Convolutional Neural Net Vs MNIST

In previous posts I used a single layer neural network to try and classify hand written digits, with mixed results. Luckily there is a neural network architecture much more suited for this task. Convolutional neural networks are based on the visual cortex and have achieved very low error rates on the MNIST dataset. The Network …

Neural Net Vs MNIST (Part 2)

In part 1 we trained a 1 layer neural net to predict the correct digit when shown a 28×28 gray scale image of a handwritten digit. After several thousand training steps the network eventually managed to get an average test error of ~10%. In this post I would like to look under the hood of …

Neural Net Vs MNIST (Part 1)

Nearly every machine learning blog inevitably does a post on the MNIST data set. I don’t want to break with tradition…so here is mine. What is MNIST ? MNIST is a data set containing 60000 images of handwritten digits (0-9) along with a label indicating what digit the image represents. Each image is 28×28 pixels …

Visualizing how a deep neural network learns

Last weekend I was working through some Tensor Flow tutorials and thought that it would be nice to be able to visualize the weight change over time in a multi layer neural network. For this I will be using the example of a pizza delivery company. Unfortunately I could not find any pizza delivery data …