Does BERT really agree ? Fine-grained Analysis of Lexical Dependence on a Syntactic Task

Karim Lasri, Alessandro Lenci, Thierry Poibeau


Abstract
Although transformer-based Neural Language Models demonstrate impressive performance on a variety of tasks, their generalization abilities are not well understood. They have been shown to perform strongly on subject-verb number agreement in a wide array of settings, suggesting that they learned to track syntactic dependencies during their training even without explicit supervision. In this paper, we examine the extent to which BERT is able to perform lexically-independent subject-verb number agreement (NA) on targeted syntactic templates. To do so, we disrupt the lexical patterns found in naturally occurring stimuli for each targeted structure in a novel fine-grained analysis of BERT’s behavior. Our results on nonce sentences suggest that the model generalizes well for simple templates, but fails to perform lexically-independent syntactic generalization when as little as one attractor is present.
Anthology ID:
2022.findings-acl.181
Volume:
Findings of the Association for Computational Linguistics: ACL 2022
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Smaranda Muresan, Preslav Nakov, Aline Villavicencio
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2309–2315
Language:
URL:
https://aclanthology.org/2022.findings-acl.181
DOI:
10.18653/v1/2022.findings-acl.181
Bibkey:
Cite (ACL):
Karim Lasri, Alessandro Lenci, and Thierry Poibeau. 2022. Does BERT really agree ? Fine-grained Analysis of Lexical Dependence on a Syntactic Task. In Findings of the Association for Computational Linguistics: ACL 2022, pages 2309–2315, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
Does BERT really agree ? Fine-grained Analysis of Lexical Dependence on a Syntactic Task (Lasri et al., Findings 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/add_acl24_videos/2022.findings-acl.181.pdf
Video:
 https://preview.aclanthology.org/add_acl24_videos/2022.findings-acl.181.mp4