Assessing the Syntactic Capabilities of Transformer-based Multilingual Language Models

被引:0
|
作者
Perez-Mayos, Laura [1 ]
Taboas Garcia, Alba [1 ]
Mille, Simon [1 ]
Wanner, Leo [1 ,2 ]
机构
[1] Univ Pompeu Fabra, TALN Res Grp, Barcelona, Spain
[2] Catalan Inst Res & Adv Studies ICREA, Barcelona, Spain
基金
欧盟地平线“2020”;
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
Multilingual Transformer-based language models, usually pretrained on more than 100 languages, have been shown to achieve outstanding results in a wide range of cross-lingual transfer tasks. However, it remains unknown whether the optimization for different languages conditions the capacity of the models to generalize over syntactic structures, and how languages with syntactic phenomena of different complexity are affected. In this work, we explore the syntactic generalization capabilities of the monolingual and multilingual versions of BERT and RoBERTa. More specifically, we evaluate the syntactic generalization potential of the models on English and Spanish tests, comparing the syntactic abilities of monolingual and multilingual models on the same language (English), and of multilingual models on two different languages (English and Spanish). For English, we use the available SyntaxGym test suite; for Spanish, we introduce SyntaxGymES, a novel ensemble of targeted syntactic tests in Spanish, designed to evaluate the syntactic generalization capabilities of language models through the SyntaxGym online platform.
引用
收藏
页码:3799 / 3812
页数:14
相关论文
共 50 条
  • [1] Ouroboros: On Accelerating Training of Transformer-Based Language Models
    Yang, Qian
    Huo, Zhouyuan
    Wang, Wenlin
    Huang, Heng
    Carin, Lawrence
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 32 (NIPS 2019), 2019, 32
  • [2] Transformer-Based Language Models for Software Vulnerability Detection
    Thapa, Chandra
    Jang, Seung Ick
    Ahmed, Muhammad Ejaz
    Camtepe, Seyit
    Pieprzyk, Josef
    Nepal, Surya
    [J]. PROCEEDINGS OF THE 38TH ANNUAL COMPUTER SECURITY APPLICATIONS CONFERENCE, ACSAC 2022, 2022, : 481 - 496
  • [3] RadBERT: Adapting Transformer-based Language Models to Radiology
    Yan, An
    McAuley, Julian
    Lu, Xing
    Du, Jiang
    Chang, Eric Y.
    Gentili, Amilcare
    Hsu, Chun-Nan
    [J]. RADIOLOGY-ARTIFICIAL INTELLIGENCE, 2022, 4 (04)
  • [4] A Comparison of Transformer-Based Language Models on NLP Benchmarks
    Greco, Candida Maria
    Tagarelli, Andrea
    Zumpano, Ester
    [J]. NATURAL LANGUAGE PROCESSING AND INFORMATION SYSTEMS (NLDB 2022), 2022, 13286 : 490 - 501
  • [5] Applications of transformer-based language models in bioinformatics: a survey
    Zhang, Shuang
    Fan, Rui
    Liu, Yuti
    Chen, Shuang
    Liu, Qiao
    Zeng, Wanwen
    [J]. NEURO-ONCOLOGY ADVANCES, 2023, 5 (01)
  • [6] TAG: Gradient Attack on Transformer-based Language Models
    Deng, Jieren
    Wang, Yijue
    Li, Ji
    Wang, Chenghong
    Shang, Chao
    Liu, Hang
    Rajasekaran, Sanguthevar
    Ding, Caiwen
    [J]. FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, EMNLP 2021, 2021, : 3600 - 3610
  • [7] Practical Transformer-based Multilingual Text Classification
    Wang, Cindy
    Banko, Michele
    [J]. 2021 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, NAACL-HLT 2021, 2021, : 121 - 129
  • [8] Multilingual Controllable Transformer-Based Lexical Simplification
    Sheang, Kim Cheng
    Saggion, Horacio
    [J]. PROCESAMIENTO DEL LENGUAJE NATURAL, 2023, (71): : 109 - 123
  • [9] Multilingual Transformer-Based Personality Traits Estimation
    Leonardi, Simone
    Monti, Diego
    Rizzo, Giuseppe
    Morisio, Maurizio
    [J]. INFORMATION, 2020, 11 (04)
  • [10] Assessing the Effectiveness of Multilingual Transformer-based Text Embeddings for Named Entity Recognition in Portuguese
    de Lima Santos, Diego Bernardes
    de Carvalho Dutra, Frederico Giffoni
    Parreiras, Fernando Silva
    Brandao, Wladmir Cardoso
    [J]. PROCEEDINGS OF THE 23RD INTERNATIONAL CONFERENCE ON ENTERPRISE INFORMATION SYSTEMS (ICEIS 2021), VOL 1, 2021, : 473 - 483