NLP, deep probabilistic and symbolic parsing, discourse modeling and
parsing, NLP-based knowledge acquisition techniques (alpage)

Type de poste : Post-doctorant

Lieu de travail : Rocquencourt

Thème de recherche : Systèmes symboliques

Projet : ALPAGE
 
 
Environnement

ALPAGE is a multidisciplinary team, regrouping researchers both from
computer science and computational linguistics, who works in the field
of Natural Language Processing (NLP).

The first goal of this team is to develop a robust and efficient
parser for French, which will be freely distributed (such a parser
doesn't exist yet for French). Both symbolic and probabilistic methods
are currently studied so as to reach an hybrid parser. Dynamic French
linguistic resources are also developed at the lexical level (lexicon
Lefff : Lexique des Formes Fléchies du Français) and at the
grammatical level, with a special expertise in writing TAG and LFG
grammars.

The second goal is to extend this sentential syntactic parser both to
the semantic level and to the discourse (supra-sentential) level.

Themes of interest are:
-anaphora resolution,
-hierarchical discourse structures based on discourse relations
 (within the SDRT framework)
-temporal and event description (adaptation of Time ML to French).
 
Missions
General themes :
-Word sense disambiguation techniques
-Hybrid (probabilistic and symbolic) approaches for parsing
-Enrichment of a syntactic parser with semantic knowledge (e.g. from
 c-structures to f-structures in LFG, syntax-semantic interface with
 synchronous TAG)
-Computational lexical semantics
-Discourse processing
 
Compétences et Profil

Computer scientists specialist in NLP or computational linguists.
Knowledge in French is welcome but not mandatory.
 
Informations complémentaires

http://alpage.inria.fr/stages.en.html

Contact : Laurence.Danlos@inria.fr