how to design a multiple layer autoencoder for "unsupervised" learning ?

Hello to the experts in deep learning.

I have been struggling to find a three-hidden layered sparse auto-encoder in matlab. I am using matlab 2017a but couldn't get any help on the same apart from stacked deepnet followed by a softmax layer at the end, which is totally different from the concept of 3 hidden layers within an autoencoder eg: [Input, N1, N2, N1, Output]

I am able to do the encoding part for the first two layers but its the follow up that becomes a problem

Is there any documentation on it in MATLAB?

[Samuel Andersson]

Hi sir,

I'm struggling with this myself. Did you find a solution in the end?

Cheers,

Sam

[BR]

Hi,

Sorry for the late response, I switched to python. Not a very convinient one but that's the only solution I found.

Thanks

BR