It is shown that the set of learning systems having a singular stationary distribution is generic in the family of all systems satisfying the average contractivity condition.

Let $\mu$ be a probability measure on $[0,1]$ which is invariant and ergodic for $T_a\left(x\right)=ax\mathrm{𝚖𝚘𝚍}1$, and $0<\mathrm{𝚍𝚒𝚖}\mu <1$. Let $f$ be a local diffeomorphism on some open set. We show that if $E\subseteq ℝ$ and $\left(f\mu \right){\left|{}_E\sim \mu \right|}_E$, then $f^{\text{'}}\left(x\right)\in \left\{\pm a^r:r\in \mathbb{Q}\right\}$ at $\mu$-a.e. point $x\in f^{-1}E$. In particular, if $g$ is a piecewise-analytic map preserving $\mu$ then there is an open $g$-invariant set $U$ containing supp $\mu$ such that $g\left|{}_U\right.$ is piecewise-linear with slopes which are rational powers of $a$. In a similar vein, for $\mu$ as above, if $b$ is another integer and $a,b$ are not powers of a common integer, and if $\nu$ is a $T_b$-invariant...

We show that the theory of graph directed Markov systems can be used to study exceptional minimal sets of some foliated manifolds. A C¹ smooth embedding of a contracting or parabolic Markov system into the holonomy pseudogroup of a codimension one foliation allows us to describe in detail the h-dimensional Hausdorff and packing measures of the intersection of a complete transversal with exceptional minimal sets.

The main goal is to use Gibbs measures in a markovian matrices context and in a more general context, to compute the Hausdorff dimension of subsets of [0, 1[ and [0, 1[². We introduce a parameter t which could be interpreted within thermodynamic framework as the variable conjugate to energy. In some particular cases we recover the Shannon-McMillan-Breiman and Eggleston theorems. Our proofs are deeply rooted in the properties of non-negative irreducible matrices and large deviations techniques as...

We study Markov shifts over countable (finite or countably infinite) alphabets, i.e. shifts generated by incidence matrices. In particular, we derive necessary and sufficient conditions for the existence of a Gibbs state for a certain class of infinite Markov shifts. We further establish a characterization of the existence, uniqueness and ergodicity of invariant Gibbs states for this class of shifts. Our results generalize the well-known results for finitely irreducible Markov shifts.

There has been a lot of interest and activity along the general lines of analysis on metric spaces recently, as in [2], [3], [26], [40], [41], [46], [48], [49], [51], [82], [83], [89], for instance. Of course this is closely related to and involves ideas concerning spaces of homogeneous type, as in [18], [19], [66], [67], [92], as well as sub-Riemannian spaces, e.g., [8], [9], [34], [47], [52], [53], [54], [55], [68], [70], [72], [73], [84], [86], [88]. In the present survey we try to give an introduction...

We obtain a lower bound for the Hausdorff dimension of the graph of a fractal interpolation function with interpolation points $(i/N,y_i):i=0,1,...,N$.

We calculate the almost sure Hausdorff dimension of the random covering set ${lim\; sup}_{n\to \infty }(g_n+{\xi }_n)$ in $d$-dimensional torus ${𝕋}^d$, where the sets $g_n\subset {𝕋}^d$ are parallelepipeds, or more generally, linear images of a set with nonempty interior, and ${\xi }_n\in {𝕋}^d$ are independent and uniformly distributed random points. The dimension formula, derived from the singular values of the linear mappings, holds provided that the sequences of the singular values are decreasing.

The aim of this paper is to calculate (deterministically) the Hausdorff dimension of the scale-sparse Weierstrass-type functions $W_s\left(x\right):={\sum }_{j\ge 1}{\rho }^{-{\gamma }^{j}s}g({\rho }^{{\gamma }^j}x+{\theta }_j)$, where ρ > 1, γ > 1 and 0 < s < 1, and g is a periodic Lipschitz function satisfying some additional appropriate conditions.

For any $x\in [0,1)$, let $x=[{ϵ}_1,{ϵ}_2,\cdots ,]$ be its dyadic expansion. Call $r_n\left(x\right):=max\{j\ge 1:{ϵ}_{i+1}=\cdots ={ϵ}_{i+j}=1$, $0\le i\le n-j\}$ the $n$-th maximal run-length function of $x$. P. Erdös and A. Rényi showed that ${lim}_{n\to \infty }{r}_n\left(x\right)/{log}_{2}n=1$ almost surely. This paper is concentrated on the points violating the above law. The size of sets of points, whose run-length function assumes on other possible asymptotic behaviors than ${log}_{2}n$, is quantified by their Hausdorff dimension.

This is a survey on transformation of fractal type sets and measures under orthogonal projections and more general mappings.

Let X be a complete metric space and write (X) for the family of all Borel probability measures on X. The local dimension $di{m}_{loc}(\mu ;x)$ of a measure μ ∈ (X) at a point x ∈ X is defined by $di{m}_{loc}(\mu ;x)=li{m}_{r\searrow 0}\left(log\mu \left(B(x,r)\right)\right)/\left(logr\right)$ whenever the limit exists, and plays a fundamental role in multifractal analysis. It is known that if a measure μ ∈ (X) satisfies a few general conditions, then the local dimension of μ exists and is equal to a constant for μ-a.a. x ∈ X. In view of this, it is natural to expect that for a fixed x ∈ X, the local dimension...