Classifying Handwritten Numbers Using Convolutional Neural Networks

Abstract

A convolutional neural network (CNN) is one of the most significant networks in the deep learning field. Since CNN has made impressive achievements in many areas, including but not limited to computer vision and natural language processing, CNNs have attracted much attention from both industry and academia in the past few years. The problem of gradient vanishing or gradient explosion tends to get worse as the depth of the model increases. In traditional neural network structures, especially in the field of image processing, since a large number of convolutional and pooling layers need to be utilized to extract features layer by layer, the model performance tends to degrade and other unfavorable situations as the number of layers accumulates. In order to solve the gradient problem that occurs during the training process of deep neural networks, the concept of residual connectivity has emerged.