Résumé de thèse. Titre : « APPROCHES QUANTITATIVES DE L’EXTRACTION DE RESSOURCES TRADUCTIONNELLES A PARTIR DE CORPUS PARALLELES ». Traitement Automatique des Langues. Vol. 45 n°3
Abstract
This research work presents the results of a series of experiments devoted to the development of new tools for intertextual textometric exploration of translation corpora. Various methods of textual statistics have been adapted for use in a multilingual context and put into practice for parallel text processing, such as: repeated segments extraction, characteristic elements computation, bi-textual topography, multiple co-occurrences, factorial analysis, automatic classification, etc. Examples of concrete applications illustrate the use of each of these methods in a multilingual context. These examples are accompanied by sample translation resources obtained on quantitative bases from the parallel French/English corpus of the Convention for the Protection of Human Rights. The suggested approach opens up new horizons for automatic exploration of lexical equivalences of translation corpora by a variety of users: translators, foreign language teachers, terminologists, lexicographers, etc.