DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation

@article{yayue2026,
  title = {DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation},
  author = {Deng, Yayue and Wang, Dingdong and Hu, Yuxuan and Li, Jinyu and Liu, Yanqing and Wang, Yuanyuan and Chen, Weidong and Meng, Helen M. and Liu, Shujie and Wu, Xixin},
  journal = {Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing},
  year = {2026},
  month = {Jan},
  doi = {10.18653/2026.emnlp-main.1525},
  url = {https://doi.org/10.18653/2026.emnlp-main.1525},
}

Deng, Yayue et al. (2026). DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation. Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing. https://doi.org/10.18653/2026.emnlp-main.1525

Deng, Yayue, et al.. "DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation." Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, 2026, https://doi.org/10.18653/2026.emnlp-main.1525.

Deng, Yayue, et al.. "DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation." Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (2026). https://doi.org/10.18653/2026.emnlp-main.1525

Deng, Yayue et al., 2026. DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation. Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing. Available at: <https://doi.org/10.18653/2026.emnlp-main.1525>.

TY  - JOUR
TI  - DuraS2ST: Chain-of-Thought and Reinforcement Learning for Duration-Aligned Speech-to-Speech Translation
AU  - Deng, Yayue
AU  - Wang, Dingdong
AU  - Hu, Yuxuan
AU  - Li, Jinyu
AU  - Liu, Yanqing
AU  - Wang, Yuanyuan
AU  - Chen, Weidong
AU  - Meng, Helen M.
AU  - Liu, Shujie
AU  - Wu, Xixin
PY  - 2026
JO  - Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
DO  - 10.18653/2026.emnlp-main.1525
UR  - https://doi.org/10.18653/2026.emnlp-main.1525
ER  -