Infinite Lewis Weights in Spectral Graph Theory
Abstract
We study the spectral implications of re-weighting a graph by the -Lewis weights of its edges. Our main motivation is the ER-Minimization problem (Saberi et al., SIAM'08): Given an undirected graph , the goal is to find positive normalized edge-weights which minimize the sum of pairwise \emph{effective-resistances} of (Kirchhoff's index). By contrast, -Lewis weights minimize the \emph{maximum} effective-resistance of \emph{edges}, but are much cheaper to approximate, especially for Laplacians. With this algorithmic motivation, we study the ER-approximation ratio obtained by Lewis weights. Our first main result is that -Lewis weights provide a constant () approximation for ER-minimization on \emph{trees}. The proof introduces a new technique, a local polarization process for effective-resistances (-congestion) on trees, which is of independent interest in electrical network analysis. For general graphs, we prove an upper bound on the approximation ratio obtained by Lewis weights, which is always , where is the condition number of the weighted Laplacian. All our approximation algorithms run in \emph{input-sparsity} time , a major improvement over Saberi et al.'s SDP for exact ER-minimization. Finally, we demonstrate the favorable effects of -LW reweighting on the \emph{spectral-gap} of graphs and on their \emph{spectral-thinness} (Anari and Gharan, 2015). En-route to our results, we prove a weighted analogue of Mohar's classical bound on , and provide a new characterization of leverage-scores of a matrix, as the gradient (w.r.t weights) of the volume of the enclosing ellipsoid.
Keywords
Cite
@article{arxiv.2302.05966,
title = {Infinite Lewis Weights in Spectral Graph Theory},
author = {Amit Suliman and Omri Weinstein},
journal= {arXiv preprint arXiv:2302.05966},
year = {2023}
}