Fully Character-Level Neural Machine Translation without Explicit Segmentation

@article{Lee2017FullyCN,
  title={Fully Character-Level Neural Machine Translation without Explicit Segmentation},
  author={Jason Lee and Kyunghyun Cho and Thomas Hofmann},
  journal={Transactions of the Association for Computational Linguistics},
  year={2017},
  volume={5},
  pages={365-378}
}
  • Jason Lee, Kyunghyun Cho, Thomas Hofmann
  • Published 2017
  • Computer Science
  • Transactions of the Association for Computational Linguistics
  • Most existing machine translation systems operate at the level of words, relying on explicit segmentation to extract tokens. We introduce a neural machine translation (NMT) model that maps a source character sequence to a target character sequence without any segmentation. We employ a character-level convolutional network with max-pooling at the encoder to reduce the length of source representation, allowing the model to be trained at a speed comparable to subword-level models while capturing… CONTINUE READING

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 256 CITATIONS

    Deep Character-Level Neural Machine Translation By Learning Morphology

    VIEW 4 EXCERPTS
    CITES BACKGROUND & METHODS
    HIGHLY INFLUENCED

    Character-level Transformer-based Neural Machine Translation

    VIEW 17 EXCERPTS
    CITES RESULTS, METHODS & BACKGROUND
    HIGHLY INFLUENCED

    Neural Machine Translation with Byte-Level Subwords

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Learning to Segment Inputs for NMT Favors Character-Level Processing

    VIEW 4 EXCERPTS
    CITES METHODS & BACKGROUND
    HIGHLY INFLUENCED

    FILTER CITATIONS BY YEAR

    2016
    2020

    CITATION STATISTICS

    • 39 Highly Influenced Citations

    • Averaged 65 Citations per year from 2018 through 2020

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 36 REFERENCES