Multi-word Tokenization for Sequence Compression

Leonidas Gee, Leonardo Rigutini, Marco Ernandes, Andrea Zugarini


Abstract
Large Language Models have proven highly successful at modelling a variety of tasks. However, this comes at a steep computational cost that hinders wider industrial uptake. In this paper, we present MWT: a Multi-Word Tokenizer that goes beyond word boundaries by representing frequent multi-word expressions as single tokens. MWTs produce a more compact and efficient tokenization that yields two benefits: (1) Increase in performance due to a greater coverage of input data given a fixed sequence length budget; (2) Faster and lighter inference due to the ability to reduce the sequence length with negligible drops in performance. Our results show that MWT is more robust across shorter sequence lengths, thus allowing for major speedups via early sequence truncation.
Anthology ID:
2023.emnlp-industry.58
Volume:
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track
Month:
December
Year:
2023
Address:
Singapore
Editors:
Mingxuan Wang, Imed Zitouni
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
612–621
Language:
URL:
https://aclanthology.org/2023.emnlp-industry.58
DOI:
10.18653/v1/2023.emnlp-industry.58
Bibkey:
Cite (ACL):
Leonidas Gee, Leonardo Rigutini, Marco Ernandes, and Andrea Zugarini. 2023. Multi-word Tokenization for Sequence Compression. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 612–621, Singapore. Association for Computational Linguistics.
Cite (Informal):
Multi-word Tokenization for Sequence Compression (Gee et al., EMNLP 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.emnlp-industry.58.pdf
Video:
 https://aclanthology.org/2023.emnlp-industry.58.mp4