Latent Representations

Latent representations are the summary of features that could fully describe an input. It's the sparse representation generated by, for example, an auto-encoder. In other words, it is the simplified model of the input data.

"Considering an auto-encoder, the central layer of this network (after training) will contain a simplified representation of the input data (i.e. a summary of key features), which can be used to reconstruct the output."