@inproceedings{tresoldi-2022-approaching,
    title = "Approaching Reflex Predictions as a Classification Problem Using Extended Phonological Alignments",
    author = "Tresoldi, Tiago",
    editor = "Vylomova, Ekaterina  and
      Ponti, Edoardo  and
      Cotterell, Ryan",
    booktitle = "Proceedings of the 4th Workshop on Research in Computational Linguistic Typology and Multilingual NLP",
    month = jul,
    year = "2022",
    address = "Seattle, Washington",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2022.sigtyp-1.11/",
    doi = "10.18653/v1/2022.sigtyp-1.11",
    pages = "86--93",
    abstract = "This work describes an implementation of the {\textquotedblleft}extended alignment{\textquotedblright} model for cognate reflex prediction submitted to the {\textquotedblleft}SIGTYP 2022 Shared Task on the Prediction of Cognate Reflexes{\textquotedblright}. Similarly to List et al. (2022a), the technique involves an automatic extension of sequence alignments with multilayered vectors that encode informational tiers on both site-specific traits, such as sound classes and distinctive features, as well as contextual and suprasegmental ones, conveyed by cross-site referrals and replication. The method allows to generalize the problem of cognate reflex prediction as a classification problem, with models trained using a parallel corpus of cognate sets. A model using random forests is trained and evaluated on the shared task for reflex prediction, and the experimental results are presented and discussed along with some differences to other implementations."
}