "ontologies". I came to know about my presetation on the 16 during the flight to Malta, but It was not as bad as I exprected.
The poster Session on 20th Mai.