Uncertainty-Guided Semi-Supervised Few-Shot Class-Incremental Learning With Knowledge Distillation

• Published in: IEEE transactions on multimedia Vol. 25; pp. 1 - 14
• Main Authors: Cui, Yawen, Deng, Wanxia, Xu, Xin, Liu, Zhen, Liu, Zhong, Pietikainen, Matti, Liu, Li
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: class-incremental learning; computer vision; Data models; Deep learning; Distillation; Few-shot learning; knowledge distillation; Machine learning; object classification; Power capacitors; Semi-supervised learning; Semisupervised learning; Task analysis; Training; Uncertainty; uncertainty estimation
• ISSN: 1520-9210; 1941-0077
• DOI: 10.1109/TMM.2022.3208743

Class-Incremental Learning (CIL) aims at incrementally learning novel classes without forgetting old ones. This capability becomes more challenging when novel tasks contain one or a few labeled training samples, which leads to a more practical learning scenario, i.e ., Few-Shot Class-Incremental Learning (FSCIL). The dilemma on FSCIL lies in serious overfitting and exacerbated catastrophic forgetting caused by the limited training data from novel classes. In this paper, excited by the easy accessibility of unlabeled data, we conduct a pioneering work and focus on a Semi-Supervised Few-Shot Class-Incremental Learning (Semi-FSCIL) problem, which requires the model incrementally to learn new classes from extremely limited labeled samples and a large number of unlabeled samples. To address this problem, a simple but efficient framework is first constructed based on the knowledge distillation technique to alleviate catastrophic forgetting. To efficiently mitigate the overfitting problem on novel categories with unlabeled data, uncertainty-guided semi-supervised learning is incorporated into this framework to select unlabeled samples into incremental learning sessions considering the model uncertainty. This process provides extra reliable supervision for the distillation process and contributes to better formulating the class means. Our extensive experiments on CIFAR100, miniImageNet and CUB200 datasets demonstrate the promising performance of our proposed method, and define baselines in this new research direction.