Critical Patch-Aware Sparse Prompting with Decoupled Training for Continual Learning on the Edge

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lim, Wonseon, Lee, Jaesung, Kim, Dae-Won
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917392202858496
author Lim, Wonseon
Lee, Jaesung
Kim, Dae-Won
author_facet Lim, Wonseon
Lee, Jaesung
Kim, Dae-Won
contents Continual learning (CL) on edge devices requires not only high accuracy but also training-time efficiency to support on-device adaptation under strict memory and computational constraints. While prompt-based continual learning (PCL) is parameter-efficient and achieves competitive accuracy, prior work has focused mainly on accuracy or inference-time performance, often overlooking the memory and computational costs of on-device training. In this paper, we propose CPS-Prompt, a critical patch-aware sparse prompting framework that explicitly targets training-time memory usage and computational cost by integrating critical patch sampling (CPS) for task-aware token reduction and decoupled prompt and classifier training (DPCT) to reduce backpropagation overhead. Experiments on three public benchmarks and real edge hardware show that CPS-Prompt improves peak memory, training time, and energy efficiency by about 1.6x over the balanced CODA-Prompt baseline, while maintaining accuracy within 2% of the state-of-the-art C-Prompt on average and remaining competitive with CODA-Prompt in accuracy. The code is available at https://github.com/laymond1/cps-prompt.
format Preprint
id arxiv_https___arxiv_org_abs_2604_07399
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Critical Patch-Aware Sparse Prompting with Decoupled Training for Continual Learning on the Edge
Lim, Wonseon
Lee, Jaesung
Kim, Dae-Won
Machine Learning
Continual learning (CL) on edge devices requires not only high accuracy but also training-time efficiency to support on-device adaptation under strict memory and computational constraints. While prompt-based continual learning (PCL) is parameter-efficient and achieves competitive accuracy, prior work has focused mainly on accuracy or inference-time performance, often overlooking the memory and computational costs of on-device training. In this paper, we propose CPS-Prompt, a critical patch-aware sparse prompting framework that explicitly targets training-time memory usage and computational cost by integrating critical patch sampling (CPS) for task-aware token reduction and decoupled prompt and classifier training (DPCT) to reduce backpropagation overhead. Experiments on three public benchmarks and real edge hardware show that CPS-Prompt improves peak memory, training time, and energy efficiency by about 1.6x over the balanced CODA-Prompt baseline, while maintaining accuracy within 2% of the state-of-the-art C-Prompt on average and remaining competitive with CODA-Prompt in accuracy. The code is available at https://github.com/laymond1/cps-prompt.
title Critical Patch-Aware Sparse Prompting with Decoupled Training for Continual Learning on the Edge
topic Machine Learning
url https://arxiv.org/abs/2604.07399