-
On the regularity of time-delayed embeddings with self-intersections
Authors:
Adam Śpiewak
Abstract:
We study regularity of the time-delayed coordinate maps \[φ_{h,k}(x) = (h(x), h(Tx), \ldots, h(T^{k-1}x))\] for a diffeomorphism $T$ of a compact manifold $M$ and smooth observables $h$ on $M$. Takens' embedding theorem shows that if $k > 2\dim M$, then $φ_{h,k}$ is an embedding for typical $h$. We consider the probabilistic case, where for a given probability measure $μ$ on $M$ one allows self-in…
▽ More
We study regularity of the time-delayed coordinate maps \[φ_{h,k}(x) = (h(x), h(Tx), \ldots, h(T^{k-1}x))\] for a diffeomorphism $T$ of a compact manifold $M$ and smooth observables $h$ on $M$. Takens' embedding theorem shows that if $k > 2\dim M$, then $φ_{h,k}$ is an embedding for typical $h$. We consider the probabilistic case, where for a given probability measure $μ$ on $M$ one allows self-intersections in the time-delayed embedding to occur along a zero-measure set. We show that if $k \geq \dim M$ and $k > \dim_H(\text{supp} μ)$, then for a typical observable, $φ_{h,k}$ is injective on a full-measure set with a pointwise Lipschitz inverse. If moreover $k > \dim M$, then $φ_{h,k}$ is a local diffeomorphism at almost every point. As an application, we show that if $k > \dim M$, then the Lyapunov exponents of the original system can be approximated with arbitrary precision by almost every orbit in the time-delayed model of the system. We also give almost sure pointwise bounds on the prediction error and provide a non-dynamical analogue of the main result, which can be seen as a probabilistic version of Whitney's embedding theorem.
△ Less
Submitted 10 May, 2025;
originally announced May 2025.
-
Universal projection theorems with applications to multifractal analysis and the dimension of every ergodic measure on self-conformal sets simultaneously
Authors:
Balázs Bárány,
Károly Simon,
Adam Śpiewak
Abstract:
We prove a universal projection theorem, giving conditions on a parametrized family of maps $Π_λ: X \to \mathbb{R}^d$ and a collection $M$ of measures on $X$ under which for almost every $λ$ equality $\mathrm{dim}_H Π_λμ= \min\{d, \mathrm{dim}_H μ\}$ holds for all measures $μ\in M$ simultaneously (i.e. on a full-measure set of $λ$'s independent of $μ$). We require $Π_λ$ to satisfy a transversality…
▽ More
We prove a universal projection theorem, giving conditions on a parametrized family of maps $Π_λ: X \to \mathbb{R}^d$ and a collection $M$ of measures on $X$ under which for almost every $λ$ equality $\mathrm{dim}_H Π_λμ= \min\{d, \mathrm{dim}_H μ\}$ holds for all measures $μ\in M$ simultaneously (i.e. on a full-measure set of $λ$'s independent of $μ$). We require $Π_λ$ to satisfy a transversality condition and $M$ to satisfy a new condition called relative dimension separability. We also prove that if the Assouad dimension of $X$ is smaller than d, then for almost every $λ$, projection $Π_λ$ is nearly bi-Lipschitz at $μ$-a.e. $x$, for all measures $μ\in M$ simultaneously. Our setting can include families of orthogonal projections, natural projections for conformal, non-autonomous or random iterated functions systems.
As an application, we prove that for a parametrized family of contracting conformal IFS with natural projections $Π_λ$ satisfying the transversality condition, for almost every parameter $λ$ one has $\mathrm{dim}_H Π_λμ= \min\{ d, \frac{h(μ)}{χ(λ, μ)} \}$ for all ergodic shift-invariant measures simultaneously.
We also prove that for self-similar systems on the line with similarity dimension smaller than one, for Lebesgue almost every choice of translations the multifractal formalism holds simultaneously on the full spectrum interval $\left[\min \frac{\log p_i}{\log|λ_i|},\max \frac{\log p_i}{\log|λ_i|}\right]$ for every self-similar measure.
