Accéder directement au contenu Accéder directement à la navigation
Article dans une revue

Evaluation of Embeddings in Medication Domain for Spanish Language Using Joint Natural Language Understanding

Surya Roca 1 Sophie Rosset 2 José García 1 Álvaro Alesanco 1
2 ILES - Information, Langue Ecrite et Signée
LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur
Abstract : Word embeddings have been widely used in Natural Language Processing as the input to neural networks. Such word embeddings can help in the understanding of the final objective and the keywords in a sentence. As such, in this work, we study the impact of different word embeddings trained with general and specific corpora using Joint Natural Language Understanding in a Spanish medication domain. We generate data using templates for training the model. The model is used for intent detection and slot-filling. We compare word2vec and fastText as word embeddings and ELMo and BERT as language models. We use three different corpora to train the embeddings: the training data generated for this scenario, the Spanish Wikipedia as general domain and the Spanish drug database as specialized data. The best result was obtained with word2vec continuous bag of words model learned with Spanish Wikipedia, obtaining a 71.77% F1-score for intent detection, an intent accuracy of 69.37% and a 74.36% F1-score for slot-filling.
Type de document :
Article dans une revue
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-03294349
Contributeur : Bénédicte Daly Connectez-vous pour contacter le contributeur
Soumis le : mardi 27 juillet 2021 - 12:40:26
Dernière modification le : jeudi 29 juillet 2021 - 04:00:58
Archivage à long terme le : : jeudi 28 octobre 2021 - 18:06:54

Fichier

Surya_EMBEC.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Surya Roca, Sophie Rosset, José García, Álvaro Alesanco. Evaluation of Embeddings in Medication Domain for Spanish Language Using Joint Natural Language Understanding. IFMBE Proceedings, 2020, 8th European Medical and Biological Engineering Conference, 80, pp.510 - 517. ⟨10.1007/978-3-030-64610-3_58⟩. ⟨hal-03294349⟩

Partager

Métriques

Consultations de la notice

49

Téléchargements de fichiers

39