Deep Autoencoder is the extension of the simple Autoencoder. The first layer present in DeepAutoencoder is responsible for first-order functions in the raw input. The second layer is responsible for second-order functions corresponding to patterns in the appearance of first-order functions. Deeper layers which are available in the Deep Autoencoder tend to learn even high-order features.
A deep autoencoder is the combination of two, symmetrical deep-belief networks:
First four or five shallow layers represent the encoding half.
The other combination of four or five layers makes up the decoding half.