A Study of Residual Adapters for Multi-Domain Neural Machine Translation - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

A Study of Residual Adapters for Multi-Domain Neural Machine Translation

Minh Quang Pham
Josep-Maria Crego
  • Fonction : Auteur
  • PersonId : 1038144
François Yvon
Jean Senellart
  • Fonction : Auteur
  • PersonId : 1038145

Résumé

Domain adaptation is an old and vexing problem for machine translation systems. The most common and successful approach to supervised adaptation is to fine-tune a baseline system with in-domain parallel data. Standard fine-tuning however modifies all the network parameters, which makes this approach computationally costly and prone to overfitting. A recent, lightweight approach, instead augments a baseline model with supplementary (small) adapter layers, keeping the rest of the model unchanged. This has the additional merit to leave the baseline model intact and adaptable to multiple domains. In this paper, we conduct a thorough analysis of the adapter model in the context of a multidomain machine translation task. We contrast multiple implementations of this idea using two language pairs. Our main conclusions are that residual adapters provide a fast and cheap method for supervised multi-domain adaptation; our two variants prove as effective as the original adapter model and open perspective to also make adapted models more robust to label domain errors.
Fichier principal
Vignette du fichier
2020.wmt-1.72.pdf (417.05 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-03013197 , version 1 (19-11-2020)

Identifiants

  • HAL Id : hal-03013197 , version 1

Citer

Minh Quang Pham, Josep-Maria Crego, François Yvon, Jean Senellart. A Study of Residual Adapters for Multi-Domain Neural Machine Translation. Conference on Machine Translation, Nov 2020, Online, United States. ⟨hal-03013197⟩
255 Consultations
132 Téléchargements

Partager

Gmail Facebook X LinkedIn More