This package contains an implementation of the experimental results in the following two papers:

• [1] Yoshua Bengio, Eric Thibodeau-Laufer, Jason Yosinski. Deep Generative Stochastic Networks Trainable by Backprop. arXiv preprint arXiv:1306.1091. (PDF, BibTeX)

• [2] Yoshua Bengio, Li Yao, Guillaume Alain, Pascal Vincent. Generalized Denoising Auto-Encoders as Generative Models. NIPS, 2013. (PDF, BibTeX)

Information for setting up TensorFlow can be found in the official documentation.
The implementation has been tested with version r0.7 of TensorFlow running on Python 2.7.

1. To run a one layer Generalized Denoising Autoencoder with a walkback procedure (paper [2])

    python dae_walkback.py
   

2. To run a one layer Generalized Denoising Autoencoder without a walkback procedure (paper [2])

    python dae_no_walkback.py
   

3. To run a Generative Stochastic Network (paper [1])

    python gsn.py