/icml-rmt-2024

Primary LanguageTeX

Fréchet-Karcher Mean with low sample support: exploiting random matrix theory and Riemannian geometry

This repository contains the code and source for the following paper accepted at ICML 2024:

Random matrix theory improved Frechet mean of symmetric positive definite matrices, Florent Bouchard, Ammar Mian, Malik Tiomoko, Guillaume Ginolhac, Frederic Pascal

Preprint available at: https://arxiv.org/abs/2405.06558

It is organised as follows:

  • code/ contains the code used to produce the numerical and real-data experiments presented in the paper
  • paper contains the source of the paper (LaTeX files) to compile the paper (file main.tex)

If you use any code in this repository please cite us using:

@misc{bouchard2024random, title={Random matrix theory improved Fr'echet mean of symmetric positive definite matrices}, author={Florent Bouchard and Ammar Mian and Malik Tiomoko and Guillaume Ginolhac and Frédéric Pascal}, year={2024}, eprint={2405.06558}, archivePrefix={arXiv}, primaryClass={stat.ML} }