Improving Classification Accuracy with Graph Filtering

TL;DR

This paper introduces a graph filtering technique to reduce intra-class noise in training data, thereby improving classification accuracy, especially in few-shot learning scenarios, by asymptotically reducing intra-class variance while preserving class means.

Contribution

The paper proposes a novel graph filtering method that reduces intra-class noise and variance, enhancing classification performance in general and few-shot learning settings.

Findings

Slight improvement over state-of-the-art in vision benchmarks

Effective noise reduction in intra-class samples

Applicable to both few-shot and standard classification

Abstract

In machine learning, classifiers are typically susceptible to noise in the training data. In this work, we aim at reducing intra-class noise with the help of graph filtering to improve the classification performance. Considered graphs are obtained by connecting samples of the training set that belong to a same class depending on the similarity of their representation in a latent space. We show that the proposed graph filtering methodology has the effect of asymptotically reducing intra-class variance, while maintaining the mean. While our approach applies to all classification problems in general, it is particularly useful in few-shot settings, where intra-class noise can have a huge impact due to the small sample selection. Using standardized benchmarks in the field of vision, we empirically demonstrate the ability of the proposed method to slightly improve state-of-the-art results in both cases of few-shot and standard classification.