CCM

Code for replicating the experiments in the paper Kernel Feature Selection via Conditional Covariance Minimization by Jianbo Chen*, Mitchell Stern*, Martin J. Wainwright, Michael I. Jordan. (* indicates equal contribution)

Dependencies

The code for CCM runs with Python and requires Tensorflow of version 1.2.1 or higher. Please pip install the following packages:

• numpy
• tensorflow

Or you may run the following and in shell to install the required packages:

git clone https://github.com/Jianbo-Lab/CCM
cd CCM
sudo pip install -r requirements.txt

Running in Docker, MacOS or Ubuntu

We provide as an example the source code to run CCM on the three synthetic datasets in the paper. Run the following commands in shell:

###############################################
# Omit if already git cloned.
git clone https://github.com/Jianbo-Lab/CCM
cd CCM 
############################################### 
python examples/run_synthetic.py

See core/ccm.py and examples/run_synthetic.py for details.

Citation

If you use this code for your research, please cite our paper:

@incollection{NIPS2017_7270,
title = {Kernel Feature Selection via Conditional Covariance Minimization},
author = {Chen, Jianbo and Stern, Mitchell and Wainwright, Martin J and Jordan, Michael I},
booktitle = {Advances in Neural Information Processing Systems 30},
editor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},
pages = {6949--6958},
year = {2017},
publisher = {Curran Associates, Inc.},
url = {http://papers.nips.cc/paper/7270-kernel-feature-selection-via-conditional-covariance-minimization.pdf}
}