A Simple Autoencoder Implementation in C++
An autoencoder is a neural network, basically having the same input and output data values. Thus it tries to learn the representation of the data set. By imposing some criterias on the number of hidden layer dimensions (for example the sparsity constraint), successful representation of the data can be found.
Not: This implementation is not optimized and may have bugs. It is not recomennded to be used in a production environment.