Orthogonality regularizer for question answering

Abstract : Learning embeddings of words and knowledge base elements is a promising approach for open domain question answering. Based on the remark that relations and entities are distinct object types lying in the same embedding space, we analyze the benefit of adding a regularizer favoring the embeddings of entities to be orthogonal to those of relations. The main motivation comes from the observation that modifying the embeddings using prior knowledge often helps performance. The experiments show that incorporating the regularizer yields better results on a challenging question answering benchmark.
Type de document :
Communication dans un congrès
*SEM 2016,. The Fifth Joint Conference on Lexical and Computational Semantics, Aug 2016, Berlin, Germany. pp.142 - 147, 2016, 〈10.18653/v1/S16-2019〉
Liste complète des métadonnées

Littérature citée [20 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01623819
Contributeur : Claire Gardent <>
Soumis le : mercredi 25 octobre 2017 - 17:02:48
Dernière modification le : mardi 24 avril 2018 - 13:30:47
Document(s) archivé(s) le : vendredi 26 janvier 2018 - 14:30:32

Fichier

orthogonality.pdf
Fichiers éditeurs autorisés sur une archive ouverte

Identifiants

Collections

Citation

Chunyang Xiao, Guillaume Bouchard, Marc Dymetman, Claire Gardent. Orthogonality regularizer for question answering. *SEM 2016,. The Fifth Joint Conference on Lexical and Computational Semantics, Aug 2016, Berlin, Germany. pp.142 - 147, 2016, 〈10.18653/v1/S16-2019〉. 〈hal-01623819〉

Partager

Métriques

Consultations de la notice

212

Téléchargements de fichiers

13