Searcharxiv⌕ Search

arXiv subjects

Alexia Yavicoli

Publications and source records attributed to Alexia Yavicoli.

At least 19 recordsLinked to original sources

Falconer's conjecture through descent and chaining

OpenAI proved that a compact set $E\subset{\mathbb R}^d$ of Hausdorff dimension greater than $\frac{d}{2}$ determines a set of Euclidean distances of positive length. We give a complete proof of this theorem which follows the overall architecture of the OpenAI argument, from the construction of the initial pair measures to the final Fourier estimate, but which modifies that argument substantially at the points where the geometric costs are paid and where the estimates are reassembled. The reductions are organized by the descent and chaining viewpoint of Burstein, Iosevich, and Krause: one potential, built from two separated depths together with a penalty for insufficient separation, pays for every reduction on every interval of scales, the terminal intervals included, and a finite-record chaining lemma keeps the fixed power losses attached to the few reductions which incur them. In odd dimensions the radial improvement is obtained directly at each finite scale from a quantitative contact interval, chosen by minimizing a mass curve against lines of two slopes, with explicit margins in place of a compactness argument, and the conditional estimates on the three parts of that interval are multiplied by a lemma on full source fibers. A random sign identity represents the coarse pair filter by products of endpoint multipliers and reduces the cost of reassembling the spatial pieces inside the shell estimate. The argument gives more than positive length. Since the mass first excluded near heavy hyperplanes at each late scale is summable with a power rate, all the Fourier estimates compare with one fixed measure, and the distance set supports a probability measure with an $L^p$ density for some $p>1$ and with power Fourier decay, so that $Δ(E)$ has positive Fourier dimension.

math.CA↗

The geometric case of the Erdős Similarity Conjecture. A human exposition of the OpenAI proof

For a fixed $q\in(0,1)$, can a set of positive Lebesgue measure avoid every translated and dilated copy of $\{q^n:n\ge1\}$? OpenAI's release of October 6, 2026 shows that it can, with a compact subset of $[0,1]$ whose measure is arbitrarily close to one. We give a detailed exposition of the proof, using only finite probability and basic measure theory. The proof removes a small open set intersecting every copy, obtained from a random construction: fair random choices, attached to the cells of nested grids, route each point down a finite tree, and at the leaf reached a further random choice removes the point with probability $p$. The removed set therefore has expected density $p$, however large the tree, while the branching gives each copy many independent chances to intersect it. Each chance depends on only part of the tree, so its outcome changes at only a controlled number of dilations, and a union bound over finitely many representative dilations then covers every dilation. We also relate the proof to the earlier random-cell arguments going back to Kolountzakis (Bull. London Math. Soc. 29 (1997), 415--424) and to recent work on two-fold sumsets (arXiv:2607.03584) and on Rajchman measures (arXiv:2609.04456), and we show how the method of conditional expectations makes the construction deterministic.

math.PR↗

Improved dimension bounds for Falconer's distance problem in the plane

Let $E\subset\mathbb R^2$ be analytic with $t=\dim_{\mathrm H}E>1$. We prove $\dim_{\mathrm H}ΔE\ge 4/5+η_0$ for an absolute, non-explicit constant $η_0>0$, and \[ \dim_{\mathrm H}ΔE\ge \frac{t^2-2t+5}{2t^2-7t+10} \quad\text{for }1 1$, then \[ \underline{\dim}_{\mathrm B}Δ_xE\ge \max\left\{\frac45+η_0,\min\left\{\frac{3s+1}{5},\frac78\right\}\right\} \] for $μ$-almost every $x$. The same bound holds for pinned Hausdorff dimension under a common dyadic mass law, allowing oscillation with scale. We retain information about both endpoints in the Kolmogorov complexity estimates of Stull and Fiedler--Stull. Applying the partition estimates with the endpoints reversed yields $4/5$ and the rational bound. Equality forces both information profiles to have excess information on a common interval of scales. Encoding one source measure for all pins converts this constraint into a configuration of points and circles with many rich perpendicular bisectors. Lower bounds of Katz--Tardos and Ren--Wang contradict the upper bound of Fu--Gan--Ren; compactness then gives $η_0>0$. This rigidity argument has a precedent in the work of Héra, Shmerkin and Yavicoli on Furstenberg sets, using Bourgain's projection theorem. A separate finite-scale selection proves the pinned lower box bound.

