How Conservative are Language Models? Adapting to the Introduction of Gender-Neutral Pronouns

被引：0
作者：
Brandl, Stephanie [1 ]
Cui, Ruixiang [1 ]
Sogaard, Anders [1 ]
机构：
[1] Univ Copenhagen, Copenhagen, Denmark
来源：
NAACL 2022: THE 2022 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES | 2022年
关键词：
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Gender-neutral pronouns have recently been introduced in many languages to a) include non-binary people and b) as a generic singular. Recent results from psycholinguistics suggest that gender-neutral pronouns (in Swedish) are not associated with human processing difficulties. This, we show, is in sharp contrast with automated processing. We show that gender-neutral pronouns in Danish, English, and Swedish are associated with higher perplexity, more dispersed attention patterns, and worse downstream performance. We argue that such conservativity in language models may limit widespread adoption of gender-neutral pronouns and must therefore be resolved.
引用
收藏
页码：3624 / 3630
页数：7
相关论文
共 38 条
  • [1] Abnar S, 2020, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020), P4190
  • [2] [Anonymous], 2020, ENGLISH LANGUAGE TEA, DOI DOI 10.1186/S13045-019-0838-Y
  • [3] Barrett Maria, 2021, P 4 WORKSH COMP MOD, P63
  • [4] On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?
    Bender, Emily M.
    Gebru, Timnit
    McMillan-Major, Angelina
    Shmitchell, Shmargaret
    [J]. PROCEEDINGS OF THE 2021 ACM CONFERENCE ON FAIRNESS, ACCOUNTABILITY, AND TRANSPARENCY, FACCT 2021, 2021, : 610 - 623
  • [5] Bowman S., 2015, P 2015 C EMP METH NA, P632, DOI DOI 10.18653/V1/D15-1075
  • [6] Cao YT, 2020, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020), P4568
  • [7] What does BERT look at? An Analysis of BERT's Attention
    Clark, Kevin
    Khandelwal, Urvashi
    Levy, Omer
    Manning, Christopher D.
    [J]. BLACKBOXNLP WORKSHOP ON ANALYZING AND INTERPRETING NEURAL NETWORKS FOR NLP AT ACL 2019, 2019, : 276 - 286
  • [8] Conneau Alexis, 2020, P 58 ANN M ASS COMP, P8440, DOI [10.18653/v1/2020.acl-main.747, DOI 10.18653/V1/2020.ACL-MAIN.747]
  • [9] Dev S, 2021, 2021 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP 2021), P1968
  • [10] Devlin J, 2019, 2019 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES (NAACL HLT 2019), VOL. 1, P4171
1234