Generalized and Incremental Few-Shot Learning by Explicit Learning and Calibration without Forgetting

By Anna Kukleva, Hilde Kuehne, Bernt Schiele. 2021

76 citations recorded by the source corpus.

Both generalized and incremental few-shot learning have to deal with three major challenges: learning novel classes from only few samples per class, preventing catastrophic forgetting of base classes, and classifier calibration across novel and base classes. In this work we propose a three-stage framework that allows to explicitly and effectively address these challenges. While the first phase learns base classes with many samples, the second phase learns a calibrated classifier for novel classes from few samples while also preventing catastrophic forgetting. In the final phase, calibration is achieved across all classes. We evaluate the proposed framework on four challenging benchmark datasets for image and video few-shot classification and obtain state-of-the-art results for both generalized and incremental few shot learning.

Subfields: Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Meta & Continual Learning

1. Xingxing Zhang
2. Liyuan Wang
3. Bernt Schiele
4. Yaoyao Liu
5. Jun Zhu
6. Hang Su
7. Qianru Sun
8. Pei-Hung Chen

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “The extremely limited training samples exacerbate the overfitting of previously-learned representations to subsequent tasks, which can be alleviated by recent work such as preserving the topology of representations (TOPIC [404]), constructing an exemplar relation graph for knowledge distillation (ER”
• RMM: Reinforced Memory Management for Class-Incremental Learning (2023, background): “…task-incremental learning—each phase corresponds to a task (dataset) that contains new data of all seen classes [7, 10, 19, 29, 43, 47, 57], or class-incremental learning (CIL)—each phase contains data of a new set of classes, i.e., classes are unseen [4, 6, 18, 25, 31, 37, 40, 41, 49, 52, 55–57].”
• Few-Shot Incremental Learning for Label-to-Image Translation (2022, methodology): “Following the FSCIL setting, later works propose strategies such as vector quantization [6] and calibrated classifiers [21] to improve classification performance.”
• A Survey on Few-Shot Class-Incremental Learning (2023, methodology): “Direct replay Kukleva et al. (2021) proposed a three-stage framework, wherein the first two stages train the network on base and novel classes separately and employ a model parameter constraint method to prevent forgetting of old classes.”
• Warping the Space: Weight Space Rotation for Class-Incremental Few-Shot Learning (2023, background): “Several other works adopt different strategies to learn new classes well during incremental sessions without forgetting (Tao et al., 2020; Cheraghian et al., 2021a; Zhang et al., 2021; Kukleva et al., 2021; Chen & Lee, 2021; Akyürek et al., 2022).”