Convolutional Neural Network Approach to Classifying the CIFAR-10 Dataset

Abstract

Convolutional neural network (CNN) is a powerful tool that can be used in many applications of machine learning. This paper demonstrates the effectiveness of using a CNN to classify images in the CIFAR-10 dataset. The model achieved an accuracy of 0.6276 and a loss of 1.116452 on the validation set. It was observed that the accuracy of predictions varied from class to class, and this paper discusses the potential causes for this variation, such as similar classes sharing common features. Further research in this field could lead to improvement in driving assistance technology and eventually automated driving.