@inproceedings{lippincott-etal-2019-jhu,
    title = "{JHU} System Description for the {MADAR} {A}rabic Dialect Identification Shared Task",
    author = "Lippincott, Tom  and
      Shapiro, Pamela  and
      Duh, Kevin  and
      McNamee, Paul",
    editor = "El-Hajj, Wassim  and
      Belguith, Lamia Hadrich  and
      Bougares, Fethi  and
      Magdy, Walid  and
      Zitouni, Imed  and
      Tomeh, Nadi  and
      El-Haj, Mahmoud  and
      Zaghouani, Wajdi",
    booktitle = "Proceedings of the Fourth Arabic Natural Language Processing Workshop",
    month = aug,
    year = "2019",
    address = "Florence, Italy",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/W19-4634/",
    doi = "10.18653/v1/W19-4634",
    pages = "264--268",
    abstract = "Our submission to the MADAR shared task on Arabic dialect identification employed a language modeling technique called Prediction by Partial Matching, an ensemble of neural architectures, and sources of additional data for training word embeddings and auxiliary language models. We found several of these techniques provided small boosts in performance, though a simple character-level language model was a strong baseline, and a lower-order LM achieved best performance on Subtask 2. Interestingly, word embeddings provided no consistent benefit, and ensembling struggled to outperform the best component submodel. This suggests the variety of architectures are learning redundant information, and future work may focus on encouraging decorrelated learning."
}