How does the pre-training objective affect what large language models learn about linguistic properties?

Ahmed Alajrami, Nikolaos Aletras


Abstract
Several pre-training objectives, such as masked language modeling (MLM), have been proposed to pre-train language models (e.g. BERT) with the aim of learning better language representations. However, to the best of our knowledge, no previous work so far has investigated how different pre-training objectives affect what BERT learns about linguistics properties. We hypothesize that linguistically motivated objectives such as MLM should help BERT to acquire better linguistic knowledge compared to other non-linguistically motivated objectives that are not intuitive or hard for humans to guess the association between the input and the label to be predicted. To this end, we pre-train BERT with two linguistically motivated objectives and three non-linguistically motivated ones. We then probe for linguistic characteristics encoded in the representation of the resulting models. We find strong evidence that there are only small differences in probing performance between the representations learned by the two different types of objectives. These surprising results question the dominant narrative of linguistically informed pre-training.
Anthology ID:
2022.acl-short.16
Volume:
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
Month:
May
Year:
2022
Address:
Dublin, Ireland
Editors:
Smaranda Muresan, Preslav Nakov, Aline Villavicencio
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
131–147
Language:
URL:
https://aclanthology.org/2022.acl-short.16
DOI:
10.18653/v1/2022.acl-short.16
Bibkey:
Cite (ACL):
Ahmed Alajrami and Nikolaos Aletras. 2022. How does the pre-training objective affect what large language models learn about linguistic properties?. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 131–147, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
How does the pre-training objective affect what large language models learn about linguistic properties? (Alajrami & Aletras, ACL 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.acl-short.16.pdf
Video:
 https://aclanthology.org/2022.acl-short.16.mp4
Code
 aajrami/acl2022-pre-training-objectives-probing
Data
GLUE