Frequency Effects on Syntactic Rule Learning in Transformers

Jason Wei, Dan Garrette, Tal Linzen, Ellie Pavlick

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    Pre-trained language models perform well on a variety of linguistic tasks that require symbolic reasoning, raising the question of whether such models implicitly represent abstract symbols and rules. We investigate this question using the case study of BERT's performance on English subject-verb agreement. Unlike prior work, we train multiple instances of BERT from scratch, allowing us to perform a series of controlled interventions at pre-training time. We show that BERT often generalizes well to subject-verb pairs that never occurred in training, suggesting a degree of rule-governed behavior. We also find, however, that performance is heavily influenced by word frequency, with experiments showing that both the absolute frequency of a verb form, as well as the frequency relative to the alternate inflection, are causally implicated in the predictions BERT makes at inference time. Closer analysis of these frequency effects reveals that BERT's behavior is consistent with a system that correctly applies the SVA rule in general but struggles to overcome strong training priors and to estimate agreement features (singular vs. plural) on infrequent lexical items.

    Original languageEnglish (US)
    Title of host publicationEMNLP 2021 - 2021 Conference on Empirical Methods in Natural Language Processing, Proceedings
    PublisherAssociation for Computational Linguistics (ACL)
    Pages932-948
    Number of pages17
    ISBN (Electronic)9781955917094
    StatePublished - 2021
    Event2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021 - Virtual, Punta Cana, Dominican Republic
    Duration: Nov 7 2021Nov 11 2021

    Publication series

    NameEMNLP 2021 - 2021 Conference on Empirical Methods in Natural Language Processing, Proceedings

    Conference

    Conference2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021
    Country/TerritoryDominican Republic
    CityVirtual, Punta Cana
    Period11/7/2111/11/21

    ASJC Scopus subject areas

    • Computational Theory and Mathematics
    • Computer Science Applications
    • Information Systems

    Fingerprint

    Dive into the research topics of 'Frequency Effects on Syntactic Rule Learning in Transformers'. Together they form a unique fingerprint.

    Cite this