Abstract
Simultaneous translation is a task that requires starting translation before the speaker has finished speaking, so we face a trade-off between latency and accuracy. In this work, we focus on prefix-to-prefix translation and propose a method to extract alignment between bilingual prefix pairs. We use the alignment to segment a streaming input and fine-tune a translation model. The proposed method demonstrated higher BLEU than those of baselines in low latency ranges in our experiments on the IWSLT simultaneous translation benchmark.- Anthology ID:
- 2022.iwslt-1.3
- Volume:
- Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022)
- Month:
- May
- Year:
- 2022
- Address:
- Dublin, Ireland (in-person and online)
- Editors:
- Elizabeth Salesky, Marcello Federico, Marta Costa-jussà
- Venue:
- IWSLT
- SIG:
- SIGSLT
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 22–31
- Language:
- URL:
- https://aclanthology.org/2022.iwslt-1.3
- DOI:
- 10.18653/v1/2022.iwslt-1.3
- Cite (ACL):
- Yasumasa Kano, Katsuhito Sudoh, and Satoshi Nakamura. 2022. Simultaneous Neural Machine Translation with Prefix Alignment. In Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022), pages 22–31, Dublin, Ireland (in-person and online). Association for Computational Linguistics.
- Cite (Informal):
- Simultaneous Neural Machine Translation with Prefix Alignment (Kano et al., IWSLT 2022)
- PDF:
- https://preview.aclanthology.org/improve-issue-templates/2022.iwslt-1.3.pdf