From neural PCA to deep unsupervised learning
This work addresses a bottleneck in deep unsupervised learning for researchers, though it appears incremental as it builds on existing autoencoder and denoising frameworks.
The paper tackles the problem of inefficient learning in deep unsupervised networks by proposing an autoencoder with lateral shortcut connections, which speeds up training and enables learning of invariant features, as demonstrated in experiments.
A network supporting deep unsupervised learning is presented. The network is an autoencoder with lateral shortcut connections from the encoder to decoder at each level of the hierarchy. The lateral shortcut connections allow the higher levels of the hierarchy to focus on abstract invariant features. While standard autoencoders are analogous to latent variable models with a single layer of stochastic variables, the proposed network is analogous to hierarchical latent variables models. Learning combines denoising autoencoder and denoising sources separation frameworks. Each layer of the network contributes to the cost function a term which measures the distance of the representations produced by the encoder and the decoder. Since training signals originate from all levels of the network, all layers can learn efficiently even in deep networks. The speedup offered by cost terms from higher levels of the hierarchy and the ability to learn invariant features are demonstrated in experiments.