Semi-supervised SRL system with Bayesian inference

Alejandra Lorenzo 1 Christophe Cerisara 1
1 SYNALP - Natural Language Processing : representations, inference and semantics
LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery
Abstract : We propose a new approach to perform semi-supervised training of Semantic Role Labeling models with very few amount of initial labeled data. The proposed approach combines in a novel way supervised and unsupervised training, by forcing the supervised classifier to over-generate potential semantic candidates, and then letting unsupervised inference choose the best ones. Hence, the supervised classifier can be trained on a very small corpus and with coarse-grain features, because its precision does not need to be high: its role is mainly to constrain Bayesian inference to explore only a limited part of the full search space. This approach is evaluated on French and English. In both cases, it achieves very good performance and outperforms a strong supervised baseline when only a small number of annotated sentences is available and even without using any previously trained syntactic parser.
Type de document :
Communication dans un congrès
15th International Conference on Intelligent Text Processing and Computational Linguistics, Apr 2014, Nepal. pp.433, 2014
Liste complète des métadonnées

Littérature citée [20 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01015414
Contributeur : Christophe Cerisara <>
Soumis le : jeudi 26 juin 2014 - 14:24:02
Dernière modification le : mardi 24 avril 2018 - 13:30:41
Document(s) archivé(s) le : vendredi 26 septembre 2014 - 12:11:10

Fichier

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

Identifiants

  • HAL Id : hal-01015414, version 1

Collections

Citation

Alejandra Lorenzo, Christophe Cerisara. Semi-supervised SRL system with Bayesian inference. 15th International Conference on Intelligent Text Processing and Computational Linguistics, Apr 2014, Nepal. pp.433, 2014. 〈hal-01015414〉

Partager

Métriques

Consultations de la notice

337

Téléchargements de fichiers

153