A Single Linear Layer Yields Task-Adapted Low-Rank Matrices

Hwichan Kim, Shota Sasaki, Sho Hoshino, Ukyo Honda


Abstract
Low-Rank Adaptation (LoRA) is a widely used Parameter-Efficient Fine-Tuning (PEFT) method that updates an initial weight matrix W0 with a delta matrix 𝛥 W consisted by two low-rank matrices A and B. A previous study suggested that there is correlation between W0 and 𝛥 W. In this study, we aim to delve deeper into relationships between W0 and low-rank matrices A and B to further comprehend the behavior of LoRA. In particular, we analyze a conversion matrix that transform W0 into low-rank matrices, which encapsulates information about the relationships. Our analysis reveals that the conversion matrices are similar across each layer. Inspired by these findings, we hypothesize that a single linear layer, which takes each layer’s W0 as input, can yield task-adapted low-rank matrices. To confirm this hypothesis, we devise a method named Conditionally Parameterized LoRA (CondLoRA) that updates initial weight matrices with low-rank matrices derived from a single linear layer. Our empirical results show that CondLoRA maintains a performance on par with LoRA, despite the fact that the trainable parameters of CondLoRA are fewer than those of LoRA. Therefore, we conclude that “a single linear layer yields task-adapted low-rank matrices.” The code used in our experiments is available at https://github.com/CyberAgentAILab/CondLoRA.
Anthology ID:
2024.lrec-main.141
Volume:
Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)
Month:
May
Year:
2024
Address:
Torino, Italia
Editors:
Nicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessandro Lenci, Sakriani Sakti, Nianwen Xue
Venues:
LREC | COLING
SIG:
Publisher:
ELRA and ICCL
Note:
Pages:
1602–1608
Language:
URL:
https://aclanthology.org/2024.lrec-main.141
DOI:
Bibkey:
Cite (ACL):
Hwichan Kim, Shota Sasaki, Sho Hoshino, and Ukyo Honda. 2024. A Single Linear Layer Yields Task-Adapted Low-Rank Matrices. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 1602–1608, Torino, Italia. ELRA and ICCL.
Cite (Informal):
A Single Linear Layer Yields Task-Adapted Low-Rank Matrices (Kim et al., LREC-COLING 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.lrec-main.141.pdf