Idiap NMT System for WAT 2019 Multimodal Translation Task

Shantipriya Parida, Ondřej Bojar, Petr Motlicek


Abstract
This paper describes the Idiap submission to WAT 2019 for the English-Hindi Multi-Modal Translation Task. We have used the state-of-the-art Transformer model and utilized the IITB English-Hindi parallel corpus as an additional data source. Among the different tracks of the multi-modal task, we have participated in the “Text-Only” track for the evaluation and challenge test sets. Our submission tops in its track among the competitors in terms of both automatic and manual evaluation. Based on automatic scores, our text-only submission also outperforms systems that consider visual information in the “multi-modal translation” task.
Anthology ID:
D19-5223
Volume:
Proceedings of the 6th Workshop on Asian Translation
Month:
November
Year:
2019
Address:
Hong Kong, China
Editors:
Toshiaki Nakazawa, Chenchen Ding, Raj Dabre, Anoop Kunchukuttan, Nobushige Doi, Yusuke Oda, Ondřej Bojar, Shantipriya Parida, Isao Goto, Hidaya Mino
Venue:
WAT
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
175–180
Language:
URL:
https://aclanthology.org/D19-5223
DOI:
10.18653/v1/D19-5223
Bibkey:
Cite (ACL):
Shantipriya Parida, Ondřej Bojar, and Petr Motlicek. 2019. Idiap NMT System for WAT 2019 Multimodal Translation Task. In Proceedings of the 6th Workshop on Asian Translation, pages 175–180, Hong Kong, China. Association for Computational Linguistics.
Cite (Informal):
Idiap NMT System for WAT 2019 Multimodal Translation Task (Parida et al., WAT 2019)
Copy Citation:
PDF:
https://aclanthology.org/D19-5223.pdf