Augmented Efficient BackProp for backpropagation learning in deep autoassociative neural networks

We introduce Augmented Efficient BackProp as a strategy for applying the backpropagation algorithm to deep autoencoders, i.e., autoassociators with many hidden layers, without relying on a weight initialization using restricted Boltzmann machines. This training method is an extension of Efficient BackProp, first proposed by LeCun et al. [1], and is benchmarked on three different types of application datasets.