Hands-On Artificial Intelligence for Beginners

Network architecture

Autoencoders work by taking an input and generating a smaller vector representation for later reconstructing its own input. They do this by using an encoder to impose an information bottleneck on incoming data, and then utilizing a decoder to recreate the input data based on that representation. This is based on the idea that there are structures within data (that is, correlations, and so on) that exist, but that are not readily apparent. Autoencoders are a means of automatically learning these relationships without explicitly doing so.

Structurally, autoencoders consist of an input layer, a hidden layer, and an output layer, as demonstrated in the following diagram: 

The encoder learns to preserve as much of the relevant ...