Logarithm-transform aided Gaussian Sampling for Few-Shot Learning

TL;DR

This paper introduces a novel logarithm-transform to better approximate Gaussian distributions in representations, significantly improving few-shot image classification performance with less data.

Contribution

A new logarithm-based Gaussian transform is proposed, outperforming existing methods in transforming data for few-shot learning tasks.

Findings

Significant performance gains in few-shot classification

Effective data transformation with less sampling

Outperforms existing Gaussian approximation methods

Abstract

Few-shot image classification has recently witnessed the rise of representation learning being utilised for models to adapt to new classes using only a few training examples. Therefore, the properties of the representations, such as their underlying probability distributions, assume vital importance. Representations sampled from Gaussian distributions have been used in recent works, [19] to train classifiers for few-shot classification. These methods rely on transforming the distributions of experimental data to approximate Gaussian distributions for their functioning. In this paper, I propose a novel Gaussian transform, that outperforms existing methods on transforming experimental data into Gaussian-like distributions. I then utilise this novel transformation for few-shot image classification and show significant gains in performance, while sampling lesser data.