Discovering hidden factors of variation in deep networks

This is a Tensorflow 2.0 implementation of Discovering hidden factors of variation in deep networks by Brian Cheung et al. (2014). This repository contains reproduce of several experiments mentioned in the paper. Based on the Lasagne (RIP) example (link).

Abstract

Deep learning has enjoyed a great deal of success because of its ability to learnuseful features for tasks such as classification. But there has been less explo-ration in learning the factors of variation apart from the classification signal. Byaugmenting autoencoders with simple regularization terms during training, wedemonstrate that standard deep architectures can discover and explicitly repre-sent factors of variation beyond those relevant for categorization. We introducea cross-covariance penalty (XCov) as a method to disentangle factors like hand-writing style for digits and subject identity in faces. We demonstrate this on theMNIST handwritten digit database, the Toronto Faces Database (TFD) and theMulti-PIE dataset by generating manipulated instances of the data. Furthermore,we demonstrate these deep networks can extrapolate ‘hidden’ variation in the supervised signal.

Requirements

Installing the dependencies

Install virtualenv and creating a new virtual environment:

pip install virtualenv
virtualenv -p /usr/bin/python3 ./venv

Install dependencies

pip3 install -r requirements.txt

Disclaimer

This is not an official implementation.

Citation

If you find this code useful for your research, please cite:

@article{cheung2014discovering,
  title={Discovering hidden factors of variation in deep networks},
  author={Cheung, Brian and Livezey, Jesse A and Bansal, Arjun K and Olshausen, Bruno A},
  journal={arXiv preprint arXiv:1412.6583},
  year={2014}
}