CD^2: Constrained Dataset Distillation for Few-Shot Class-Incremental Learning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909988827430912 |
|---|---|
| author | Bao, Kexin Zhang, Daichi Zhang, Hansong Li, Yong Yue, Yutao Ge, Shiming |
| author_facet | Bao, Kexin Zhang, Daichi Zhang, Hansong Li, Yong Yue, Yutao Ge, Shiming |
| contents | Few-shot class-incremental learning (FSCIL) receives significant attention from the public to perform classification continuously with a few training samples, which suffers from the key catastrophic forgetting problem. Existing methods usually employ an external memory to store previous knowledge and treat it with incremental classes equally, which cannot properly preserve previous essential knowledge. To solve this problem and inspired by recent distillation works on knowledge transfer, we propose a framework termed \textbf{C}onstrained \textbf{D}ataset \textbf{D}istillation (\textbf{CD$^2$}) to facilitate FSCIL, which includes a dataset distillation module (\textbf{DDM}) and a distillation constraint module~(\textbf{DCM}). Specifically, the DDM synthesizes highly condensed samples guided by the classifier, forcing the model to learn compacted essential class-related clues from a few incremental samples. The DCM introduces a designed loss to constrain the previously learned class distribution, which can preserve distilled knowledge more sufficiently. Extensive experiments on three public datasets show the superiority of our method against other state-of-the-art competitors. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_08519 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CD^2: Constrained Dataset Distillation for Few-Shot Class-Incremental Learning Bao, Kexin Zhang, Daichi Zhang, Hansong Li, Yong Yue, Yutao Ge, Shiming Computer Vision and Pattern Recognition Artificial Intelligence Few-shot class-incremental learning (FSCIL) receives significant attention from the public to perform classification continuously with a few training samples, which suffers from the key catastrophic forgetting problem. Existing methods usually employ an external memory to store previous knowledge and treat it with incremental classes equally, which cannot properly preserve previous essential knowledge. To solve this problem and inspired by recent distillation works on knowledge transfer, we propose a framework termed \textbf{C}onstrained \textbf{D}ataset \textbf{D}istillation (\textbf{CD$^2$}) to facilitate FSCIL, which includes a dataset distillation module (\textbf{DDM}) and a distillation constraint module~(\textbf{DCM}). Specifically, the DDM synthesizes highly condensed samples guided by the classifier, forcing the model to learn compacted essential class-related clues from a few incremental samples. The DCM introduces a designed loss to constrain the previously learned class distribution, which can preserve distilled knowledge more sufficiently. Extensive experiments on three public datasets show the superiority of our method against other state-of-the-art competitors. |
| title | CD^2: Constrained Dataset Distillation for Few-Shot Class-Incremental Learning |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2601.08519 |