Modèles algorithmiques de l'acquisition de la syntaxe : concepts et méthodes, résultats et problèmes

Denis Bechet 1 Roberto Bonato 2, 3 Alexandre Dikovsky 1 Annie Foret 4 Yannick Le Nir 3, 5 Erwan Moreau 1 Christian Retoré 2, 3 Isabelle Tellier 6, 7
3 SIGNES - Linguistic signs, grammar and meaning: computational logic for natural language
Université Sciences et Technologies - Bordeaux 1, Inria Bordeaux - Sud-Ouest, École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), CNRS - Centre National de la Recherche Scientifique : UMR5800
4 LIS - Logical Information Systems
IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
6 MOSTRARE - Modeling Tree Structures, Machine Learning, and Information Extraction
LIFL - Laboratoire d'Informatique Fondamentale de Lille, Inria Lille - Nord Europe
Abstract : In this paper, we present our recent results on the acquistion of the syntax of natural languages, from the point of view of the theory of grammatical inference. Given a class of possible grammars, the objective is to identify, from a set of positive examples, a grammar in the class which produces the examples. The Gold model formalises the learning process and gives stringent criteria of its success: when does there exist an algorithm producing a target grammar ? what kind of structure should the examples have (strings of words, strings of tagged words, trees) ? From a theoretical point of view, our results establish the learnability or the unlearnability of various classes of categorial grammars. From a practical perspective, these results enable the extraction of syntactic information from real data. Finally, we discuss the interest of this approach for modelling child language acquisition and for automated induction of grammars from corpora.
Complete list of metadatas

Cited literature [30 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-00354043
Contributor : Christian Retoré <>
Submitted on : Monday, November 4, 2013 - 2:36:01 PM
Last modification on : Sunday, April 7, 2019 - 3:00:39 PM
Long-term archiving on : Tuesday, January 3, 2017 - 5:47:58 PM

File

gracq_final_2007_rlv.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00354043, version 1

Citation

Denis Bechet, Roberto Bonato, Alexandre Dikovsky, Annie Foret, Yannick Le Nir, et al.. Modèles algorithmiques de l'acquisition de la syntaxe : concepts et méthodes, résultats et problèmes. Recherches linguistiques de Vincennes, Presses Universitaires de Vincennes, 2007, 36, pp.123--152. ⟨hal-00354043⟩

Share

Metrics

Record views

1595

Files downloads

833