arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2610.05404

A Fraction Dimensional Kakeya Compression

Jiahan Du
View original source ↗Read source document ↗Export metadata ↓

Abstract

Recently, it was shown in arXiv:2607.27629 that under a very general condition, a generic family of curved Kakeya sets in $\mathbb{R}^3$ not trivially contained in a hypersurface has Hausdorff dimension strictly greater than 2. In this paper we leverage large language models to give an explicit construction of a family of curves failing the worst compression phenomenon in arXiv:2607.27629, but whose union has Hausdorff dimension and upper box dimension $5/2$. The example shows that there are Kakeya compression phenomena besides the trivial example of concentrating near a hypersurface. This answers a question raised in section 1.9 of arXiv:2607.27629. The models used are ChatGPT 5.6 Sol and ChatGPT 6 Astra.

Explore related subjects

math.CA

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jiahan Du. 2026-10-04. A Fraction Dimensional Kakeya Compression. https://arxiv.org/abs/2610.05404

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

KEEP EXPLORING

Related papers

arXiv2026

Positive-measure intersections of Bedford-McMullen carpets under similitudes

Let $K=K(n,m,Γ)$ be a Bedford-McMullen carpet with $n>m\geq2$ and $1<\#Γ 0$, then its orthogonal part is diagonal or anti-diagonal. This holds without any arithmetic restrictions on the defining bases. For isometries, we further prove that if $μ(K\cap f(K))>0$ then $f(K)$ and $K$ differ only by a translation, thereby establishing a strengthened form of the 2010 conjecture of Elekes, Keleti and Máthé in this planar self-affine context.

math.CA↗
arXiv2026

Sparse domination modulo polynomials

We develop a sparse domination theory that captures both local polynomial approximation and cancellation of the input below the scale of each sparse cube. Its coefficients are percentiles of localized smooth maximal functions of the input minus its local polynomial approximation, combining features of mean-oscillation sparse bounds and of the cancellative sparse bounds of Conde-Alonso, Lorist and Rey. Using wavelet representations, we characterize the Calderón-Zygmund operators whose derivatives admit these bounds by explicit polynomial testing conditions. In the homogeneous setting, the same conditions characterize weighted Hardy-Sobolev inequalities with linear dependence on the $A_\infty$ characteristic. This quantitative dependence detects the transition from BMO to $L^\infty$ testing conditions that Lerner found for pointwise bounds by Riesz potentials, a distinction that qualitative weighted estimates cannot recover. In the inhomogeneous setting, we prove corresponding weighted inequalities and show that they are strictly weaker than pointwise sparse domination from the second derivative order onward. Consequences include Hardy-Sobolev estimates below the Banach range, sharp weighted Sobolev inequalities, and pointwise bounds by Riesz potentials.

math.CA↗
arXiv2026

On an infinitely divisible distribution involving modified Bessel functions

Mourad E.H. Ismail and Kenneth S. Miller in 1982 conjectured that for $ν>μ\geq0$ and $b>a>0$ the function $$x\mapsto G(x;μ,ν)=\left(\frac ba\right)^{μ-ν}\frac{I_μ(a\sqrt{x})I_ν(b\sqrt{x})}{I_μ(b\sqrt{x})I_ν(a\sqrt{x})},$$ where $I_ν$ stands for the modified Bessel function of the first kind, is the Laplace transform of an infinitely divisible probability distribution. Their conjecture is equivalent to the monotonicity property of some exponential sums over the zeros $j_{ν,n}$ with respect to the order, where $j_{ν,n}$ denotes the $n$th positive zero of the Bessel function of the first kind of order $ν$. In this paper our aim is to reduce the problem further to the monotonicity, with respect to $ν$, of \[Q_ν(t)=\sum_{n\geq1}j_{ν,n}^2e^{-t j_{ν,n}^2},\qquad t>0,\] and to prove this monotonicity. In this way we show that indeed the above function $x\mapsto G(x;μ,ν)$ is the Laplace transform of an infinitely divisible probability distribution, and the above conjecture is true. The key ingredient in our proof is an integral representation for the derivative of a quotient of modified Bessel functions of the first kind with respect to the order, combined with Bernstein's theorem and a probability tail argument.

math.CA↗