V. Claveau > Publications > TAL 42(3) Abstract OLSTOLST
Version française

Vincent Claveau, Pascale Sébillot, Pierrette Bouillon, Cécile Fabre,
Acquérir des éléments du lexique génératif : quels résultats et à quels coûts ?,
TAL (traitement automatique des langues), special issue on Semantic Lexicons in NLP applications, Hermès, Vol. 42, No. 3, 2001,
Document (pdf)

Abstract This paper demonstrates the feasibility of automatic acquisition of generative lexicons from corpora through the report of four experiments in machine learning in which various levels of word tagging (categorial and semantic) are handled. The lexical information that is learnt consists of lists of noun-verb couples related by one of the roles of the qualia structure. They provide linguistic knowledge useful in many applications such as information retrieval. We first show that satisfactory results are obtained on the basis of categorial information only, exploited by means of a learning method in the Inductive Logic Programming framework. We further demonstrate that the balance between quality and cost of the learning method is reached by the combination of a categorial tagging and a semantic tagging of words others than nouns.