Neuronal networks: Autoencoder

In the videos I speak a bit of the self-coder or Autoencoder a simple method of learning with images, the goal of the network is to learn from the examples given to them of input and response.
In the basic example the network learns the same image and has to solve a similar one. But if the example is completed, you can invest or solve images very different from the originals.

Source Code

Posted in Tutorials and tagged , , .