EUVIMEDEUVIMED entdecken
Crossref

Model Attention Expansion for Few-Shot Class-Incremental Learning

Xuan Wang, Zhong Ji, Yunlong Yu, Yanwei Pang, Jungong Han

IEEE Transactions on Image Processing · 2024 · 33 · 4419-4431

DOI: 10.1109/tip.2024.3434475 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Image Processing, 2024. Verfasst von Xuan Wang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 51 Literaturangaben. Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Few-Shot Class-Incremental Learning (FSCIL) aims at incrementally learning new knowledge from limited training examples without forgetting previous knowledge. However, we observe that existing methods face a challenge known as supervision collapse, where the model disproportionately emphasizes class-specific features of base classes at the detriment of novel class representations, leading to restricted cognitive capabilities. To alleviate this issue, we propose a new framework, Model aTtention Expansion for Few-Shot Class-Incremental Learning (MTE-FSCIL), aimed at expanding the model attention fields to improve transferability without compromising the discriminative capability for base classes. Specifically, the framework adopts a dual-stage training strategy, comprising pre-training and meta-training stages. In the pre-training stage, we present a new regularization technique, named the Reserver (RS) loss, to expand the global perception and reduce over-reliance on class-specific features by amplifying feature map activations. During the meta-training stage, we introduce the Repeller (RP) loss, a novel pair-based loss that promotes variation in representations and improves the model's recognition of sample uniqueness by scattering intra-class samples within the embedding space. Furthermore, we propose a Transformational Adaptation (TA) strategy to enable continuous incorporation of new knowledge from downstream tasks, thus facilitating cross-task knowledge transfer. Extensive experimental results on mini-ImageNet, CIFAR100, and CUB200 datasets demonstrate that our proposed framework consistently outperforms the state-of-the-art methods.

Förderung

  • National Key Research and Development Program of China
  • National Natural Science Foundation of China

Zitieren

APA

Wang, X., Ji, Z., Yu, Y., Pang, Y., & Han, J. (2024). Model Attention Expansion for Few-Shot Class-Incremental Learning. IEEE Transactions on Image Processing, 33, 4419-4431. https://doi.org/10.1109/tip.2024.3434475

Vancouver

Wang X, Ji Z, Yu Y, Pang Y, Han J. Model Attention Expansion for Few-Shot Class-Incremental Learning. IEEE Transactions on Image Processing. 2024;33:4419-4431. doi:10.1109/tip.2024.3434475

BibTeX
@article{Wang2024,
  author = {Wang, Xuan and Ji, Zhong and Yu, Yunlong and Pang, Yanwei and Han, Jungong},
  title = {Model Attention Expansion for Few-Shot Class-Incremental Learning},
  journal = {IEEE Transactions on Image Processing},
  year = {2024},
  volume = {33},
  pages = {4419--4431},
  doi = {10.1109/tip.2024.3434475}
}
Kostenlos starten →