math.CA↗

Minkowski sums with convex curves without pointwise Fourier decay

Let $Γ\subset\mathbb R^2$ be a compact convex graph and define \[ T(Γ) = \inf \left\{ t: \dim_{\mathrm H}(E)>t \Longrightarrow |E+Γ|>0 \text{ for every compact }E\subset\mathbb R^2 \right\}. \] For a graph over an interval of positive length the smallest possible value is $T(Γ)=1$. We ask whether this optimal conclusion can hold when pointwise Fourier decay of arclength is unavailable. The answer is yes, even for strictly convex curves. We use the Fourier transform convention $\widehatν(ξ)=\int e^{-2πi x\cdotξ}\,dν(x)$. We construct a strictly convex Lipschitz graph $Γ$ with $T(Γ)=1$ such that, for every nontrivial subarc $Γ_0$ and every $α>0$, \[ \limsup_{|ξ|\to\infty} |ξ|^α\left| \widehat{H^1|_{Γ_0}}(ξ) \right|= \infty. \] We also give a convex example for which arclength on every nontrivial subarc fails even to be a Rajchman measure. The geometric mechanism behind these examples is a positive curved trace: if $Γ$ contains a positive-length subset of a $C^2$ curve whose curvature is bounded away from zero, then $|E+Γ|>0$ whenever $\dim_{\mathrm H}(E)>1$. For a nondegenerate graph this gives $T(Γ)=1$. For convex graphs it implies, in particular, that $T(Γ)=1$ whenever the curvature measure has a nonzero absolutely continuous part. The positive-measure proofs are in physical space and use translated-tube intersections and elementary facts about convex functions. The same overlap estimates give Mattila-type lower bounds for the average lengths of the associated curve projections of neighborhoods under the positive curved-trace hypothesis. We also prove a dimension-one endpoint result for sets with a positive-length rectifiable part and formulate the main remaining question: whether every strictly convex Lipschitz graph has the optimal threshold $T(Γ)=1$.

math.CA↗

Arithmetic-progression gap sets in Cantor sets

We address the question of which common differences can arise in arithmetic progressions contained in fractal sets. For a compact set $C\subset\mathbb R$, we investigate not only whether arithmetic progressions occur in $C$, but the full collection of their common differences. More generally, for a finite pattern $P$, we study the set of scales at which affine copies of $P$ appear in $C$. For affine self-similar sets satisfying strong separation, we obtain explicit restrictions on admissible common differences. Specializing to middle-$\varepsilon$ Cantor sets, we prove that the longest arithmetic progression has length four whenever $3-2\sqrt2<\varepsilon\le 1/3$, showing that the maximal progression length drops immediately from six at the critical parameter $\varepsilon=3-2\sqrt2$. We further develop recursive bounds for the sets of admissible common differences and derive explicit blackout intervals, namely ranges of scales for which arithmetic progressions cannot occur. On the positive side, sufficiently thick Cantor sets exhibit the opposite behavior. Combining a refinement of the Hunt-Kan-Yorke construction with the Newhouse Gap Lemma, we prove that every sufficiently small common difference occurs in a three-term arithmetic progression. In particular, if the largest bounded gap of a Cantor set is at most $0.067 diam(C)$ and its thickness is at least $6.96268\ldots$, then every common difference in $(0,0.435 diam(C)]$ occurs in a three-term arithmetic progression contained in $C$. Analogous interval results are obtained for four-term arithmetic progressions and asymmetric three-point patterns.

math.CA↗

Additive and multiplicative densities, prime valuations and symbolic models

We study additive and multiplicative densities of subsets of $\N$ along prescribed Følner sequences. We prove that additive upper density one implies multiplicative density one along a suitable multiplicative Følner sequence. We also prove independence in the following sense: given an additive Følner sequence $(G_n)_n$, a multiplicative Følner sequence $(F_n)_n$, and any $(α,β)\in[0,1]^2$, we construct a single set $A\subseteq\N$ such that $\dens_{(G_n)_n}(A)=α$ and $\md_{(F_n)_n}(A)=β$. Prime-valuation coordinates yield exact density formulas and random models for one local condition and, under summability assumptions, countably many; in the finite-coordinate case they also give exact higher-order correlations. Finally, we realize these multiplicative correlations as correlations in symbolic dynamical systems and obtain criteria for ergodicity and mixing.

math.NT↗

Intersections of thick compact sets in $\mathbb{R}^d$

We introduce a definition of thickness in $\mathbb{R}^d$ and obtain a lower bound for the Hausdorff dimension of the intersection of finitely or countably many thick compact sets using a variant of Schmidt's game. As an application we prove that given any compact set in $\mathbb{R}^d$ with thickness $τ$, there is a number $N(τ)$ such that the set contains a translate of all sufficiently small similar copies of every set in $\mathbb{R}^d$ with at most $N(τ)$ elements; indeed the set of such translations has positive Hausdorff dimension. We also prove a gap lemma and bounds relating Hausdorff dimension and thickness.

math.CA↗

Full measure universality for Cantor Sets

We investigate variants of the Erdős similarity problem for Cantor sets. We prove that under a mild Hausdorff or packing logarithmic dimension assumption, Cantor sets are not full measure universal, significantly improving the known fact that sets of positive Hausdorff dimension are not measure universal. We prove a weaker result for all Cantor sets $A$: there is a dense $G_δ$ set of full measure $X\subset\mathbb{R}^d$, such that for any bi-Lipschitz function $f:\mathbb{R}^d\to \mathbb{R}^d$, the set of translations $t$ such that $f(A)+t\subseteq X$ is of measure zero. Equivalently, there is a null set $B\subset\mathbb{R}^d$ such that $\mathbb{R}^d\setminus (f(A)+B)$ is null for all bi-Lipschitz functions $f$.

