SearcharxivSearch

arXiv subjects

Tim Austin

Publications and source records attributed to Tim Austin.

At least 37 records · Page 2Linked to original sources

A CAT(0)-valued pointwise ergodic theorem

In this note we prove the a pointwise ergodic theorem for functions taking values in a separable complete CAT(0)-space, analogous to Lindenstrauss' pointwise ergodic theorem for real-valued integrable functions on a probability space subject to a probability-preserving action of an amenable l.c.s.c. group, where in the CAT(0) setting the role of ergodic averages is played by the barycentres of the empirical distributions of a CAT(0)-valued map along an orbit of the group action. The proof rests on an approximation argument and an appeal to that result for real-valued maps.

math.GT

Entropy of probability kernels from the backwards tail boundary

A number of recent works have sought to generalize the Kolmogorov-Sinai entropy of probability-preserving transformations to the setting of Markov operators acting on the integrable functions on a probability space $(X,μ)$. These have culminated in a proof by Downarovicz and Frej that these definitions all coincide, and that the resulting quantity is uniquely characterized by certain properties. On the other hand, Makarov has shown that this `operator entropy' is always dominated by the Kolmogorov-Sinai entropy of a classical system that may be constructed from a Markov operator, and that these numbers coincide under certain extra assumptions. This note proves that equality in all cases.

math.DS

Quantitative equidistribution for certain quadruples in quasi-random groups

In a recent paper (arXiv:1211.6372), Bergelson and Tao proved that if $G$ is a $D$-quasi-random group, and $x$,$g$ are drawn uniformly and independently from $G$, then the quadruple $(g,x,gx,xg)$ is roughly equidistributed in the subset of $G^4$ defined by the constraint that the last two coordinates lie in the same conjugacy class. Their proof gives only a qualitative version of this result. The present notes gives a rather more elementary proof which improves this to an explicit polynomial bound in $D^{-1}$.

math.CO

Exchangeable random measures

Let A be a standard Borel space, and consider the space A^{\bbN^{(k)}} of A-valued arrays indexed by all size-k subsets of \bbN. This paper concerns random measures on such a space whose laws are invariant under the natural action of permutations of \bbN. The main result is a representation theorem for such `exchangeable' random measures, obtained using the classical representation theorems for exchangeable arrays due to de Finetti, Hoover, Aldous and Kallenberg. After proving this representation, two applications of exchangeable random measures are given. The first is a short new proof of the Dovbysh-Sudakov Representation Theorem for exchangeable PSD matrices. The second is in the formulation of a natural class of limit objects for dilute mean-field spin glass models, retaining more information than just the limiting Gram-de Finetti matrix used in the study of the Sherrington-Kirkpatrick model.

math.PR

Ajtai-Szemerédi Theorems over quasirandom groups

Two versions of the Ajtai-Szemerédi Theorem are considered in the Cartesian square of a finite non-Abelian group $G$. In case $G$ is sufficiently quasirandom, we obtain strong forms of both versions: if $E \subseteq G\times G$ is fairly dense, then $E$ contains a large number of the desired patterns for most individual choices of `common difference'. For one of the versions, we also show that this set of good common differences is syndetic.

math.CO

Integrable measure equivalence for groups of polynomial growth

Bader, Furman and Sauer have introduced the notion of integrable measure equivalence for finitely-generated groups. This is the sub-equivalence relation of measure equivalence obtained by insisting that the relevant cocycles satisfy an integrability condition. They have used it to prove new classification results for hyperbolic groups. The present work shows that groups of polynomial growth are also quite rigid under integrable measure equivalence, in that if two such groups are equivalent then they must have bi-Lipschitz asymptotic cones. This will follow by proving that the cocycles arising from an integrable measure equivalence converge, albeit in a very weak sense, to bi-Lipschitz maps of asymptotic cones.

math.GR

Partial difference equations over compact Abelian groups, I: modules of solutions

Consider a compact Abelian group $Z$ and closed subgroups $U_1$, \ldots, $U_k \leq Z$. Let $\mathbb{T} := \mathbb{R}/\mathbb{Z}$. This paper examines two kinds of functional equation for measurable functions $Z\to \mathbb{T}$. First, given $f:Z\to \mathbb{T}$ and $w \in Z$, the resulting differenced function is \[d_wf(z) := f(z-w) - f(z).\] In this notation, we study solutions to the system of difference equations \[d_{u_1}\cdots d_{u_k}f \equiv 0 \quad \forall u_1 \in U_1,\ u_2 \in U_2,\ \ldots\ u_k \in U_k.\] Second, we study tuples of measurable functions $f_i:Z\to \mathbb{T}$ such that $f_i$ is invariant under translation by $U_i$ and also \[f_1 + \cdots + f_k = 0.\] For these equations, the solutions form a subgroup of $\mathcal{F}(Z)$ or $\mathcal{F}(Z)^k$, where $\mathcal{F}(Z)$ is the group of measurable functions $Z\to \mathbb{T}$ modulo Haar-a.e. equality. The subgroup of solutions is closed under convergence in probability and is globally invariant under rotations of $Z$, so it is a complete metrizable $Z$-module. We will give a recursive description of the structure of this $Z$-module relative to the solution-modules of lower-order equations of the same kind. These results are obtained as applications of an abstract theory of a special class of $Z$-modules. Most of our work will go into showing that this class of modules is closed under various natural operations. Knowing that, the above descriptions follow as easy consequences. Partial difference equations of the above kind can be seen as an extremal version of the inverse problem for the higher-dimensional, directional analogs of Gowers' uniformity norms. Our methods also give some information about the `stability' version of this inverse problem, which concerns functions whose Gowers norm is sufficiently close to being maximal.