We prove that the dimension part of the Marstrand-Mattila projection theorem holds simultaneously for the collection of all ergodic measures on a self-conformal set with the strong separation condition and without any separation for the collection of all Gibbs measures.
△ Less
Submitted 7 February, 2025; v1 submitted 4 December, 2024;
originally announced December 2024.
-
Typical dimension and absolute continuity for classes of dynamically defined measures, Part II : exposition and extensions
Authors:
Balázs Bárány,
Károly Simon,
Boris Solomyak,
Adam Śpiewak
Abstract:
This paper is partly an exposition, and partly an extension of our work [1] to the multiparameter case. We consider certain classes of parametrized dynamically defined measures. These are push-forwards, under the natural projection, of ergodic measures for parametrized families of smooth iterated function systems (IFS) on the line. Under some assumptions, most crucially, a transversality condition…
▽ More
This paper is partly an exposition, and partly an extension of our work [1] to the multiparameter case. We consider certain classes of parametrized dynamically defined measures. These are push-forwards, under the natural projection, of ergodic measures for parametrized families of smooth iterated function systems (IFS) on the line. Under some assumptions, most crucially, a transversality condition, we obtain formulas for the Hausdorff dimension of the measure and absolute continuity for almost every parameter in the appropriate parameter region. The main novelty of [1] and the present paper is that not only the IFS, but also the ergodic measure in the symbolic space, whose push-forward we consider, depends on the parameter. This includes many interesting families of measures, in particular, invariant measures for IFS's with place-dependent probabilities and natural (equilibrium) measures for smooth IFS's. One of the goals of this paper is to present an exposition of [1] in a more reader-friendly way, emphasizing the ideas and proof strategies, but omitting the more technical parts. This exposition/survey is based in part on the series of lectures by Károly Simon at the Summer School "Dynamics and Fractals" in 2023 at the Banach Center, Warsaw. The main new feature, compared to [1], is that we consider multi-parameter families; in other words, the set of parameters is allowed to be multi-dimensional. This broadens the scope of applications. A new application considered here is to a class of Furstenberg-like measures.
[1] B. Bárány, K. Simon, B. Solomyak and A. Śpiewak: Typical absolute continuity for classes of dynamically defined measures. Advances in Mathematics, Volume 399, 2022, 108258, ISSN 0001-8708, https://doi.org/10.1016/j.aim.2022.108258.
△ Less
Submitted 10 May, 2024;
originally announced May 2024.
-
Predicting dynamical systems with too few time-delay measurements: error estimates
Authors:
Krzysztof Barański,
Yonatan Gutman,
Adam Śpiewak
Abstract:
We study the problem of reconstructing and predicting the future of a dynamical system by the use of time-delay measurements of typical observables. Considering the case of too few measurements, we prove that for Lipschitz systems on compact sets in Euclidean spaces, equipped with an invariant Borel probability measure $μ$ of Hausdorff dimension $d$, one needs at least $d$ measurements of a typica…
▽ More
We study the problem of reconstructing and predicting the future of a dynamical system by the use of time-delay measurements of typical observables. Considering the case of too few measurements, we prove that for Lipschitz systems on compact sets in Euclidean spaces, equipped with an invariant Borel probability measure $μ$ of Hausdorff dimension $d$, one needs at least $d$ measurements of a typical (prevalent) Lipschitz observable for $μ$-almost sure reconstruction and prediction. Consequently, the Hausdorff dimension of $μ$ is the precise threshold for the minimal delay (embedding) dimension for such systems in a probabilistic setting. Furthermore, we establish a lower bound postulated in the Schroer--Sauer--Ott--Yorke prediction error conjecture from 1998, after necessary modifications (whereas the upper estimates were obtained in our previous work). To this aim, we prove a general theorem on the dimensions of conditional measures of $μ$ with respect to time-delay coordinate maps.
△ Less
Submitted 28 January, 2024;
originally announced January 2024.
