Continuing with the theme of unsupervised learning and generative models, while researching GANs I came across the concept of autoencoding. … More

Multi-Layer Perceptron (MLP)/Deep Neural Networks (DNN)

Last time, we briefly introduced MNIST, and provided some performance metrics of the SLP model on classifying the digits. A subset of MNIST characters and their associated labels While I was pleasantly (and perhaps naively) surprised as to how well the SLP performed, the classification accuracy achieved (around 86%) is considered unthinkable, maybe even borderline… Continue Reading →