The NiuTrans Machine Translation Systems for WMT21

This paper describes NiuTrans neural machine translation systems of the WMT 2021 news translation tasks. We made submissions to 9 language directions, including English$\leftrightarrow$$\{\(Chinese, Japanese, Russian, Icelandic\)\}\( and English\)\rightarrow$Hausa tasks. Our primary systems are buil...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2021-09
Main Authors: Zhou, Shuhan, Zhou, Tao, Wei, Binghao, Luo, Yingfeng, Mu, Yongyu, Zhou, Zefan, Wang, Chenglong, Zhou, Xuanjun, Lv, Chuanhao, Yi, Jing, Wang, Laohu, Zhang, Jingnan, Huang, Canan, Yan, Zhongxiang, Hu, Chi, Li, Bei, Tong, Xiao, Zhu, Jingbo
Format: Article
Language:English
Subjects:
Distillation
English language
Iterative methods
Machine translation
System effectiveness
Transformers
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper describes NiuTrans neural machine translation systems of the WMT 2021 news translation tasks. We made submissions to 9 language directions, including English$\leftrightarrow$$\{\(Chinese, Japanese, Russian, Icelandic\)\}\( and English\)\rightarrow$Hausa tasks. Our primary systems are built on several effective variants of Transformer, e.g., Transformer-DLCL, ODE-Transformer. We also utilize back-translation, knowledge distillation, post-ensemble, and iterative fine-tuning techniques to enhance the model performance further.
ISSN:2331-8422