Abstract
Machine translation is generally understood as generating one target text from an input source document. In this paper, we consider a stronger requirement: to jointly generate two texts so that each output side effectively depends on the other. As we discuss, such a device serves several practical purposes, from multi-target machine translation to the generation of controlled variations of the target text. We present an analysis of possible implementations of dual decoding, and experiment with four applications. Viewing the problem from multiple angles allows us to better highlight the challenges of dual decoding and to also thoroughly analyze the benefits of generating matched, rather than independent, translations.- Anthology ID:
- 2021.emnlp-main.671
- Volume:
- Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
- Month:
- November
- Year:
- 2021
- Address:
- Online and Punta Cana, Dominican Republic
- Editors:
- Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
- Venue:
- EMNLP
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 8533–8546
- Language:
- URL:
- https://aclanthology.org/2021.emnlp-main.671
- DOI:
- 10.18653/v1/2021.emnlp-main.671
- Cite (ACL):
- Jitao Xu and François Yvon. 2021. One Source, Two Targets: Challenges and Rewards of Dual Decoding. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 8533–8546, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
- Cite (Informal):
- One Source, Two Targets: Challenges and Rewards of Dual Decoding (Xu & Yvon, EMNLP 2021)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-2/2021.emnlp-main.671.pdf
- Code
- jitao-xu/dual-decoding