Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.36917

The sharp exponent for the Banach--Mazur distance to the cube

Omer Friedland
View original source ↗Read source document ↗Export metadata ↓

Abstract

Let $R_\infty(n)$ be the maximal Banach--Mazur distance from an $n$-dimensional real normed space to $\ell_\infty^n$. We prove that $R_\infty(n)\ge cn^{2/3}(\log(2n))^{-10}$. Together with the upper bound of Friedland, Naor, and Youssef, this gives $R_\infty(n) = n^{2/3+o(1)}$. The examples are polars of symmetric Gaussian polytopes with $2n$ generators, and the lower bound holds with exponentially high probability. A joint minimization over coefficient representations and positive weights yields an exact second-variation identity and sparse coefficients. A Gaussian projection argument controls all low-rank completions of the recorded matrix. An adaptive choice of exceptional generators then reduces the remaining case to a fixed-shift Gaussian small-ball estimate over two finite nets.

Explore related subjects

math.FA

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Omer Friedland. 2026-10-27. The sharp exponent for the Banach--Mazur distance to the cube. https://arxiv.org/abs/2610.36917

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

A Framework for Supervised and Unsupervised Learning via Reproducing Kernel Hilbert Spaces

We introduce a unified framework for supervised and unsupervised learning through density estimation using reproducing kernel Hilbert spaces (RKHSs). To this end, we introduce the notion of a nested RKHS, that is, an increasing sequence of RKHSs whose union is dense in an ambient $L^2$-space. This construction allows us to exploit the computational structure of RKHS methods at each finite approximation level while preserving the approximation power of the entire $L^2$-space. We first study statistical learning in a fixed RKHS. In both unsupervised and supervised settings, we establish strong convergence results for empirical averages of random elements taking values in an RKHS continuously embedded in $L^2$. We then use the nested RKHS structure to combine these statistical convergence results with the density of the union of the RKHSs in the ambient $L^2$-space. We apply this construction to probability density estimation and binary Bayesian classification. For density estimation, we construct explicit empirical estimators of the orthogonal projections of the target density onto the RKHSs and establish their almost sure convergence. A diagonal selection procedure then yields a sequence of estimators converging almost surely to the target density in $L^2$. For binary Bayesian classification, we formulate the approximation of the Bayes target as the minimization of suitable population and empirical risk functionals. We prove consistency of the empirical minimizers and use the nested structure to obtain convergence towards the Bayes target in $L^2$. We further provide a geometric interpretation in terms of hyperplanes and weighted distances in the ambient $L^2$-space. Finally, we show that the population risk admits a direct probabilistic interpretation in terms of the probability of misclassification.

math.FA↗
arXiv2026

From one particle to N particles: uniform transport-information and log-Sobolev inequalities for Coulomb gases on the torus

We study how a functional inequality satisfied by the one-particle conditional distributions of a Gibbs measure, with a constant independent of the positions of the other particles, can be propagated to the whole N -particle system, with a constant independent of N , when the interaction is singular. Our model is the Coulomb gas on the torus T 3 in the mean-field regime. We distinguish two mechanisms. The first one is a dependent tensorization, governed by Dobrushin-Wu interdependence coefficients. It yields uniform Poincar{é} and transport-information W 2 I inequalities with fully explicit constants. We explain why this mechanism does not extend to the log-Sobolev inequality: the Zegarlinski-type coefficients, which require pointwise bounds on the interaction, are infinite. The second mechanism is a tensorization with a defect. Using a uniform bound on the partition function due to Duerinckx and Jabin, it yields a uniform log-Sobolev inequality for the Coulomb gas on T 3 at small charge. We identify the assumptions on the interaction under which each mechanism applies. All the results are then extended to the two-dimensional Coulomb gas and to periodic Riesz gases 1/|x|^s , 0 < s $\le$ 1. The extension to d $\ge$ 4 is also discussed.

math.FA↗