yumatsuoka / Auto_Encoder

Auto Encoder(include sparse and denoiging one)

Implement auto encoder and observe hiddin layer and output layer's weight and output in auto encoder.
That's because I wanna know auto encoder actually shrinks data from datasets in its hidden layer.
I also implemented nomal 'auto encoder', 'sparse auto encoder' and 'denoising auto encoder' with MNIST dataset.

The result on the experiments is summarized in the following slide.
http://www.slideshare.net/YumaMatsuoka/auto-encoder

Requirements

• Python 3.x (checked Python 3.4.3)
• numpy(checked numpy 1.11.0)
• pylab(checked matplotlib 1.5.0)
• sklearn(checked scikit-learn 0.17)

How to use

python autoencoder_multi.py