Abstract
This paper describes the Xiaomi’s submissions to the IWSLT20 shared open domain translation task for Chinese<->Japanese language pair. We explore different model ensembling strategies based on recent Transformer variants. We also further strengthen our systems via some effective techniques, such as data filtering, data selection, tagged back translation, domain adaptation, knowledge distillation, and re-ranking. Our resulting Chinese->Japanese primary system ranked second in terms of character-level BLEU score among all submissions. Our resulting Japanese->Chinese primary system also achieved a competitive performance.- Anthology ID:
- 2020.iwslt-1.18
- Volume:
- Proceedings of the 17th International Conference on Spoken Language Translation
- Month:
- July
- Year:
- 2020
- Address:
- Online
- Editors:
- Marcello Federico, Alex Waibel, Kevin Knight, Satoshi Nakamura, Hermann Ney, Jan Niehues, Sebastian Stüker, Dekai Wu, Joseph Mariani, Francois Yvon
- Venue:
- IWSLT
- SIG:
- SIGSLT
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 149–157
- Language:
- URL:
- https://aclanthology.org/2020.iwslt-1.18
- DOI:
- 10.18653/v1/2020.iwslt-1.18
- Cite (ACL):
- Yuhui Sun, Mengxue Guo, Xiang Li, Jianwei Cui, and Bin Wang. 2020. Xiaomi’s Submissions for IWSLT 2020 Open Domain Translation Task. In Proceedings of the 17th International Conference on Spoken Language Translation, pages 149–157, Online. Association for Computational Linguistics.
- Cite (Informal):
- Xiaomi’s Submissions for IWSLT 2020 Open Domain Translation Task (Sun et al., IWSLT 2020)
- PDF:
- https://preview.aclanthology.org/proper-vol2-ingestion/2020.iwslt-1.18.pdf