Stacked Denoising Auto-encoders

[pluskid]

(Deep) auto-encoders and variants are important ways of doing unsupervised deep learning. Since they share some properties with current Mocha architecture, it would be good to add some auto-encoders to Mocha.

• Denoising Auto-encoders ((denoising) auto-encoder #28)
• Stacked auto-encoders, an interface to do layer-wise training (Freeze Layer Parameters #31, Example of pre-training with stacked denoising auto-encoder #32)
• Tutorial documentation (Tutorial document on denoising auto-encoders #35)