Unsupervised and reinforcement learning in neural networks EPFL CS-434
Joachim Muth and Sebastien Speierer
November 17, 2015
For this project, our task is to understand the Kohonen algorithm and apply it to a data set of hand-written digits. A Kohonen map is an artificial neural network trained using unsupervised learning to produce a low- dimensional representation of the input space. For this exercise, our inputs consists of vectorized 28x28 images of digits so it has a dimension of 784. Thus, the Kohonen map will be very useful allowing us to look at this data on a two dimensional map.