Learning by Reconstruction

Sparse autoencoder

Autoencoder incorporating a sparsity constraint on hidden layer activations, encouraging the model to use only a subset of neurons to represent each input.

← Geri