2011 Memisevic, R.
Gradient-based learning of higher-order image features.
International Conference on Computer Vision (ICCV 2011).
[pdf] © IEEE

Python implementation of a factored gated autoencoder.

Example code that applies the model to shifted random dots.

The model allows you to learn filters like these or these.

datafile 1 and datafile 2 used by the script.

The implementation is based on theano. (There are also older versions of the model and script.)