Simple K-star Categorial Dependency Grammars and their Inference

Denis Béchet (1) , Annie Foret (2, 3, 4)

1 LINA - Laboratoire d'Informatique de Nantes Atlantique
2 LIS - Logical Information Systems
3 UR - Université de Rennes
4 IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires

We propose a novel subclass in the family of Categorial Dependency Grammars (CDG), based on a syntactic criterion on categorial types associated to words in the lexicon and study its learnability. This proposal relies on a linguistic principle and relates to a former non-constructive condition on iterated dependencies. We show that the projective CDG in this subclass are incrementally learnable in the limit from dependency structures. In contrast to previous proposals, our criterion is both syntactic and does not impose a (rigidity) bound on the number of categorial types associated to a word.

Mots clés: Categorial grammar; Dependency grammar; Incremental learning; Grammatical inference; Context-free languages; Iterated dependencies; Dependency Treebanks; Computational linguistics

Domaines: Informatique et langage [cs.CL]; Logique en informatique [cs.LO]; Traitement du texte et du document; Linguistique; Recherche d'information [cs.IR]; Intelligence artificielle [cs.AI]; Environnements Informatiques pour l'Apprentissage Humain; Apprentissage [cs.LG]

Denis Béchet, Annie Foret. Simple K-star Categorial Dependency Grammars and their Inference. The 13th International Conference on Grammatical Inference (ICGI), Oct 2016, Delft, Netherlands. ⟨hal-01363393⟩