Affiliation:
1. University of California Berkeley Statistics, Evans Hall, Berkeley, CA
2. University of California Berkeley Mathematics, Evans Hall, Berkeley CA
Abstract
Abstract
We prove improved bounds on how localized an eigenvector of a high girth regular graph can be, and present examples showing that these bounds are close to sharp. This study was initiated by Brooks and Lindenstrauss [6] who relied on the observation that certain suitably normalized averaging operators o nhigh girth graphs are hyper-contractive and can be used to approximate projectors onto the eigenspaces of such graphs. Informally, their delocalization result in the contrapositive states that for any $\varepsilon \in (0,1)$ and positive integer $k,$ if a $(d+1)-$regular graph has an eigenvector that supports $\varepsilon $ fraction of the $\ell _2^2$ mass on a subset of $k$ vertices, then the graph must have a cycle of size $\log _{d}(k)/\varepsilon ^2)$, up to multiplicative universal constants and additive logarithmic terms in $1/\varepsilon $. In this paper, we improve the upper bound to $\log _{d}(k)/\varepsilon $ up to similar logarithmic correction terms; and present a construction showing a lower bound of $\log _d(k)/\varepsilon $ up to multiplicative constants. Our construction is probabilistic and involves gluing together a pair of trees while maintaining high girth as well as control on the eigenvectors and could be of independent interest.
Funder
Miller Research Fellowship
National Science Foundation
Sloan Research Fellowship
Publisher
Oxford University Press (OUP)
Reference19 articles.
1. Non-backtracking random walks mix faster;Alon;Commun. Contemp. Math.,2007
2. Quantum ergodicity on large regular graphs;Anantharaman;Duke Math. J.,2015
3. Local Kesten–Mckay law for random regular graphs.;Bauerschmidt,2016
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献