Metraipolis › archive › dc_4fdf56932d04

Width Expansion as a Method for Class Incremental Learning — arXiv:2609.37702

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.37702
License (as archived)
CC0-1.0
Archived
2026-10-07T16:27:36.058Z
SHA-256 of the archived text
c8bf9863f4efcd7b475f7030f0a970fdcb76651781bd40e4365bf20a73affa82
Stable URL of this page
https://metraipolis.com/documento/dc_4fdf56932d04

How to cite

Width Expansion as a Method for Class Incremental Learning — arXiv:2609.37702 — arXiv (en), https://arxiv.org/abs/2609.37702 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4fdf56932d04 (Metraipolis archive dc_4fdf56932d04, sha256 c8bf9863f4efcd7b475f7030f0a970fdcb76651781bd40e4365bf20a73affa82)
BibTeX
@misc{dc4fdf56932d04,
  title = {Width Expansion as a Method for Class Incremental Learning — arXiv:2609.37702},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.37702},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4fdf56932d04; sha256 c8bf9863f4efcd7b475f7030f0a970fdcb76651781bd40e4365bf20a73affa82},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.37702. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Class Incremental Learning (Class-IL) requires models to learn new classes over time while preserving previously acquired knowledge without access to past data or task identity. This setting intensifies the stability-plasticity dilemma and makes catastrophic forgetting a central challenge. Existing approaches include regularization, knowledge distillation, replay, and architectural expansion. However, many expansion methods rely on explicit task identifiers or predefined growth strategies, limiting their applicability when task boundaries are unavailable at inference time. This work proposes a dynamic width expansion method that increases the number of neurons within existing layers according to a normalized loss criterion, without requiring task-specific information. An attention mechanism with persistent key-value memory is also incorporated to stabilize feature representations and reduce interference between previously learned and newly introduced classes. The approach is evaluated on Split MNIST and Split CIFAR-100 under the standard Class-IL protocol. Experiments compare fixed-capacity and dynamically expanding architectures, both with and without attention, combined with established continual learning methods including EWC, LwF, and A-GEM. Results show that progressive width expansion consistently improves performance over fixed architectures, particularly when combined with functional methods and A-GEM. The combination of width expansion and attention provides the most consistent gains. Overall, dynamic width expansion based on representational demand provides an effective and flexible strategy for Class-IL, although uncontrolled growth may increase overfitting and computational cost.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.37702 — 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 2026-10-07T15:54:54.016Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).