Source code for our EMNLP 2022 main conference paper "Information-Transport-based Policy for Simultaneous Translation"

fig

Our method is implemented based on the open-source toolkit Fairseq.

• Python version = 3.8

• PyTorch version = 1.7.1

• Install requirements:

  git clone https://github.com/ictnlp/ITST.git
  cd ITST
  pip install -r requirements.txt
  pip install --editable ./

Information-transport-based simultaneous translation (ITST) achieves good results on both text-to-text simultaneous translation and speech-to-text simultaneous translation (a.k.a., streaming speech translation). Detailed introductions refer to:

• Text-to-text simultaneous translation
• Speech-to-text simultaneous translation with fixed pre-decision
• Speech-to-text simultaneous translation with flexible pre-decision

All example shell scripts refer to shell_scripts/.

If you have any questions, feel free to contact me with: zhangshaolei20z@ict.ac.cn.

If this repository is useful for you, please cite as:

@inproceedings{ITST,
 title = "Information-Transport-based Policy for Simultaneous Translation",
 author = "Zhang, Shaolei and
 Feng, Yang",
 booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
 month = dec,
 year = "2022",
 address = "Online and Abu Dhabi",
 publisher = "Association for Computational Linguistics",
 url="https://arxiv.org/pdf/2210.12357.pdf",
}