On Furstenberg's intersection conjecture, self-similar measures, and the L^q norms of convolutions

DOI10.4007/ANNALS.2019.189.2.1MaRDI QIDQ1732984zbMATH OpenOpenAlexWikidataFDO

Publication date 26 March 2019

Published in Annals of Mathematics. Second Series (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1609.07802

$\times p$-invariant sets Bernoulli convolutions dynamical rigidity intersections of Cantor sets self-similar measures

Mathematics Subject Classification ID

Metric theory of other algorithms and expansions; measure and Hausdorff dimension (11K55) Fractals (28A80) Dimension theory of smooth dynamical systems (37C45)

Abstract: We study a class of measures on the real line with a kind of self-similar structure, which we call dynamically driven self-similar measures, and contain proper self-similar measures such as Bernoulli convolutions as special cases. Our main result gives an expression for the

L^{q}

-dimensions of such dynamically driven self-similar measures, under certain conditions. As an application, we settle Furstenberg's long-standing conjecture on the dimension of the intersections of

i m e s p

and

i m e s q

-invariant sets. Among several other applications, we also show that Bernoulli convolutions have an

L^{q}

density for all finite

q

, outside of a zero-dimensional set of exceptions. The proof of the main result is inspired by M. Hochman's approach to the dimensions of self-similar measures and his inverse theorem for entropy. Our method can be seen as an extension of Hochman's theory from entropy to

L^{q}

norms, and likewise relies on an inverse theorem for the decay of

L^{q}

norms of discrete measures under convolution. This central piece of our approach may be of independent interest, and is an application of well-known methods and results in additive combinatorics: the asymmetric version of the Balog-Szemer'{e}di-Gowers Theorem due to Tao-Vu, and some constructions of Bourgain.

Review text

In this paper, a class of measures on the real line with a kind of self-similar structure is studied, which are called dynamically driven self-similar measures, and proper self-similar measures such as Bernoulli convolutions which are contained as a special case. The strong version of Furstenberg's conjecture [\textit{H. Furstenberg}, Math. Syst. Theory 1, 1--49 (1967; Zbl 0146.28502); Probl. Analysis, Sympos. in Honor of Salomon Bochner, Princeton Univ. 1969, 41--59 (1970; Zbl 0208.32203)] is proved, which, in his terminology, says that the maps $T_p$ and $T_q$ on the circle are strongly transverse. The following is the main result of the paper Let $(X;\mathbf{T};\Delta;\lambda)$ be a pleasant model with exponential separation, and denote the unique invariant measure by $\mathbb P$. Assume further that the map $x\mapsto\delta(x)$ is continuous $\mathbb P$-almost everywhere, and the number of atoms of $\Delta(x)$ is uniformly bounded. Then for all $q\in(1,+\infty)$, \[ \lim_{m\to\infty}-\frac{\log\sum_{I\in\mathcal D_m}\mu_x(I)^q}{(q-1)m}=\min\left(\frac{\int_X\log\|\Delta(x)\|_q^q d\mathbb P(x)}{(q-1)\log\lambda},1\right), \] uniformly in $x\in X$. That is, the limit in the definition of $L^q$ dimension of $\mu_x$ exists and equals the constant value on the right-hand side for all $x\in X$. The overall strategy is inspired by the ideas of the deep work of \textit{M. Hochman} [Ann. of Math. (2) 180 no. 2 (2014), 773--822. Zbl 1337.28015]. Additional complications are caused by the fact that our model allows measures that are not strictly self-similar; this is dealt with by the help of a cocycle. The key difference is that Hochman's method is based on entropy, while this paper needs to deal with $L^q$ norms. This forces substantial changes in the implementation of the general strategy. Sections 2--5 are devoted to the proof of the main theorem. In Section 2, the inverse theorem for the $L^q$ norms of convolutions of discrete measures is stated and discussed. The inverse theorem is proved in Section 3. The two main tools in the proof of the inverse theorem come from additive combinatorics: an asymmetric version of the Balog-Szemerédi-Gowers Theorem, due to Tao and Vu, and a structure result due to Bourgain. Section 4 develops some properties of dynamically driven self-similar measures. In Section 5, these are combined with the inverse theorem to conclude the proof of the main theorem. The remaining Sections 6--9 contain the applications of the main theorem. In Section 6 the main theorem is applied to study $L^q$ dimensions and Frostman exponents of self-similar measures on the line. In particular, the theorem which generalizes to homogeneous self-similar measures on $R$, and the theorem which extends this to arbitrary self-similar measures on the line (not necessarily homogeneous) are proved. In Section 7, the $L^q$ dimensions of convolutions of self-similar measures are studied and a variant of Furstenberg's conjecture for self-similar sets is deduced. Section 8 contains further applications of the main theorem to projections and sections of planar self-similar sets and measures. In particular, an upper bound for the dimensions of arbitrary linear sections of some self-similar sets on the plane is proved. Finally, in Section 9 focusing on the densities of the measures studied in the previous sections, a general result in the framework of dynamically defined measures is presented, as well as several other applications.

Recommendations

Cites work

Cited in

(only showing first 100 items - show all)

This page was built for publication: On Furstenberg's intersection conjecture, self-similar measures, and the $L^q$ norms of convolutions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1732984)