Tokenization with Factorized Subword Encoding

David Samuel, Lilja Øvrelid


Abstract
In recent years, language models have become increasingly larger and more complex. However, the input representations for these models continue to rely on simple and greedy subword tokenization methods. In this paper, we propose a novel tokenization method that factorizes subwords onto discrete triplets using a VQ-VAE model. The effectiveness of the proposed tokenization method, referred to as the Factorizer, is evaluated on language modeling and morpho-syntactic tasks for 7 diverse languages. Results indicate that this method is more appropriate and robust for morphological tasks than the commonly used byte-pair encoding (BPE) tokenization algorithm.
Anthology ID:
2023.findings-acl.890
Volume:
Findings of the Association for Computational Linguistics: ACL 2023
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Anna Rogers, Jordan Boyd-Graber, Naoaki Okazaki
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
14143–14161
Language:
URL:
https://aclanthology.org/2023.findings-acl.890
DOI:
10.18653/v1/2023.findings-acl.890
Bibkey:
Cite (ACL):
David Samuel and Lilja Øvrelid. 2023. Tokenization with Factorized Subword Encoding. In Findings of the Association for Computational Linguistics: ACL 2023, pages 14143–14161, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Tokenization with Factorized Subword Encoding (Samuel & Øvrelid, Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-3/2023.findings-acl.890.pdf