Improving Temporal Generalization of Pre-trained Language Models with Lexical Semantic Change

Zhaochen Su, Zecheng Tang, Xinyan Guan, Lijun Wu, Min Zhang, Juntao Li


Abstract
Recent research has revealed that neural language models at scale suffer from poor temporal generalization capability, i.e., language model pre-trained on static data from past years performs worse over time on emerging data. Existing methods mainly perform continual training to mitigate such a misalignment. While effective to some extent but is far from being addressed on both the language modeling and downstream tasks. In this paper, we empirically observe that temporal generalization is closely affiliated with lexical semantic change, which is one of the essential phenomena of natural languages. Based on this observation, we propose a simple yet effective lexical-level masking strategy to post-train a converged language model. Experiments on two pre-trained language models, two different classification tasks, and four benchmark datasets demonstrate the effectiveness of our proposed method over existing temporal adaptation methods, i.e., continual training with new data. Our code is available at https://github.com/zhaochen0110/LMLM.
Anthology ID:
2022.emnlp-main.428
Volume:
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6380–6393
Language:
URL:
https://aclanthology.org/2022.emnlp-main.428
DOI:
10.18653/v1/2022.emnlp-main.428
Bibkey:
Cite (ACL):
Zhaochen Su, Zecheng Tang, Xinyan Guan, Lijun Wu, Min Zhang, and Juntao Li. 2022. Improving Temporal Generalization of Pre-trained Language Models with Lexical Semantic Change. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6380–6393, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Improving Temporal Generalization of Pre-trained Language Models with Lexical Semantic Change (Su et al., EMNLP 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/add_acl24_videos/2022.emnlp-main.428.pdf
Software:
 2022.emnlp-main.428.software.zip