-
Regularity of almost-surely injective projections in Euclidean spaces
Authors:
Krzysztof Barański,
Yonatan Gutman,
Adam Śpiewak
Abstract:
In a previous work we proved that if a finite Borel measure $μ$ in a Euclidean space has Hausdorff dimension smaller than a positive integer $k$, then the orthogonal projection onto almost every $k$-dimensional linear subspace is injective on a set of full $μ$-measure. In this paper we study the regularity of the inverses of these projections and prove that if $μ$ has a compact support $X$ such th…
▽ More
In a previous work we proved that if a finite Borel measure $μ$ in a Euclidean space has Hausdorff dimension smaller than a positive integer $k$, then the orthogonal projection onto almost every $k$-dimensional linear subspace is injective on a set of full $μ$-measure. In this paper we study the regularity of the inverses of these projections and prove that if $μ$ has a compact support $X$ such that (respectively) the Hausdorff, upper box-counting or Assouad dimension of $X$ is smaller than $k$, then the inverse is (respectively) continuous, pointwise $α$-Hölder for some $α\in (0,1)$ or pointwise $α$-Hölder for every $α\in (0,1)$. The results generalize to the case of typical linear perturbations of Lipschitz maps and strengthen previously known ones in the lossless analog compression literature. We provide examples showing the sharpness of the statements. Additionally, we construct a non-trivial measure on the plane which admits almost-surely injective projections in every direction, and show that no homogeneous self-similar measure has this property.
△ Less
Submitted 23 June, 2023; v1 submitted 27 January, 2023;
originally announced January 2023.
-
Absolute continuity of self-similar measures on the plane
Authors:
Boris Solomyak,
Adam Śpiewak
Abstract:
Consider an iterated function system consisting of similarities on the complex plane of the form $g_{i}(z) = λ_i z + t_i,\ λ_i, t_i \in \mathbb{C},\ |λ_i|<1, i=1,\ldots, k$. We prove that for almost every choice of $(λ_1, \ldots, λ_k)$ in the super-critical region (with fixed translations and probabilities), the corresponding self-similar measure is absolutely continuous. This extends results of S…
▽ More
Consider an iterated function system consisting of similarities on the complex plane of the form $g_{i}(z) = λ_i z + t_i,\ λ_i, t_i \in \mathbb{C},\ |λ_i|<1, i=1,\ldots, k$. We prove that for almost every choice of $(λ_1, \ldots, λ_k)$ in the super-critical region (with fixed translations and probabilities), the corresponding self-similar measure is absolutely continuous. This extends results of Shmerkin-Solomyak (in the homogenous case) and Saglietti-Shmerkin-Solomyak (in the one-dimensional non-homogeneous case). As the main steps of the proof, we obtain results on the dimension and power Fourier decay of random self-similar measures on the plane, which may be of independent interest.
△ Less
Submitted 30 August, 2023; v1 submitted 25 January, 2023;
originally announced January 2023.
-
Prediction of dynamical systems from time-delayed measurements with self-intersections
Authors:
Krzysztof Barański,
Yonatan Gutman,
Adam Śpiewak
Abstract:
In the context of predicting the behaviour of chaotic systems, Schroer, Sauer, Ott and Yorke conjectured in 1998 that if a dynamical system defined by a smooth diffeomorphism $T$ of a Riemannian manifold $X$ admits an attractor with a natural measure $μ$ of information dimension smaller than $k$, then $k$ time-delayed measurements of a one-dimensional observable $h$ are generically sufficient for…
▽ More
In the context of predicting the behaviour of chaotic systems, Schroer, Sauer, Ott and Yorke conjectured in 1998 that if a dynamical system defined by a smooth diffeomorphism $T$ of a Riemannian manifold $X$ admits an attractor with a natural measure $μ$ of information dimension smaller than $k$, then $k$ time-delayed measurements of a one-dimensional observable $h$ are generically sufficient for $μ$-almost sure prediction of future measurements of $h$. In a previous paper we established this conjecture in the setup of injective Lipschitz transformations $T$ of a compact set $X$ in Euclidean space with an ergodic $T$-invariant Borel probability measure $μ$. In this paper we prove the conjecture for all (also non-invertible) Lipschitz systems on compact sets with an arbitrary Borel probability measure, and establish an upper bound for the decay rate of the measure of the set of points where the prediction is subpar. This partially confirms a second conjecture by Schroer, Sauer, Ott and Yorke related to empirical prediction algorithms as well as algorithms estimating the dimension and number of required delayed measurements (the so-called embedding dimension) of an observed system. We also prove general time-delay prediction theorems for locally Lipschitz or Hölder systems on Borel sets in Euclidean space.
