Robust Sparse Mean Estimation

A python implementation of algorithms in the paper https://arxiv.org/abs/1911.08085.

Prerequisites

• Jupyter

Explanation of Files

• robustlib.py: Library containing code for robust mean estimation and robust PCA algorithms as well as to plot comparisons between the performance of these algorithms across various noise models.

• experiments_notebook.ipynb: A Jupyter notebook with code to reproduce plots that are seen in the paper.

Reference

This repository is an implementation of our paper "Outlier-Robust High-Dimensional Sparse Estimation via Iterative Filtering", authored by Ilias Diakonikolas, Daniel M. Kane, Sushrut Karmalkar, Eric Price and Alistair Stewart.

If you use the code for our paper, we ask that you please cite

@inproceedings{diakonikolas2019outlier,
  title={Outlier-robust high-dimensional sparse estimation via iterative filtering},
  author={Diakonikolas, Ilias and Kane, Daniel and Karmalkar, Sushrut and Price, Eric and Stewart, Alistair},
  booktitle={Advances in Neural Information Processing Systems},
  pages={10689--10700},
  year={2019}
}