1 |
FlauBERT: Unsupervised Language Model Pre-training for French
|
|
Le, Hang; Vial, Loïc; Frej, Jibril; Segonne, Vincent; Coavoux, Maximin; Lecouteux, Benjamin; Allauzen, Alexandre; Crabbe, Benoit; Besacier, Laurent; Schwab, Didier
|
|
In: Proceedings of the 12th Language Resources and Evaluation Conference ; LREC ; https://hal.archives-ouvertes.fr/hal-02890258 ; LREC, 2020, Marseille, France (2020)
|
|
Abstract:
International audience ; Language models have become a key step to achieve state-of-the art results in many different Natural Language Processing (NLP) tasks. Leveraging the huge amount of unlabeled texts nowadays available, they provide an efficient way to pre-train continuous word representations that can be fine-tuned for a downstream task, along with their contextualization at the sentence level. This has been widely demonstrated for English using contextualized representations (Dai and Le, 2015; Peters et al., 2018; Howard and Ruder, 2018; Radford et al., 2018; Devlin et al., 2019; Yang et al., 2019b). In this paper, we introduce and share FlauBERT, a model learned on a very large and heterogeneous French corpus. Models of different sizes are trained using the new CNRS (French National Centre for Scientific Research) Jean Zay supercomputer. We apply our French language models to diverse NLP tasks (text classification, paraphrasing, natural language inference, parsing, word sense disambiguation) and show that most of the time they outperform other pre-training approaches. Different versions of FlauBERT as well as a unified evaluation protocol for the downstream tasks, called FLUE (French Language Understanding Evaluation), are shared to the research community for further reproducible experiments in French NLP.
|
|
Keyword:
[INFO.INFO-CL]Computer Science [cs]/Computation and Language [cs.CL]; BERT; FlauBERT; FLUE; French; language model; natural language inference; NLP benchmark; paraphrase; parsing; pre-training; text classification; Transformer; word sense disambiguation
|
|
URL: https://hal.archives-ouvertes.fr/hal-02890258/file/Flaubert.pdf https://hal.archives-ouvertes.fr/hal-02890258 https://hal.archives-ouvertes.fr/hal-02890258/document
|
|
BASE
|
|
Hide details
|
|
2 |
FlauBERT : Unsupervised Language Model Pre-training for French ; FlauBERT : des modèles de langue contextualisés pré-entraînés pour le français
|
|
|
|
In: Actes de la 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles ; 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles ; https://hal.archives-ouvertes.fr/hal-02784776 ; 6e conférence conjointe Journées d'Études sur la Parole (JEP, 33e édition), Traitement Automatique des Langues Naturelles (TALN, 27e édition), Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues (RÉCITAL, 22e édition). Volume 2 : Traitement Automatique des Langues Naturelles, Jun 2020, Nancy, France. pp.268-278 (2020)
|
|
BASE
|
|
Show details
|
|
4 |
The STEM-ECR Dataset: Grounding Scientific Entity References in STEM Scholarly Content to Authoritative Encyclopedic and Lexicographic Sources ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Общий взгляд на криминалистическое следоведение ... : GENERAL VIEW ON FORENSIC INVESTIGATION ...
|
|
|
|
BASE
|
|
Show details
|
|
6 |
A novel framework for biomedical entity sense induction
|
|
|
|
In: ISSN: 1532-0464 ; EISSN: 1532-0480 ; Journal of Biomedical Informatics ; https://hal-lirmm.ccsd.cnrs.fr/lirmm-01851988 ; Journal of Biomedical Informatics, Elsevier, 2018, 84, pp.31-41. ⟨10.1016/j.jbi.2018.06.007⟩ (2018)
|
|
BASE
|
|
Show details
|
|
7 |
Cross-domain polarity classification using a knowledge-enhanced meta-classifier
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Senso Comune as a Knowledge Base of Italian language: The Resource and its Development
|
|
|
|
In: Proceedings of the First Italian Conference on Computational Linguistics CLiC-it 2014 ; First Italian Conference on Computational Linguistics - CLiC-it 2014 ; https://hal.archives-ouvertes.fr/hal-01134621 ; First Italian Conference on Computational Linguistics - CLiC-it 2014, Dec 2014, Pisa, Italy. pp. 93-97 (2014)
|
|
BASE
|
|
Show details
|
|
9 |
Applying Lexical Semantics to Improve Text Classification
|
|
|
|
In: http://web2py.iiit.ac.in/publications/default/download/inproceedings.pdf.9ecb6867-0fb0-48a5-8020-0310468d3275.pdf (2008)
|
|
BASE
|
|
Show details
|
|
10 |
Hybrid Methods for Acquisition of Lexical Information: the Case for Verbs
|
|
|
|
In: http://rave.ohiolink.edu/etdc/view?acc_num=osu1228259857 (2008)
|
|
BASE
|
|
Show details
|
|
11 |
Desambiguación léxica mediante marcas de especificidad
|
|
Montoyo, Andres. - : Sociedad Española para el Procesamiento del Lenguaje Natural, 2002
|
|
BASE
|
|
Show details
|
|
12 |
Proceedings of the Twenty-Fourth Conference on Computational Linguistics and Speech Processing (ROCLING 2012) Associating Collocations with WordNet Senses Using Hybrid Models �
|
|
|
|
In: http://wing.comp.nus.edu.sg/~antho/O/O12/O12-1006.pdf
|
|
BASE
|
|
Show details
|
|
13 |
Ontology-Supported Text Classification Based on Cross-Lingual Word Sense Disambiguation
|
|
|
|
In: http://www.racai.ro/~tufis/papers/Tufis-Koeva-wilf2007.pdf
|
|
BASE
|
|
Show details
|
|
14 |
Many tasks in Natural Language Processing – Word
|
|
|
|
In: http://www.plwordnet.pwr.wroc.pl/main/content/files/publications/ltc07-125-piasecki.pdf
|
|
BASE
|
|
Show details
|
|
|
|