[go: up one dir, main page]

On Surgical Fine-tuning for Language Encoders

Abhilasha Lodha, Gayatri Belapurkar, Saloni Chalkapurkar, Yuanming Tao, Reshmi Ghosh, Samyadeep Basu, Dmitrii Petrov, Soundararajan Srinivasan


Abstract
Fine-tuning all the layers of a pre-trained neural language encoder (either using all the parameters or using parameter-efficient methods) is often the de-facto way of adapting it to a new task. We show evidence that for different downstream language tasks, fine-tuning only a subset of layers is sufficient to obtain performance that is close to and often better than fine-tuning all the layers in the language encoder. We propose an efficient metric based on the diagonal of the Fisher information matrix (FIM score), to select the candidate layers for selective fine-tuning. We show, empirically on GLUE and SuperGLUE tasks and across distinct language encoders, that this metric can effectively select layers leading to a strong downstream performance. Our work highlights that task-specific information corresponding to a given downstream task is often localized within a few layers, and tuning only those is sufficient for strong performance. Additionally, we demonstrate the robustness of the FIM score to rank layers in a manner that remains constant during the optimization process.
Anthology ID:
2023.findings-emnlp.204
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:
3105–3113
Language:
URL:
https://aclanthology.org/2023.findings-emnlp.204
DOI:
10.18653/v1/2023.findings-emnlp.204
Bibkey:
Cite (ACL):
Abhilasha Lodha, Gayatri Belapurkar, Saloni Chalkapurkar, Yuanming Tao, Reshmi Ghosh, Samyadeep Basu, Dmitrii Petrov, and Soundararajan Srinivasan. 2023. On Surgical Fine-tuning for Language Encoders. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 3105–3113, Singapore. Association for Computational Linguistics.
Cite (Informal):
On Surgical Fine-tuning for Language Encoders (Lodha et al., Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-emnlp.204.pdf