小样本增量学习论文阅读
Graph-Based Methods
TOPIC
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123640256.pdf

CEC
https://openaccess.thecvf.com/content/CVPR2021/papers/Zhang_Few-Shot_Incremental_Learning_With_Continually_Evolved_Classifiers_CVPR_2021_paper.pdf

Clustering-Based Methods
IDL-VQ
https://openreview.net/pdf?id=3SV-ZePhnZM
增量深度学习向量化利用高斯混合将学习类别的视觉特征量化为参考向量质心。随后,任何传入的新类都由它们与现有参考向量的 soft similarity 进行表示。为了减少灾难性遗忘的现象,作者额外存储了每个类别的一个样本并在增量训练的期间进行回放。
SA-KD
https://openaccess.thecvf.com/content/CVPR2021/papers/Cheraghian_Semantic-Aware_Knowledge_Distillation_for_Few-Shot_Class-Incremental_Learning_CVPR_2021_paper.pdf

SUB-REG
https://openreview.net/forum?id=boJy41J-tnQ

FACT
https://openaccess.thecvf.com/content/CVPR2022/papers/Ramanujan_Forward_Compatible_Training_for_Large-Scale_Embedding_Retrieval_Systems_CVPR_2022_paper.pdf

Architectural Methods
FSLL
https://ojs.aaai.org/index.php/AAAI/article/view/16334/16141

C-FSCIL
https://openaccess.thecvf.com/content/CVPR2022/papers/Hersche_Constrained_Few-Shot_Class-Incremental_Learning_CVPR_2022_paper.pdf
