Reading questions

If you are unsure about any of these concepts, post a question on the discussion forum!

  • What is the relationship between an autoencoder (or autoassociator) and a regular feed-forward neural network?
  • What is the reconstruction error?
  • Why is training an autoencoder with more hidden units than input units not interesting?
  • What is a denoising autoencoder?
  • What is a contractive autoencoder?
  • Why does it make sense to train a denoising autoencoder or a contractive autoencoder with more hidden units than input units (unlike a regular autoencoder)?