@inproceedings{peters-martins-2020-one,
    title = "One-Size-Fits-All Multilingual Models",
    author = "Peters, Ben  and
      Martins, Andr{\'e} F. T.",
    editor = "Nicolai, Garrett  and
      Gorman, Kyle  and
      Cotterell, Ryan",
    booktitle = "Proceedings of the 17th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology",
    month = jul,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2020.sigmorphon-1.4/",
    doi = "10.18653/v1/2020.sigmorphon-1.4",
    pages = "63--69",
    abstract = "This paper presents DeepSPIN{'}s submissions to Tasks 0 and 1 of the SIGMORPHON 2020 Shared Task. For both tasks, we present multilingual models, training jointly on data in all languages. We perform no language-specific hyperparameter tuning {--} each of our submissions uses the same model for all languages. Our basic architecture is the sparse sequence-to-sequence model with entmax attention and loss, which allows our models to learn sparse, local alignments while still being trainable with gradient-based techniques. For Task 1, we achieve strong performance with both RNN- and transformer-based sparse models. For Task 0, we extend our RNN-based model to a multi-encoder set-up in which separate modules encode the lemma and inflection sequences. Despite our models' lack of language-specific tuning, they tie for first in Task 0 and place third in Task 1."
}