Experiments with Hinton's recirculation algorithm for training autoencoders with Python & NumPy - no TensorFlow required!
A basic description for the training algorithm ('recirculation') and the outline for the first experiment here can be found in this paper:
Hinton, G. E. and McClelland, J. L. Learning representations by recirculation. Neural Information Processing Systems, 1988, pages 358-366. http://www.cs.toronto.edu/~hinton/absps/recirculation.pdf