math.CA↗

Numbers omitting digits in certain base expansions

In DOI:10.1017/etds.2022.2 the author proved that for each integer $k$ there is an implicit number $M > 0$ such that if $b_1, \cdots , b_k$ are multiplicatively independent integers greater than $M$, there are infinitely many integers whose base $b_1, b_2, \cdots , b_k$ expansions all do not have zero digits. In this paper we don't require the multiplicative independence condition and make the result quantitative, getting an explicit value for $M$. We also obtain a result for the case when the missing digit(s) may not be zero. Finally, we extend our method to study various missing-digit sets in an algebraic setting.

math.NT↗

On the volumes of simplices determined by a subset of $\mathbb{R}^d$

We prove that for $1\le k<d$, if $E$ is a Borel subset of $\mathbb{R}^d$ of Hausdorff dimension strictly larger than $k$, the set of $(k+1)$-volumes determined by $k+2$ points in $E$ has positive one-dimensional Lebesgue measure. In the case $k=d-1$, we obtain an essentially sharp lower bound on the dimension of the set of tuples in $E$ generating a given volume. We also establish a finer version of the classical slicing theorem of Marstrand-Mattila in terms of dimension functions, and use it to extend our results to sets of ``dimension logarithmically larger than $k$''.

math.CA↗

Thickness and a gap lemma in $\mathbb{R}^d$

We give a definition of thickness in $\mathbb{R}^d$ that is useful even for totally disconnected sets, and prove a Gap Lemma type result. We also guarantee an interval of distances in any direction in thick compact sets, relate thick sets (for this definition of thickness) with winning sets, give a lower bound for the Hausdorff dimension of the intersection of countably many of them, a result guaranteeing the presence of large patterns, and lower bounds for the Hausdorff dimension of a set in relationship with its thickness.

math.CA↗

A survey on Newhouse thickness, fractal intersections and patterns

In this article, we introduce a notion of size for sets called thickness that can be used to guarantee that two Cantor sets intersect (the Gap Lemma), and show a connection among Thickness, Schmidt Games and Patterns. We work mostly in the real line, but we also introduce the topic in higher dimensions.

math.AP↗

The density of sets containing large similar copies of finite sets

We prove that if $E \subseteq \mathbb{R}^d$ ($d\geq 2$) is a Lebesgue-measurable set with density larger than $\frac{n-2}{n-1}$, then $E$ contains similar copies of every $n$-point set $P$ at all sufficiently large scales. Moreover, `sufficiently large' can be taken to be uniform over all $P$ with prescribed size, minimum separation and diameter. On the other hand, we construct an example to show that the density required to guarantee all large similar copies of $n$-point sets tends to $1$ at a rate $1- O(n^{-1/5}\log n)$.

math.CA↗

Improved bounds on the dimensions of sets that avoid approximate arithmetic progressions

We provide quantitative estimates for the supremum of the Hausdorff dimension of sets in the real line which avoid $\varepsilon$-approximations of arithmetic progressions. Some of these estimates are in terms of Szemerédi bounds. In particular, we answer a question of Fraser, Saito and Yu (IMRN, 2019) and considerably improve their bounds. We also show that Hausdorff dimension is equivalent to box or Assouad dimension for this problem, and obtain a lower bound for Fourier dimension.

math.CA↗

Patterns in thick compact sets

We introduce a connection between Newhouse thickness and patterns through a variant of Schmidt's game introduced by Broderick, Fishman and Simmons. This yields an explicit, robust and checkable condition that ensures the presence of patterns in compact sets, in particular in Cantor sets.

math.CA↗

An improved bound for the dimension of $(α,2α)$-Furstenberg sets

We show that given $α\in (0, 1)$ there is a constant $c=c(α) > 0$ such that any planar $(α, 2α)$-Furstenberg set has Hausdorff dimension at least $2α+ c$. This improves several previous bounds, in particular extending a result of Katz-Tao and Bourgain. We follow the Katz-Tao approach with suitable changes, along the way clarifying, simplifying and/or quantifying many of the steps.

math.CA↗

Small Sets containing any Pattern

Given any dimension function $h$, we construct a perfect set $E \subseteq \mathbb{R}$ of zero $h$-Hausdorff measure, that contains any finite polynomial pattern. This is achieved as a special case of a more general construction in which we have a family of functions $\mathcal{F}$ that satisfy certain conditions and we construct a perfect set $E$ in $\mathbb{R}^N$, of $h$-Hausdorff measure zero, such that for any finite set $\{ f_1,\ldots,f_n\}\subseteq \mathcal{F}$, $E$ satisfies that $\bigcap_{i=1}^n f^{-1}_i(E)\neq\emptyset$. We also obtain an analogous result for the images of functions. Additionally we prove some related results for countable (not necessarily finite) intersections, obtaining, instead of a perfect set, an $\mathcal{F}_σ$ set without isolated points.

math.CA↗

Large sets avoiding linear patterns

We prove that for any dimension function $h$ with $h \prec x^d$ and for any countable set of linear patterns, there exists a compact set $E$ with $\mathcal{H}^h(E)>0$ avoiding all the given patterns. We also give several applications and recover results of Keleti, Maga, and Máthé.

math.CA↗