math.FA

Partial difference equations over compact Abelian groups, II: step-polynomial solutions

This paper continues an earlier work on the structure of solutions to two classes of functional equation. Let $Z$ be a compact Abelian group and $U_1$, \ldots, $U_k \leq Z$ be closed subgroups. Given $f:Z\to\mathbb{T}$ and $w \in Z$, one defines the differenced function \[d_wf(z) := f(z+w) - f(z).\] In this notation, we shall study solutions to the system of difference equations \[d_{u_1}\cdots d_{u_k}f \equiv 0 \quad \forall (u_1,\ldots,u_k) \in \prod_{i\leq k}U_i,\] and to the zero-sum problem \[f_1 + \cdots + f_k = 0\] for functions $f_i:Z\to \mathbb{T}$ that are $U_i$-invariant for each $i$. Part I of this work showed that the $Z$-modules of solutions to these problems can be described using a general theory of `almost modest $¶$-modules'. Much of the global structure of these solution $Z$-modules could then be extracted from results about the closure of this general class under certain natural operations, such as forming cohomology groups. The main result of the present paper is that solutions to either problem can always be decomposed into summands which either solve a simpler system of equations, or have some special `step polynomial' structure. This will be proved by augmenting the definition of `almost modest $\mathcal{P}$-modules' further, to isolate a subclass in which elements can be represented by the desired `step polynomials'. We will then find that this subclass is closed under the same operations.

math.FA

Pleasant extensions retaining algebraic structure, II

In this paper we combine the general tools developed in (arXiv:0905.0518) with several ideas taken from earlier work on one-dimensional nonconventional ergodic averages by Furstenberg and Weiss, Host and Kra and Ziegler to study the averages $\frac{1}{N}\sum_{n=1}^N(f_1\circ T^{n\bf{p}_1})(f_2\circ T^{n\bf{p}_2})(f_3\circ T^{n\bf{p}_3})$ for $f_1,f_2,f_3 \in L^\infty(μ)$ associated to a triple of directions $\bf{p}_1,\bf{p}_2,\bf{p}_3 \in \mathbb{Z}^2$ that lie in general position along with $0 \in \mathbb{Z}^2$. We will show how to construct a `pleasant' extension of an initially-given $\mathbb{Z}^2$-system for which these averages admit characteristic factors with a very concrete description, involving one-dimensional isotropy factors and two-step pro-nilsystems. We also use this analysis to construct pleasant extensions and then prove norm convergence for the polynomial nonconventional ergodic averages $\frac{1}{N}\sum_{n=1}^N(f_1\circ T_1^{n^2})(f_2\circ T_1^{n^2}T_2^n)$ associated to two commuting transformations $T_1$, $T_2$.

math.DS

Pleasant extensions retaining algebraic structure, I

In two recent papers we introduced some new techniques for constructing an extension of a probability-preserving system $T:\mathbb{Z}^d\curvearrowright (X,μ)$ that enjoys certain desirable properties in connexion with the asymptotic behaviour of some related nonconventional ergodic averages. The present paper is the first of two that will explore various refinements and extensions of these ideas. This first part is dedicated to some much more general machinery for the construction of extensions that can be used to recover various earlier results. It also contains two relatively simple new applications of this machinery to the study of certain families of nonconventional averages, one in discrete and one in continuous time (convergence being a new result for the latter). In the forthcoming second part (arXiv:0910.0907) we will introduce the problem of describing the characteristic factors and the limit of the linear nonconventional averages $\frac{1}{N}\sum_{n=1}^N \prod_{i=1}^kf_i\circ T^{n\bf{p}_i}$ when the directions $\bf{p}_1$, $\bf{p}_2$, \ldots, $\bf{p}_k \in \mathbb{Z}^d$ are not assumed to be linearly independent, and provide a fairly detailed solution in the case when k = 3, d = 2 and any pair of directions is linearly independent. This will then be used to prove the convergence in $L^2(μ)$ of the quadratic nonconventional averages $\frac{1}{N}\sum_{n=1}^N (f_1\circ T_1^{n^2})(f_2\circ T_1^{n^2}T_2^n)$.

math.DS

On the failure of concentration for the \ell_\infty-ball

Let $(X,d)$ be a compact metric space and $μ$ a Borel probability on $X$. For each $N\geq 1$ let $d^N_\infty$ be the $\ell_\infty$-product on $X^N$ of copies of $d$, and consider $1$-Lipschitz functions $X^N\to\mathbb{R}$ for $d^N_\infty$. If the support of $μ$ is connected and locally connected, then all such functions are close in probability to juntas: that is, functions that depend on only a few coordinates of $X^N$. This describes the failure of measure concentration for these product spaces, and can be seen as a Lipschitz-function counterpart of the celebrated result of Friedgut that Boolean functions with small influences are close to juntas.

