Lexical Disambiguation in LTAG using Left Context

Claire Gardent (1) , Yannick Parmentier (2) , Guy Perrier (3) , Sylvain Schmitz (4)

1 SYNALP - Natural Language Processing : representations, inference and semantics
2 LIFO - Laboratoire d'Informatique Fondamentale d'Orléans
3 SEMAGRAMME - Semantic Analysis of Natural Language
4 LSV - Laboratoire Spécification et Vérification [Cachan]

In this paper, we present an optimization for parsing with Lexicalized Tree-Adjoining Grammar which takes inspiration from previous work on polarity based grammar abstraction (Bonfante et al., 2004). We illustrate the impact of this optimization on two benchmarks and we relate our approach to the more general optimization framework proposed for Interaction Grammars by (Bonfante et al., 2009) and (Morey et al., 2011).

Mots clés: parsing; lexical disambiguation; lexicalized tree adjoining grammar; supertagging

Domaines: Informatique et langage [cs.CL]

Claire Gardent, Yannick Parmentier, Guy Perrier, Sylvain Schmitz. Lexical Disambiguation in LTAG using Left Context. Zygmunt Vetulani; Joseph Mariani. Human Language Technology. Challenges for Computer Science and Linguistics. 5th Language and Technology Conference, LTC 2011, Poznan, Poland, November 25-27, 2011, Revised Selected Papers, 8387, Springer, pp.67-79, 2014, Lecture Notes in Computer Science (LNCS) series / Lecture Notes in Artificial Intelligence (LNAI) subseries, ISBN 978-3-319-08957-7. ⟨10.1007/978-3-319-08958-4_6⟩. ⟨hal-00921246⟩