Learning Class-level Prototypes for Few-shot Learning

TL;DR

This paper introduces a novel prototype generation framework for few-shot learning that mitigates outlier influence, leading to improved classification performance on standard benchmarks.

Contribution

It proposes an episodic prototype generator module that produces more representative prototypes from limited support data, enhancing few-shot learning accuracy.

Findings

Significant performance improvement over baseline models.

Competitive results on miniImageNet and tieredImageNet datasets.

Effective cross-domain adaptation from miniImageNet to CUB-200-2011.

Abstract

Few-shot learning aims to recognize new categories using very few labeled samples. Although few-shot learning has witnessed promising development in recent years, most existing methods adopt an average operation to calculate prototypes, thus limited by the outlier samples. In this work, we propose a simple yet effective framework for few-shot classification, which can learn to generate preferable prototypes from few support data, with the help of an episodic prototype generator module. The generated prototype is meant to be close to a certain \textit{\targetproto{}} and is less influenced by outlier samples. Extensive experiments demonstrate the effectiveness of this module, and our approach gets a significant raise over baseline models, and get a competitive result compared to previous methods on \textit{mini}ImageNet, \textit{tiered}ImageNet, and cross-domain (\textit{mini}ImageNet $\rightarrow$ CUB-200-2011) datasets.