△ Less
Submitted 15 October, 2024; v1 submitted 27 December, 2022;
originally announced December 2022.
-
On the dimension of stationary measures for random piecewise affine interval homeomorphisms
Authors:
Krzysztof Barański,
Adam Śpiewak
Abstract:
We study stationary measures for iterated function systems (considered as random dynamical systems) consisting of two piecewise affine interval homeomorphisms, called Alsedà--Misiurewicz (AM) systems. We prove that for an open set of parameters, the unique non-atomic stationary measure for an AM-system has Hausdorff dimension strictly smaller than $1$. In particular, we obtain singularity of these…
▽ More
We study stationary measures for iterated function systems (considered as random dynamical systems) consisting of two piecewise affine interval homeomorphisms, called Alsedà--Misiurewicz (AM) systems. We prove that for an open set of parameters, the unique non-atomic stationary measure for an AM-system has Hausdorff dimension strictly smaller than $1$. In particular, we obtain singularity of these measures, answering partially a question of Alsedà and Misiurewicz from 2014.
△ Less
Submitted 10 July, 2023; v1 submitted 12 August, 2022;
originally announced August 2022.
-
On the Shroer-Sauer-Ott-Yorke predictability conjecture for time-delay embeddings
Authors:
Krzysztof Barański,
Yonatan Gutman,
Adam Śpiewak
Abstract:
Shroer, Sauer, Ott and Yorke conjectured in 1998 that the Takens delay embedding theorem can be improved in a probabilistic context. More precisely, their conjecture states that if $μ$ is a natural measure for a smooth diffeomorphism of a Riemannian manifold and $k$ is greater than the information dimension of $μ$, then $k$ time-delayed measurements of a one-dimensional observable $h$ are generica…
▽ More
Shroer, Sauer, Ott and Yorke conjectured in 1998 that the Takens delay embedding theorem can be improved in a probabilistic context. More precisely, their conjecture states that if $μ$ is a natural measure for a smooth diffeomorphism of a Riemannian manifold and $k$ is greater than the information dimension of $μ$, then $k$ time-delayed measurements of a one-dimensional observable $h$ are generically sufficient for a predictable reconstruction of $μ$-almost every initial point of the original system. This reduces by half the number of required measurements, compared to the standard (deterministic) setup. We prove the conjecture for ergodic measures and show that it holds for a generic smooth diffeomorphism, if the information dimension is replaced by the Hausdorff one. To this aim, we prove a general version of predictable embedding theorem for injective Lipschitz maps on compact sets and arbitrary Borel probability measures. We also construct an example of a $C^\infty$-smooth diffeomorphism with a natural measure, for which the conjecture does not hold in its original formulation.
△ Less
Submitted 3 February, 2022; v1 submitted 16 July, 2021;
originally announced July 2021.
