Domain Adaptation for Named Entity Recognition Using CRFs

Tian Tian (1, 2, 3, 4, 5) , Marco Dinarelli (1, 2, 3, 4) , Isabelle Tellier (1, 2, 3, 4) , Pedro Dias Cardoso (5)

1 Lattice - Lattice - Langues, Textes, Traitements informatiques, Cognition - UMR 8094
2 ENS-PSL - École normale supérieure - Paris
3 USPC - Université Sorbonne Paris Cité
4 PSL - Université Paris Sciences et Lettres
5 Synthesio

In this paper we explain how we created a labelled corpus in English for a Named Entity Recognition (NER) task from multi-source and multi-domain data, for an industrial partner. We explain the specificities of this corpus with examples and describe some baseline experiments. We present some results of domain adaptation on this corpus using a labelled Twitter corpus (Ritter et al., 2011). We tested a semi-supervised method from (Garcia-Fernandez et al., 2014) combined with a supervised domain adaptation approach proposed in (Raymond and Fayolle, 2010) for machine learning experiments with CRFs (Conditional Random Fields). We use the same technique to improve the NER results on the Twitter corpus (Ritter et al., 2011). Our contributions thus consist in an industrial corpus creation and NER performance improvements.

Mots clés: Domain Adaptation; Social Media; CRFs; Machine Learning

Domaines: Informatique; Linguistique

Tian Tian, Marco Dinarelli, Isabelle Tellier, Pedro Dias Cardoso. Domain Adaptation for Named Entity Recognition Using CRFs. LREC 2016, May 2016, Portoroz, Slovenia. ⟨hal-01473397⟩