math.MG

Non-conventional ergodic averages for several commuting actions of an amenable group

Let $(X,μ)$ be a probability space, $G$ a countable amenable group and $(F_n)_n$ a left Følner sequence in $G$. This paper analyzes the non-conventional ergodic averages \[\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^d (f_i\circ T_1^g\cdots T_i^g)\] associated to a commuting tuple of $μ$-preserving actions $T_1$, ..., $T_d:G\curvearrowright X$ and $f_1$, ..., $f_d \in L^\infty(μ)$. We prove that these averages always converge in $\|\cdot\|_2$, and that they witness a multiple recurrence phenomenon when $f_1 = \ldots = f_d = 1_A$ for a non-negligible set $A\subseteq X$. This proves a conjecture of Bergelson, McCutcheon and Zhang. The proof relies on an adaptation from earlier works of the machinery of sated extensions.

math.DS

Scenery entropy as an invariant of RWRS processes

Probabilistic models of random walks in random sceneries give rise to examples of probability-preserving dynamical systems. A point in the state spaces consists of a walk-trajectory and a scenery, and its `motion' corresponds to shifting the time-origin. These models were proposed as natural examples of non-Bernoulli K-automorphisms by Adler, Ornstein and Weiss. This was proved in a famous analysis by Kalikow using Ornstein's Very Weak Bernoulli characterization of Bernoulli processes. Since then, various authors have generalized this construction to give other examples, including some smooth examples due to Katok and Rudolph. However, the methods used to prove non-Bernoullicity do not obviously show that these examples are distinct from one another. This paper introduces a new isomorphism-invariant of probability-preserving systems, and shows that in a large class of the above examples it essentially captures the Kolmogorov-Sinai entropy of the scenery process alone. As a result, constructions that use different scenery-entropies give continuum-many non-isomorphic examples. Conditionally on an invariance principle for certain local times, these include a continuum of distinct smooth non-Bernoulli K-automorphisms on a fixed compact manifold.

math.DS

A proof of Walsh's convergence theorem using couplings

Walsh has recently proved the norm convergence of all nonconventional ergodic averages involving polynomial sequences in discrete nilpotent acting groups. He deduces this convergence from an equivalent, `finitary' assertion of stability over arbitrarily long time-intervals for these averages, which is proved by essentially finitary means. The present paper shows how the induction at the heart of Walsh's proof can also be implemented using more classical notions of ergodic theory: in particular, couplings and characteristic factors.

math.DS

Ergodic-theoretic implementations of the Roth density-increment argument

We exhibit proofs of two ergodic-theoretic results in the study of multiple recurrence using an analog of the density-increment argument of Roth and Gowers: Furstenberg's Multiple Recurrence Theorem (which implies Szemerédi's Theorem), and a two-dimensional special case of Furstenberg and Katznelson's multidimensional version of this theorem. The second of these requires also an analog of some recent finitary work by Shkredov. Many proofs of these multiple recurrence theorems are now known, but our main goal is to shed some further light on the heuristic correspondence principle that has grown up between the ergodic-theoretic and combinatorial aspects of multiple recurrence and Szemerédi's Theorem. Focusing on the density-increment strategy highlights several close points of connection between these settings.

math.DS

A hierarchical version of the de Finetti and Aldous-Hoover representations

We consider random arrays indexed by the leaves of an infinitary rooted tree of finite depth, with the distribution invariant under the rearrangements that preserve the tree structure. We call such arrays hierarchically exchangeable and prove that they satisfy an analogue of de Finetti's theorem. We also prove a more general result for arrays indexed by several trees, which includes a hierarchical version of the Aldous-Hoover representation.

math.PR

Continuity properties of measurable group cohomology

A version of group cohomology for locally compact groups and Polish modules has previously been developed using a bar resolution restricted to measurable cochains. That theory was shown to enjoy analogs of most of the standard algebraic properties of group cohomology, but various analytic features of those cohomology groups were only partially understood. This paper re-examines some of those issues. At its heart is a simple dimension-shifting argument which enables one to `regularize' measurable cocycles, leading to some simplifications in the description of the cohomology groups. A range of consequences are then derived from this argument. First, we prove that for target modules that are Fréchet spaces, the cohomology groups agree with those defined using continuous cocycles, and hence they vanish in positive degrees when the acting group is compact. Using this, we then show that for Fréchet, discrete or toral modules the cohomology groups are continuous under forming inverse limits of compact base groups, and also under forming direct limits of discrete target modules. Lastly, these results together enable us to establish various circumstances under which the measurable-cochains cohomology groups coincide with others defined using sheaves on a semi-simplicial space associated to the underlying group, or sheaves on a classifying space for that group. We also prove in some cases that the natural quotient topologies on the measurable-cochains cohomology groups are Hausdorff.

math.GR