@inproceedings{jean-cho-2020-log,
    title = "Log-Linear Reformulation of the Noisy Channel Model for Document-Level Neural Machine Translation",
    author = "Jean, S{\'e}bastien  and
      Cho, Kyunghyun",
    editor = "Agrawal, Priyanka  and
      Kozareva, Zornitsa  and
      Kreutzer, Julia  and
      Lampouras, Gerasimos  and
      Martins, Andr{\'e}  and
      Ravi, Sujith  and
      Vlachos, Andreas",
    booktitle = "Proceedings of the Fourth Workshop on Structured Prediction for NLP",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2020.spnlp-1.11/",
    doi = "10.18653/v1/2020.spnlp-1.11",
    pages = "95--101",
    abstract = "We seek to maximally use various data sources, such as parallel and monolingual data, to build an effective and efficient document-level translation system. In particular, we start by considering a noisy channel approach (CITATION) that combines a target-to-source translation model and a language model. By applying Bayes' rule strategically, we reformulate this approach as a log-linear combination of translation, sentence-level and document-level language model probabilities. In addition to using static coefficients for each term, this formulation alternatively allows for the learning of dynamic per-token weights to more finely control the impact of the language models. Using both static or dynamic coefficients leads to improvements over a context-agnostic baseline and a context-aware concatenation model."
}