ABM-LoRA: Activation Boundary Matching for Fast Convergence in Low-Rank Adaptation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lee, Dongha, Park, Jinhee, Kim, Minjun, Kwon, Junseok
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914173207707648
author Lee, Dongha
Park, Jinhee
Kim, Minjun
Kwon, Junseok
author_facet Lee, Dongha
Park, Jinhee
Kim, Minjun
Kwon, Junseok
contents We propose Activation Boundary Matching for Low-Rank Adaptation (ABM-LoRA), a principled initialization strategy that substantially accelerates the convergence of low-rank adapters. While LoRA offers high parameter efficiency, its random initialization restricts gradient updates to a mismatched tangent space, causing significant information loss and hindering early convergence. Our ABM-LoRA addresses this by aligning the adapter's activation boundaries with those of the pretrained model before downstream training, thereby maximizing the projection of full-parameter gradients into the adapter subspace. This alignment sharply reduces information loss at initialization, yields a lower starting loss, and accelerates convergence. We demonstrate ABM-LoRA's effectiveness across diverse architectures and tasks: language understanding (T5-Base on GLUE), dialogue generation (LLaMA2-7B on WizardLM), and vision recognition (ViT-B/16 on VTAB-1K). On VTAB-1K, it achieves the highest accuracy among all methods, with strong gains on structured reasoning tasks requiring geometric understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2511_19145
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ABM-LoRA: Activation Boundary Matching for Fast Convergence in Low-Rank Adaptation
Lee, Dongha
Park, Jinhee
Kim, Minjun
Kwon, Junseok
Computer Vision and Pattern Recognition
We propose Activation Boundary Matching for Low-Rank Adaptation (ABM-LoRA), a principled initialization strategy that substantially accelerates the convergence of low-rank adapters. While LoRA offers high parameter efficiency, its random initialization restricts gradient updates to a mismatched tangent space, causing significant information loss and hindering early convergence. Our ABM-LoRA addresses this by aligning the adapter's activation boundaries with those of the pretrained model before downstream training, thereby maximizing the projection of full-parameter gradients into the adapter subspace. This alignment sharply reduces information loss at initialization, yields a lower starting loss, and accelerates convergence. We demonstrate ABM-LoRA's effectiveness across diverse architectures and tasks: language understanding (T5-Base on GLUE), dialogue generation (LLaMA2-7B on WizardLM), and vision recognition (ViT-B/16 on VTAB-1K). On VTAB-1K, it achieves the highest accuracy among all methods, with strong gains on structured reasoning tasks requiring geometric understanding.
title ABM-LoRA: Activation Boundary Matching for Fast Convergence in Low-Rank Adaptation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.19145