HAL: inria-00617068, version 1

See detailed view

BibTeX,EndNote,... BibTeX - EndNote - TEI - RefWorks -

TALN'2011 - Traitement Automatique des Langues Naturelles, Montpellier : France (2011)

Coopération de méthodes statistiques et symboliques pour l'adaptation non-supervisée d'un système d'étiquetage en entités nommées

Frédéric Béchet 1, Benoît Sagot 2, Rosa Stern 2, 3

(2011)

Named entity recognition and typing is achieved both by symbolic and probabilistic systems. We report on an experiment for making the rule-based system NP, a high-precision system developed on AFP news corpora and relies on the Aleda named entity database, interact with LIANE, a high-recall probabilistic system trained on oral transcriptions from the ESTER corpus. We show that a probabilistic system such as LIANE can be adapted to a new type of corpus in a non-supervized way thanks to large-scale corpora automatically annotated by NP. This adaptation does not require any additional manual anotation and illustrates the complementarity between numeric and symbolic techniques for tackling linguistic tasks.

1:	Laboratoire d'informatique Fondamentale de Marseille (LIF)
	CNRS : UMR6166 – Université de la Méditerranée - Aix-Marseille II – Université de Provence - Aix-Marseille I
2:	ALPAGE (INRIA Rocquencourt)
	INRIA – Université Paris VII - Paris Diderot
3:	Medialab AFP (Medialab AFP)
	Agence France-Presse

Domain

:

Computer Science/Computation and Language

Keywords: Named entity recognition – domain adaptation – cooperation between probabilistic and symbolic approaches

inria-00617068, version 1

http://hal.inria.fr/inria-00617068

oai:hal.inria.fr:inria-00617068

From: Benoît Sagot <>

Submitted on: Thursday, 25 August 2011 22:40:32

Updated on: Monday, 5 September 2011 14:49:55