Metraipolis › archive › dc_f83ea91d4662

SPARS: A Stability-Plasticity-Aware Region Selection Framework for Continual Learning — arXiv:2610.19876

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19876
License (as archived)
CC0-1.0
Archived
2026-10-15T14:18:46.151Z
SHA-256 of the archived text
28597c21fd4d16c26ff16ed0baf28b1697e861c1c5bc4c436606e733cf98344d
Stable URL of this page
https://metraipolis.com/documento/dc_f83ea91d4662

How to cite

SPARS: A Stability-Plasticity-Aware Region Selection Framework for Continual Learning — arXiv:2610.19876 — arXiv (en), https://arxiv.org/abs/2610.19876 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f83ea91d4662 (Metraipolis archive dc_f83ea91d4662, sha256 28597c21fd4d16c26ff16ed0baf28b1697e861c1c5bc4c436606e733cf98344d)
BibTeX
@misc{dcf83ea91d4662,
  title = {SPARS: A Stability-Plasticity-Aware Region Selection Framework for Continual Learning — arXiv:2610.19876},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19876},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f83ea91d4662; sha256 28597c21fd4d16c26ff16ed0baf28b1697e861c1c5bc4c436606e733cf98344d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19876. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Continual learning aims to acquire new knowledge while retaining previously learned knowledge. However, updating shared parameters when learning new tasks often overwrites existing knowledge, causing catastrophic forgetting. Memory replay methods mitigate catastrophic forgetting but rely on previous samples, raising privacy, memory, and scalability concerns. On the other hand, parameter isolation-based methods preserve learned knowledge but may limit knowledge transfer, whereas regularization promotes adaptability but remains susceptible to interference between old and new tasks. To maintain the stability-plasticity trade-off, this paper proposes Stability-Plasticity-Aware Region Selection (SPARS), a replay-free hybrid framework combining global parameter isolation with localized parameter regularization. SPARS estimates parameter importance using either the Fisher Information Matrix (FIM) or a novel Confidence Score (CS) estimator that leverages logit sensitivity and achieves comparable performance to FIM without requiring labelled data. Based on parameter importance, SPARS partitions parameters into three importance-aware regions namely, winning, neighboring, and non-important parameter regions. Each region receives a different adaptation strategy: selective freezing, soft regularization, or unrestricted learning based on its role in the stability-plasticity trade-off. To mitigate classifier bias in class-incremental learning, we employ Quadratic Discriminant Analysis to improve performance on previously learned classes .Experiments on CIFAR-10 and TinyImageNet demonstrate consistent improvement in average incremental accuracy over existing isolation- and regularization-based methods across task- and class-incremental learning. On TinyImageNet, SPARS achieves the best performance with 6% improvement for class-incremental learning while retaining over 60% of parameters for future tasks.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.19876 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T13:53:35.662Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).