Tree Representations in Probabilistic Models for Extended Named Entities Detection

Marco Dinarelli (1, 2, 3, 4) , Sophie Rosset (5)

1 ENS-PSL - École normale supérieure - Paris
2 USPC - Université Sorbonne Paris Cité
3 PSL - Université Paris Sciences et Lettres
4 Lattice - Lattice - Langues, Textes, Traitements informatiques, Cognition - UMR 8094
5 LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur

In this paper we deal with Named Entity Recognition (NER) on transcriptions of French broadcast data. Two aspects make the task more difficult with respect to previous NER tasks: i) named entities annotated used in this work have a tree structure, thus the task cannot be tackled as a sequence labelling task; ii) the data used are more noisy than data used for previous NER tasks. We approach the task in two steps, involving Conditional Random Fields and Probabilistic Context-Free Grammars, integrated in a single parsing algorithm. We analyse the effect of using several tree representations. Our system outperforms the best system of the evaluation campaign by a significant margin.

Domaines: Informatique et langage [cs.CL]; Apprentissage [cs.LG]

Marco Dinarelli, Sophie Rosset. Tree Representations in Probabilistic Models for Extended Named Entities Detection. European Chapter of the Association for Computational Linguistics, Apr 2012, Avignon, France. ⟨hal-01490007⟩