Empirical Study of Easy and Hard Examples in CNN Training

TL;DR

This study investigates the properties of easy and hard examples in CNN training, revealing their visual similarities, contributions to generalization, and potential for dataset compression.

Contribution

It provides a detailed analysis of easy and hard examples across CNN architectures, highlighting their roles in generalization and dataset efficiency.

Findings

Easy examples are visually similar across CNNs.

Hard examples are visually diverse.

Removing easy examples impairs generalization.

Abstract

Deep Neural Networks (DNNs) generalize well despite their massive size and capability of memorizing all examples. There is a hypothesis that DNNs start learning from simple patterns and the hypothesis is based on the existence of examples that are consistently well-classified at the early training stage (i.e., easy examples) and examples misclassified (i.e., hard examples). Easy examples are the evidence that DNNs start learning from specific patterns and there is a consistent learning process. It is important to know how DNNs learn patterns and obtain generalization ability, however, properties of easy and hard examples are not thoroughly investigated (e.g., contributions to generalization and visual appearances). In this work, we study the similarities of easy and hard examples respectively for different Convolutional Neural Network (CNN) architectures, assessing how those examples contribute to generalization. Our results show that easy examples are visually similar to each other and hard examples are visually diverse, and both examples are largely shared across different CNN architectures. Moreover, while hard examples tend to contribute more to generalization than easy examples, removing a large number of easy examples leads to poor generalization. By analyzing those results, we hypothesize that biases in a dataset and Stochastic Gradient Descent (SGD) are the reasons why CNNs have consistent easy and hard examples. Furthermore, we show that large scale classification datasets can be efficiently compressed by using easiness proposed in this work.