Code for the paper: Wirth, E. S., & Pokutta, S. (2022, May). Conditional gradients for the approximately vanishing ideal. In Proceedings of the International Conference on Artificial Intelligence and Statistics (pp. 2191-2209). PMLR.

and

Wirth, E. and Pokutta, S., 2022. Conditional Gradients for the Approximate Vanishing Ideal. arXiv preprint arXiv:2202.03349.

This project is an extension of the previously published release and Git repository cgavi and avi_at_scale, respectively.

Download the repository and store it in your preferred location, say ~/tmp.

Open your terminal and navigate to ~/tmp.

Run the command:

$ conda env create --file environment.yml

This will create the conda environment cgavi.

Activate the conda environment with:

$ conda activate cgavi

Run the tests:

>>> python3 -m unittest

No errors should occur.

Execute the experiments:

>>> python3 experiments_cgavi.py

This will create folders named data_frames and plots, which contain subfolders containing the experiment results and the plots, respectively.

The performance experiments can be displayed as latex_code by executing:

>>> experiments_to_latex_cgavi.py