Following on from my last post I have been looking for Octave code for the denoising autoencoder to avoid reinventing the wheel and writing it myself from scratch, and luckily I have found two options. The first is a tutorial on autoencoders, by a Piotr Mirowski, which has a link to a Github page with code. The second is a toolbox for dimensionality reduction, by Laurens van der Maaten, which has autoencoders as one of its options.
As of now I'm not sure which one I'll use, and perhaps I might yet write my own code heavily reusing elements from both of the above. More in due course.
