Character-level Representations Improve DRS-based Semantic Parsing Even in the Age of BERT

    Onderzoeksoutput: Conference contributionAcademicpeer review

    13 Downloads (Pure)

    Samenvatting

    We combine character-level and contextual language model representations to improve performance on Discourse Representation Structure parsing. Character representations can easily be added in a sequence-to-sequence model in either one encoder or as a fully separate encoder, with improvements that are robust to different language models, languages and data sets. For English, these improvements are larger than adding individual sources of linguistic information or adding non-contextual embeddings. A new method of analysis based on semantic tags demonstrates that the character-level representations improve performance across a subset of selected semantic phenomena.
    Originele taal-2English
    TitelProceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
    UitgeverijAssociation for Computational Linguistics (ACL)
    Pagina's4587-4603
    Aantal pagina's17
    DOI's
    StatusPublished - 16-nov-2020

    Citeer dit