POSTECH Submission on Duolingo Shared Task

Junsu Park, Hongseok Kwon, Jong-Hyeok Lee


Abstract
In this paper, we propose a transfer learning based simultaneous translation model by extending BART. We pre-trained BART with Korean Wikipedia and a Korean news dataset, and fine-tuned with an additional web-crawled parallel corpus and the 2020 Duolingo official training dataset. In our experiments on the 2020 Duolingo test dataset, our submission achieves 0.312 in weighted macro F1 score, and ranks second among the submitted En-Ko systems.
Anthology ID:
2020.ngt-1.16
Volume:
Proceedings of the Fourth Workshop on Neural Generation and Translation
Month:
July
Year:
2020
Address:
Online
Venues:
ACL | NGT | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
139–143
Language:
URL:
https://aclanthology.org/2020.ngt-1.16
DOI:
10.18653/v1/2020.ngt-1.16
Bibkey:
Cite (ACL):
Junsu Park, Hongseok Kwon, and Jong-Hyeok Lee. 2020. POSTECH Submission on Duolingo Shared Task. In Proceedings of the Fourth Workshop on Neural Generation and Translation, pages 139–143, Online. Association for Computational Linguistics.
Cite (Informal):
POSTECH Submission on Duolingo Shared Task (Park et al., NGT 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.ngt-1.16.pdf
Video:
 http://slideslive.com/38929830
Data
Duolingo STAPLE Shared Task