LFD-ProtoNet: Prototypical Network Based on Local Fisher Discriminant Analysis for Few-shot Learning

TL;DR

This paper introduces LFD-ProtoNet, an enhancement of the prototypical network for few-shot learning that incorporates local Fisher discriminant analysis to improve classification accuracy by better handling support set variance.

Contribution

It combines ProtoNet with local Fisher discriminant analysis to reduce within-class variance and enhance between-class separation, improving few-shot learning performance.

Findings

Achieves higher classification accuracy on miniImageNet and tieredImageNet datasets.

Provides theoretical risk bounds supporting the method's effectiveness.

Demonstrates robustness to high variance in support sets.

Abstract

The prototypical network (ProtoNet) is a few-shot learning framework that performs metric learning and classification using the distance to prototype representations of each class. It has attracted a great deal of attention recently since it is simple to implement, highly extensible, and performs well in experiments. However, it only takes into account the mean of the support vectors as prototypes and thus it performs poorly when the support set has high variance. In this paper, we propose to combine ProtoNet with local Fisher discriminant analysis to reduce the local within-class covariance and increase the local between-class covariance of the support set. We show the usefulness of the proposed method by theoretically providing an expected risk bound and empirically demonstrating its superior classification accuracy on miniImageNet and tieredImageNet.