Abstract
Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-efficient fine-tuning with pre-trained models reduces parameter overhead but can suffer from cumulative interference and suboptimal alignment between inference samples and specialized modules. We propose Dynamic LoRA-Experts and Prototype-Ensemble Matching (DLEPEM), a two-stage rehearsal-free framework. DLEPEM allocates a task-specific LoRA-Expert for each incremental task to reduce cross-task interference, then combines frozen pre-trained-model prototypes with task-adaptive LoRA-Expert prototypes for reliable task-level discrimination. Experiments on standard CIL and Few-Shot CIL benchmarks demonstrate strong performance under the evaluated protocols.
Keywords
Publication details
- DOI
- 10.3390/app16126153
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Zhao, H., Liu, R., & Liu, Y. (2026). Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning. https://doi.org/10.3390/app16126153
MLA 9
Zhao, Hongwei, et al. "Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning." https://doi.org/10.3390/app16126153.
Chicago (author–date)
Zhao, Hongwei, Rui Liu, and Yansong Liu. 2026. "Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning." https://doi.org/10.3390/app16126153.
Harvard
Zhao, H., Liu, R. and Liu, Y. (2026) 'Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning', doi:10.3390/app16126153.
Vancouver
Zhao H, Liu R, Liu Y. Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning. doi:10.3390/app16126153
IEEE
H. Zhao, R. Liu, and Y. Liu, "Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning," doi: 10.3390/app16126153.