Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2409.11286 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910608138436608 |
|---|---|
| author | Chen, Bingzhi Zhou, Haoming Liu, Yishu Zeng, Biqing Pan, Jiahui Lu, Guangming |
| author_facet | Chen, Bingzhi Zhou, Haoming Liu, Yishu Zeng, Biqing Pan, Jiahui Lu, Guangming |
| contents | Most recent few-shot learning approaches are based on meta-learning with episodic training. However, prior studies encounter two crucial problems: (1) \textit{the presence of inductive bias}, and (2) \textit{the occurrence of catastrophic forgetting}. In this paper, we propose a novel Multi-Level Contrastive Constraints (MLCC) framework, that jointly integrates within-episode learning and across-episode learning into a unified interactive learning paradigm to solve these issues. Specifically, we employ a space-aware interaction modeling scheme to explore the correct inductive paradigms for each class between within-episode similarity/dis-similarity distributions. Additionally, with the aim of better utilizing former prior knowledge, a cross-stage distribution adaption strategy is designed to align the across-episode distributions from different time stages, thus reducing the semantic gap between existing and past prediction distribution. Extensive experiments on multiple few-shot datasets demonstrate the consistent superiority of MLCC approach over the existing state-of-the-art baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_11286 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Enhancing Few-Shot Classification without Forgetting through Multi-Level Contrastive Constraints Chen, Bingzhi Zhou, Haoming Liu, Yishu Zeng, Biqing Pan, Jiahui Lu, Guangming Multimedia Most recent few-shot learning approaches are based on meta-learning with episodic training. However, prior studies encounter two crucial problems: (1) \textit{the presence of inductive bias}, and (2) \textit{the occurrence of catastrophic forgetting}. In this paper, we propose a novel Multi-Level Contrastive Constraints (MLCC) framework, that jointly integrates within-episode learning and across-episode learning into a unified interactive learning paradigm to solve these issues. Specifically, we employ a space-aware interaction modeling scheme to explore the correct inductive paradigms for each class between within-episode similarity/dis-similarity distributions. Additionally, with the aim of better utilizing former prior knowledge, a cross-stage distribution adaption strategy is designed to align the across-episode distributions from different time stages, thus reducing the semantic gap between existing and past prediction distribution. Extensive experiments on multiple few-shot datasets demonstrate the consistent superiority of MLCC approach over the existing state-of-the-art baselines. |
| title | Enhancing Few-Shot Classification without Forgetting through Multi-Level Contrastive Constraints |
| topic | Multimedia |
| url | https://arxiv.org/abs/2409.11286 |