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

TL;DR

This paper introduces a three-stage framework for generalized and incremental few-shot learning that effectively learns new classes, prevents forgetting, and calibrates classifiers, achieving state-of-the-art results on multiple benchmarks.

Contribution

The work presents a novel three-stage approach that explicitly addresses learning, forgetting prevention, and calibration in few-shot learning scenarios.

Findings

Achieved state-of-the-art results on four benchmark datasets.

Effectively prevents catastrophic forgetting of base classes.

Provides calibrated classifiers across all classes.

Abstract

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.