SemTAG : une plate-forme pour le calcul sémantique à partir de Grammaires d'Arbres Adjoints

Abstract : In this thesis, we propose a software architecture (namely SemTAG) allowing for semantic construction with Tree Adjoining Grammars (TAG). SemTAG provides with an environment allowing to build an underspecified semantic representation (Predicate Logic Unplugged of (Bos, 95)) from a TAG lexicon and a sentence. In order to facilitate lexicon management, the SemTAG platforms uses a metagrammar compiler. The goal of the compiler is to automatically produce a TAG from a reduced description. This description corresponds to (a) a hierarchy of tree fragments and (b) combinations of these fragments defined using a control language. Furthermore, each tree produced may be equipped with a syntax / semantic interface following (Gardent and Kallmeyer, 03). The semantic construction is based on the result of the syntactic parsing. Parsing is done by a tabular TAG parser generated automatically from the input TAG by the DyALog system of (De La Clergerie, 05). This parser outputs a derivation forest, which encodes all TAG derivations, and from which the unifications of semantic indices are extracted. This platforms has been evaluated in terms of semantic coverage on the TSNLP test-suite.
Document type :
Theses
Complete list of metadatas

Cited literature [24 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01748165
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 11:27:00 AM
Last modification on : Tuesday, April 24, 2018 - 1:52:08 PM
Long-term archiving on : Friday, September 14, 2018 - 2:22:18 AM

File

SCD_T_2007_0005_PARMENTIER.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01748165, version 1

Collections

Citation

Yannick Parmentier. SemTAG : une plate-forme pour le calcul sémantique à partir de Grammaires d'Arbres Adjoints. Autre [cs.OH]. Université Henri Poincaré - Nancy 1, 2007. Français. ⟨NNT : 2007NAN10005⟩. ⟨tel-01748165⟩

Share

Metrics

Record views

17

Files downloads

26