@inproceedings{dai-yamaguchi-2019-compact,
    title = "Compact and Robust Models for {J}apanese-{E}nglish Character-level Machine Translation",
    author = "Dai, Jinan  and
      Yamaguchi, Kazunori",
    editor = "Nakazawa, Toshiaki  and
      Ding, Chenchen  and
      Dabre, Raj  and
      Kunchukuttan, Anoop  and
      Doi, Nobushige  and
      Oda, Yusuke  and
      Bojar, Ond{\v{r}}ej  and
      Parida, Shantipriya  and
      Goto, Isao  and
      Mino, Hidaya",
    booktitle = "Proceedings of the 6th Workshop on Asian Translation",
    month = nov,
    year = "2019",
    address = "Hong Kong, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest_wac_2008/D19-5202/",
    doi = "10.18653/v1/D19-5202",
    pages = "36--44",
    abstract = "Character-level translation has been proved to be able to achieve preferable translation quality without explicit segmentation, but training a character-level model needs a lot of hardware resources. In this paper, we introduced two character-level translation models which are mid-gated model and multi-attention model for Japanese-English translation. We showed that the mid-gated model achieved the better performance with respect to BLEU scores. We also showed that a relatively narrow beam of width 4 or 5 was sufficient for the mid-gated model. As for unknown words, we showed that the mid-gated model could somehow translate the one containing Katakana by coining out a close word. We also showed that the model managed to produce tolerable results for heavily noised sentences, even though the model was trained with the dataset without noise."
}