We present a framework and its implementation relying on Natural Language Processing methods, which aims at the identification of exercise item candidates from corpora. The hybrid system combining heuristics and machine learning methods includes a number of relevant selection criteria. We focus on two fundamental aspects: linguistic complexity and the dependence of the extracted sentences on th...

RÉSUMÉ ____________________________________________________________________________________________________________ 
ROCme! permet une gestion rationalisée, autonome et dématérialisée de l’enregistrement de corpus oraux. Il dispose notamment d’une interface pour le recueil de métadonnées sur les locuteurs totalement paramétrable via des balises XML. Les locuteurs peuvent gérer les réponses au q...

To produce coherent linear documents, Natural Language Generation systems have traditionally exploited the structuring role of textual discourse markers such as relational and referential phrases. These coherence markers of the traditional notion of text, however, do not work in non-linear documents: a new set of graphical devices is needed together with formation rules to govern their usage, s...

RÉSUMÉ Ce travail présente une nouvelle approche pour injecter des dépendances profondes (sujet des verbes à contrôle, partage du sujet en cas d’ellipses, . . .) dans un corpus arboré présentant un schéma d’annotation surfacique et projectif. Nous nous appuyons sur un système de réécriture de graphes utilisant des techniques de programmation par contraintes pour produire des règles génériques q...

In natural language generation using symbolic grammars, state-of-the-art realisation rankers use statistical models incorporating both language model and structural features. The rankers depend on multiple structures produced by the particular large-scale symbolic grammars to rank the output; for languages with smaller resources and in-development grammars, we look at the feasibility of an alte...

Welsh noun phrases have had much less attention than clauses, and there are unresolved issues about the nature of possessors, attributive adjectives, definite article agreement clitics. There is evidence, especially from agreement, that possessors complements, evidence adjectives adjoined to a preceding [LEX+] nominal constituent, clitics specifiers. The last these positions makes it fairly sim...