-
Typical absolute continuity for classes of dynamically defined measures
Authors:
Balázs Bárány,
Károly Simon,
Boris Solomyak,
Adam Śpiewak
Abstract:
We consider one-parameter families of smooth uniformly contractive iterated function systems $\{f^λ_j\}$ on the real line. Given a family of parameter dependent measures $\{μ_λ\}$ on the symbolic space, we study geometric and dimensional properties of their images under the natural projection maps $Π^λ$. The main novelty of our work is that the measures $μ_λ$ depend on the parameter, whereas up ti…
▽ More
We consider one-parameter families of smooth uniformly contractive iterated function systems $\{f^λ_j\}$ on the real line. Given a family of parameter dependent measures $\{μ_λ\}$ on the symbolic space, we study geometric and dimensional properties of their images under the natural projection maps $Π^λ$. The main novelty of our work is that the measures $μ_λ$ depend on the parameter, whereas up till now it has been usually assumed that the measure on the symbolic space is fixed and the parameter dependence comes only from the natural projection. This is especially the case in the question of absolute continuity of the projected measure $(Π^λ)_*μ_λ$, where we had to develop a new approach in place of earlier attempt which contains an error. Our main result states that if $μ_λ$ are Gibbs measures for a family of Hölder continuous potentials $φ^λ$, with Hölder continuous dependence on $λ$ and $\{Π^λ\}$ satisfy the transversality condition, then the projected measure $(Π^λ)_*μ_λ$ is absolutely continuous for Lebesgue a.e.\ $λ$, such that the ratio of entropy over the Lyapunov exponent is strictly greater than $1$. We deduce it from a more general almost sure lower bound on the Sobolev dimension for families of measures with regular enough dependence on the parameter. Under less restrictive assumptions, we also obtain an almost sure formula for the Hausdorff dimension. As applications of our results, we study stationary measures for iterated function systems with place-dependent probabilities (place-dependent Bernoulli convolutions and the Blackwell measure for binary channel) and equilibrium measures for hyperbolic IFS with overlaps (in particular: natural measures for non-homogeneous self-similar IFS and certain systems corresponding to random continued fractions).
△ Less
Submitted 3 February, 2022; v1 submitted 8 July, 2021;
originally announced July 2021.
-
Around the variational principle for metric mean dimension
Authors:
Yonatan Gutman,
Adam Śpiewak
Abstract:
We study variational principles for metric mean dimension. First we prove that in the variational principle of Lindenstrauss and Tsukamoto it suffices to take supremum over ergodic measures. Second we derive a variational principle for metric mean dimension involving growth rates of measure-theoretic entropy of partitions decreasing in diameter which holds in full generality and in particular does…
▽ More
We study variational principles for metric mean dimension. First we prove that in the variational principle of Lindenstrauss and Tsukamoto it suffices to take supremum over ergodic measures. Second we derive a variational principle for metric mean dimension involving growth rates of measure-theoretic entropy of partitions decreasing in diameter which holds in full generality and in particular does not necessitate the assumption of tame growth of covering numbers. The expressions involved are a dynamical version of Renyi information dimension. Third we derive a new expression for Geiger-Koch information dimension rate for ergodic shift-invariant measures. Finally we develop a lower bound for metric mean dimension in terms of Brin-Katok local entropy.
△ Less
Submitted 22 February, 2021; v1 submitted 27 October, 2020;
originally announced October 2020.
-
New Uniform Bounds for Almost Lossless Analog Compression
Authors:
Yonatan Gutman,
Adam Śpiewak
Abstract:
Wu and Verdú developed a theory of almost lossless analog compression, where one imposes various regularity conditions on the compressor and the decompressor with the input signal being modelled by a (typically infinite-entropy) stationary stochastic process. In this work we consider all stationary stochastic processes with trajectories in a prescribed set $\mathcal{S} \subset [0,1]^\mathbb{Z}$ of…
▽ More
Wu and Verdú developed a theory of almost lossless analog compression, where one imposes various regularity conditions on the compressor and the decompressor with the input signal being modelled by a (typically infinite-entropy) stationary stochastic process. In this work we consider all stationary stochastic processes with trajectories in a prescribed set $\mathcal{S} \subset [0,1]^\mathbb{Z}$ of (bi)infinite sequences and find uniform lower and upper bounds for certain compression rates in terms of metric mean dimension and mean box dimension. An essential tool is the recent Lindenstrauss-Tsukamoto variational principle expressing metric mean dimension in terms of rate-distortion functions.
△ Less
Submitted 18 June, 2019;
originally announced June 2019.
