Better exploiting BERT for few-shot event detection - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Better exploiting BERT for few-shot event detection

Résumé

Recent approaches for event detection rely on deep supervised learning, which requires large annotated corpora. Few-shot learning approaches, such as the meta-learning paradigm, can be used to address this issue. We focus in this paper on the use of prototypical networks with a BERT encoder for event detection. More specifically, we optimize the use of the information contained in the different layers of a pre-trained BERT model and show that simple strategies for combining BERT layers can outperform the current state-of-the-art for this task.
Fichier principal
Vignette du fichier
tuo-nldb_author-22.pdf (1.22 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

cea-04363098 , version 1 (24-12-2023)

Identifiants

Citer

Aboubacar Tuo, Romaric Besancon, Olivier Ferret, Julien Tourille. Better exploiting BERT for few-shot event detection. 27th International Conference on Applications of Natural Language to Information Systems (NLDB 2022), Jun 2022, Valencia (Espagne), Spain. pp.291-298, ⟨10.1007/978-3-031-08473-7_26⟩. ⟨cea-04363098⟩
18 Consultations
30 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More