Deep Convolutional Generative Adversial Network on MNIST Dataset

Abstract

In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning. Training on various image datasets, we show convincing evidence that our deep convolutional adversarial pair learns a hierarchy of representations from object parts to scenes in both the generator and discriminator. The generator uses tf.keras.layers.conv2Dtranspose (up sampling) layers to produce an image from a seed (random noise). Start with a dense layer that takes this seed as input, then up sample several times until you reach the desired image size of 28x28x1. The discriminator is a CNN-based image classifier. The model will be trained to output positive values for real images, and negative values for fake images. We define the Generator loss and the discriminator loss and we finally we get new images that look similar to our input(MNIST) images.