Grow, Assess, Compress: Adaptive Backbone Scaling for Memory-Efficient Class Incremental Learning

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Garcia-Castañeda, Adrian, Irureta, Jon, Imaz, Jon, Lojo, Aizea
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917376662962176
author Garcia-Castañeda, Adrian
Irureta, Jon
Imaz, Jon
Lojo, Aizea
author_facet Garcia-Castañeda, Adrian
Irureta, Jon
Imaz, Jon
Lojo, Aizea
contents Class Incremental Learning (CIL) poses a fundamental challenge: maintaining a balance between the plasticity required to learn new tasks and the stability needed to prevent catastrophic forgetting. While expansion-based methods effectively mitigate forgetting by adding task-specific parameters, they suffer from uncontrolled architectural growth and memory overhead. In this paper, we propose a novel dynamic scaling framework that adaptively manages model capacity through a cyclic "GRow, Assess, ComprEss" (GRACE) strategy. Crucially, we supplement backbone expansion with a novel saturation assessment phase that evaluates the utilization of the model's capacity. This assessment allows the framework to make informed decisions to either expand the architecture or compress the backbones into a streamlined representation, preventing parameter explosion. Experimental results demonstrate that our approach achieves state-of-the-art performance across multiple CIL benchmarks, while reducing memory footprint by up to a 73% compared to purely expansionist models.
format Preprint
id arxiv_https___arxiv_org_abs_2603_08426
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Grow, Assess, Compress: Adaptive Backbone Scaling for Memory-Efficient Class Incremental Learning
Garcia-Castañeda, Adrian
Irureta, Jon
Imaz, Jon
Lojo, Aizea
Machine Learning
Computer Vision and Pattern Recognition
Class Incremental Learning (CIL) poses a fundamental challenge: maintaining a balance between the plasticity required to learn new tasks and the stability needed to prevent catastrophic forgetting. While expansion-based methods effectively mitigate forgetting by adding task-specific parameters, they suffer from uncontrolled architectural growth and memory overhead. In this paper, we propose a novel dynamic scaling framework that adaptively manages model capacity through a cyclic "GRow, Assess, ComprEss" (GRACE) strategy. Crucially, we supplement backbone expansion with a novel saturation assessment phase that evaluates the utilization of the model's capacity. This assessment allows the framework to make informed decisions to either expand the architecture or compress the backbones into a streamlined representation, preventing parameter explosion. Experimental results demonstrate that our approach achieves state-of-the-art performance across multiple CIL benchmarks, while reducing memory footprint by up to a 73% compared to purely expansionist models.
title Grow, Assess, Compress: Adaptive Backbone Scaling for Memory-Efficient Class Incremental Learning
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.08426