-
Singular stationary measures for random piecewise affine interval homeomorphisms
Authors:
Krzysztof Barański,
Adam Śpiewak
Abstract:
We show that the stationary measure for some random systems of two piecewise affine homeomorphisms of the interval is singular, verifying partially a conjecture by Alsedà and Misiurewicz and contributing to a question of Navas on the absolute continuity of stationary measures, considered in the setup of semigroups of piecewise affine circle homeomorphisms. We focus on the case of resonant boundary…
▽ More
We show that the stationary measure for some random systems of two piecewise affine homeomorphisms of the interval is singular, verifying partially a conjecture by Alsedà and Misiurewicz and contributing to a question of Navas on the absolute continuity of stationary measures, considered in the setup of semigroups of piecewise affine circle homeomorphisms. We focus on the case of resonant boundary derivatives.
△ Less
Submitted 27 May, 2019;
originally announced May 2019.
-
Metric mean dimension and analog compression
Authors:
Yonatan Gutman,
Adam Śpiewak
Abstract:
Wu and Verdú developed a theory of almost lossless analog compression, where one imposes various regularity conditions on the compressor and the decompressor with the input signal being modelled by a (typically infinite-entropy) stationary stochastic process. In this work we consider all stationary stochastic processes with trajectories in a prescribed set of (bi-)infinite sequences and find unifo…
▽ More
Wu and Verdú developed a theory of almost lossless analog compression, where one imposes various regularity conditions on the compressor and the decompressor with the input signal being modelled by a (typically infinite-entropy) stationary stochastic process. In this work we consider all stationary stochastic processes with trajectories in a prescribed set of (bi-)infinite sequences and find uniform lower and upper bounds for certain compression rates in terms of metric mean dimension and mean box dimension. An essential tool is the recent Lindenstrauss-Tsukamoto variational principle expressing metric mean dimension in terms of rate-distortion functions. We obtain also lower bounds on compression rates for a fixed stationary process in terms of the rate-distortion dimension rates and study several examples.
△ Less
Submitted 1 May, 2020; v1 submitted 2 December, 2018;
originally announced December 2018.
-
A probabilistic Takens theorem
Authors:
Krzysztof Barański,
Yonatan Gutman,
Adam Śpiewak
Abstract:
Let $X \subset \mathbb{R}^N$ be a Borel set, $μ$ a Borel probability measure on $X$ and $T:X \to X$ a Lipschitz and injective map. Fix $k \in \mathbb{N}$ greater than the (Hausdorff) dimension of $X$ and assume that the set of $p$-periodic points has dimension smaller than $p$ for $p=1, \ldots, k-1$. We prove that for a typical polynomial perturbation $\tilde{h}$ of a given Lipschitz map…
▽ More
Let $X \subset \mathbb{R}^N$ be a Borel set, $μ$ a Borel probability measure on $X$ and $T:X \to X$ a Lipschitz and injective map. Fix $k \in \mathbb{N}$ greater than the (Hausdorff) dimension of $X$ and assume that the set of $p$-periodic points has dimension smaller than $p$ for $p=1, \ldots, k-1$. We prove that for a typical polynomial perturbation $\tilde{h}$ of a given Lipschitz map $h : X \to \mathbb{R}$, the $k$-delay coordinate map $x \mapsto (\tilde{h}(x), \tilde{h}(Tx), \ldots, \tilde{h}(T^{k-1}x))$ is injective on a set of full measure $μ$. This is a probabilistic version of the Takens delay embedding theorem as proven by Sauer, Yorke and Casdagli. We also provide a non-dynamical probabilistic embedding theorem of similar type, which strengthens a previous result by Alberti, Bölcskei, De Lellis, Koliander and Riegler. In both cases, the key improvements compared to the non-probabilistic counterparts are the reduction of the number of required measurements from $2\dim X$ to $\dim X$ and using Hausdorff dimension instead of the box-counting one. We present examples showing how the use of the Hausdorff dimension improves the previously obtained results.
△ Less
Submitted 11 August, 2020; v1 submitted 14 November, 2018;
originally announced November 2018.
