When is multitask learning effective? Semantic sequence prediction under varying data conditions - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2017

When is multitask learning effective? Semantic sequence prediction under varying data conditions

Résumé

Multitask learning has been applied successfully to a range of tasks, mostly mor-phosyntactic. However, little is known on when MTL works and whether there are data characteristics that help to determine its success. In this paper we evaluate a range of semantic sequence labeling tasks in a MTL setup. We examine different auxiliary tasks, amongst which a novel setup, and correlate their impact to data-dependent conditions. Our results show that MTL is not always effective, significant improvements are obtained only for 1 out of 5 tasks. When successful, auxiliary tasks with compact and more uniform label distributions are preferable.
Fichier principal
Vignette du fichier
eacl2017.pdf (306.92 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01677427 , version 1 (08-01-2018)

Identifiants

  • HAL Id : hal-01677427 , version 1

Citer

Hector Martinez Alonso, Barbara Plank. When is multitask learning effective? Semantic sequence prediction under varying data conditions. EACL 2017 - 15th Conference of the European Chapter of the Association for Computational Linguistics, Apr 2017, Valencia, Spain. pp.1-10. ⟨hal-01677427⟩
264 Consultations
136 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More