Étiquetage multilingue en parties du discours avec MElt

Benoît Sagot 1
1 ALPAGE - Analyse Linguistique Profonde à Grande Echelle ; Large-scale deep linguistic processing
Inria Paris-Rocquencourt, UPD7 - Université Paris Diderot - Paris 7
Abstract : We describe recent evolutions of MElt, a discriminative part-of-speech tagging system. MElt is targeted at the optimal exploitation of information provided by external lexicons for improving its performance over models trained solely on annotated corpora. We have trained MElt on more than 40 datasets covering over 30 languages. Compared with the state-of-the-art system MarMoT, MElt's results are slightly worse on average when no external lexicon is used, but slightly better when such resources are available, resulting in state-of-the-art taggers for a number of languages.
Document type :
Conference papers
Complete list of metadatas

Cited literature [13 references]  Display  Hide  Download

https://hal.inria.fr/hal-01352243
Contributor : Benoît Sagot <>
Submitted on : Saturday, August 6, 2016 - 3:46:48 PM
Last modification on : Friday, January 4, 2019 - 5:33:24 PM
Long-term archiving on : Tuesday, November 8, 2016 - 6:19:00 PM

File

taln16melt.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01352243, version 1

Collections

Citation

Benoît Sagot. Étiquetage multilingue en parties du discours avec MElt. 23ème Conférence sur le Traitement Automatique des Langues Naturelles, Jul 2016, Paris, France. ⟨hal-01352243⟩

Share

Metrics

Record views

344

Files downloads

215