Episode Adaptive Embedding Networks for Few-shot Learning

TL;DR

This paper introduces EAEN, a novel few-shot learning method that creates episode-specific embeddings to improve classification accuracy by capturing discriminative features unique to each episode.

Contribution

The paper proposes EAEN, a new approach that learns episode-specific embeddings using distributional information, enhancing generalization and reducing overfitting in few-shot learning.

Findings

EAEN improves classification accuracy by 10-20% over state-of-the-art methods.

EAEN effectively captures episode-specific discriminative features.

Extensive experiments validate EAEN's robustness across datasets and backbones.

Abstract

Few-shot learning aims to learn a classifier using a few labelled instances for each class. Metric-learning approaches for few-shot learning embed instances into a high-dimensional space and conduct classification based on distances among instance embeddings. However, such instance embeddings are usually shared across all episodes and thus lack the discriminative power to generalize classifiers according to episode-specific features. In this paper, we propose a novel approach, namely \emph{Episode Adaptive Embedding Network} (EAEN), to learn episode-specific embeddings of instances. By leveraging the probability distributions of all instances in an episode at each channel-pixel embedding dimension, EAEN can not only alleviate the overfitting issue encountered in few-shot learning tasks, but also capture discriminative features specific to an episode. To empirically verify the effectiveness and robustness of EAEN, we have conducted extensive experiments on three widely used benchmark datasets, under various combinations of different generic embedding backbones and different classifiers. The results show that EAEN significantly improves classification accuracy about $10\%$ to $20\%$ in different settings over the state-of-the-art methods.