I am currently working on autoencoders using R. I have tried with the autoencoder package in R, but I am unable to interpret the results. So I am trying to write a R code for autoencoders. My question is how encoding and decoding is done in autoencoders. Is there any formula or algorithm for it? Can anyone please help me with it?
Asked
Active
Viewed 162 times
2
-
you can refer my answer over here https://stackoverflow.com/questions/47632208/differene-between-autoencoder-network-and-fully-convolution-network/47966014#47966014... the formula is feedforward and backprop formulas that's it – Jai Dec 26 '17 at 22:43
-
Friendly advice, work on understanding how to read those results properly. It'll be much better. also, you need to figure out the autoencoders architecture on the basis of your dataset. e.g if you're dealing with images, then its better to build autoencoders with CNNs or if its textual data, then with LSTMs or GRU. – Asad Mahmood Oct 12 '18 at 17:48