Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning

Page view(s)
0
Checked on
Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning
Title:
Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning
Journal Title:
2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
Keywords:
Publication Date:
08 April 2025
Citation:
Gu, Y., Yang, M., Yang, X., Wei, K., Zhu, H., Goenawan, G. J., & Deng, C. (2025). Dynamic Adapter Tuning for Long-Tailed Class-Incremental Learning. 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 8176–8185. https://doi.org/10.1109/wacv61041.2025.00793
Abstract:
Long-tailed class-incremental learning (LT-CIL) aims to learn new classes continuously from a long-tailed data stream, while simultaneously dealing with challenges such as imbalanced learning of tail classes and catastrophic for-getting. To address these challenges, most existing methods employ a two-stage strategy by initializing model training from scratch with further balanced knowledge driven cali-bration. This strategy faces challenges in deriving discrim-inative features from cold-started backbones for the long-tailed distribution of data, consequently leading to relatively diminished performance. In this paper, with the pow-erful feature extraction capability of pre-trained foundation models, we have achieved a one-stage approach that de-livers superior performance. Specifically, we propose Dy-namic Adapter Tuning (DAT), which employs a dynamic adapter cache mechanism to adapt a pre-trained model to learn tasks sequentially. The adapter in the cache is either dynamically selected or created according to task similar-ity, and further compactified with the new task's adapter to mitigate cross-task and cross-class gaps in LT-CIL, sig-nificantly alleviating catastrophic forgetting and imbalance learning issues, respectively. With extensive experimental validation, our method consistently achieves state-of-the-art performance under the challenging LT-CIL setting.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the Economic Development Board - Space Technology Development Programme
Grant Reference no. : S22-19016-STDP.
Description:
© 2025 IEEE.  Personal use of this material is permitted.  Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works
ISSN:
2642-9381
Files uploaded:

File Size Format Action
gu-dynamic-adapter-tuning-for-long-tailed-class-incremental-learning-wacv-2025-paper.pdf 746.09 KB PDF Request a copy