Spiked Covariance Estimation from Modulo-Reduced Measurements
Abstract
Consider the rank-1 spiked model: , where is the spike intensity, is an unknown direction and . Motivated by recent advances in analog-to-digital conversion, we study the problem of recovering from i.i.d. modulo-reduced measurements , focusing on the high-dimensional regime (). We develop and analyze an algorithm that, for most directions and , estimates to high accuracy using measurements, provided that . Up to constants, our algorithm accurately estimates at the smallest possible that allows (in an information-theoretic sense) to recover from . A key step in our analysis involves estimating the probability that a line segment of length in a random direction passes near a point in the lattice . Numerical experiments show that the developed algorithm performs well even in a non-asymptotic setting.
Keywords
Cite
@article{arxiv.2110.01150,
title = {Spiked Covariance Estimation from Modulo-Reduced Measurements},
author = {Elad Romanov and Or Ordentlich},
journal= {arXiv preprint arXiv:2110.01150},
year = {2022}
}
Comments
AISTATS, 2022