TRAMS: Training-free Memory Selection for Long-range Language Modeling

Haofei Yu, Cunxiang Wang, Yue Zhang, Wei Bi


Abstract
The Transformer architecture is crucial for numerous AI models, but it still faces challenges in long-range language modeling. Though several specific transformer architectures have been designed to tackle issues of long-range dependencies, existing methods like Transformer-XL are plagued by a high percentage of ineffective memories. In this study, we present a plug-and-play strategy, known as TRAining-free Memory Selection (TRAMS), that selects tokens participating in attention calculation based on one simple metric. This strategy allows us to keep tokens that are likely to have a high attention score with the current queries and ignore the other ones. We have tested our approach on the word-level benchmark (WikiText-103) and the character-level benchmark (enwik8), and the results indicate an improvement without having additional training or adding additional parameters.
Anthology ID:
2023.findings-emnlp.331
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2023
Month:
December
Year:
2023
Address:
Singapore
Editors:
Houda Bouamor, Juan Pino, Kalika Bali
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4966–4972
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.331
DOI:
10.18653/v1/2023.findings-emnlp.331
Bibkey:
Cite (ACL):
Haofei Yu, Cunxiang Wang, Yue Zhang, and Wei Bi. 2023. TRAMS: Training-free Memory Selection for Long-range Language Modeling. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 4966–4972, Singapore. Association for Computational Linguistics.
Cite (Informal):
TRAMS: Training-free Memory Selection for Long-range Language Modeling (Yu et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-5/2023.findings-emnlp.331.pdf