-
Bounds for the sampling discretization error and their applications to the universal sampling discretization
Authors:
E. D. Kosov,
V. N. Temlyakov
Abstract:
In the first part of the paper we study absolute error of sampling discretization of the integral $L_p$-norm for function classes of continuous functions. We use basic approaches from chaining technique to provide general upper bounds for the error of sampling discretization of the $L_p$-norm on a given function class in terms of entropy numbers in the uniform norm of this class. As an example we…
▽ More
In the first part of the paper we study absolute error of sampling discretization of the integral $L_p$-norm for function classes of continuous functions. We use basic approaches from chaining technique to provide general upper bounds for the error of sampling discretization of the $L_p$-norm on a given function class in terms of entropy numbers in the uniform norm of this class. As an example we apply these general results to obtain new error bounds for sampling discretization of the $L_p$-norms on classes of multivariate functions with mixed smoothness. In the second part of the paper we apply our general bounds to study the problem of universal sampling discretization.
△ Less
Submitted 9 August, 2024; v1 submitted 9 December, 2023;
originally announced December 2023.
-
Sampling discretization of the uniform norm and applications
Authors:
E. D. Kosov,
V. N. Temlyakov
Abstract:
Discretization of the uniform norm of functions from a given finite dimensional subspace of continuous functions is studied. Previous known results show that for any $N$-dimensional subspace of the space of continuous functions it is sufficient to use $e^{CN}$ sample points for an accurate upper bound for the uniform norm by the discrete norm and that one cannot improve on the exponential growth o…
▽ More
Discretization of the uniform norm of functions from a given finite dimensional subspace of continuous functions is studied. Previous known results show that for any $N$-dimensional subspace of the space of continuous functions it is sufficient to use $e^{CN}$ sample points for an accurate upper bound for the uniform norm by the discrete norm and that one cannot improve on the exponential growth of the number of sampling points for a good discretization theorem in the uniform norm. In this paper we focus on two types of results, which allow us to obtain good discretization of the uniform norm with polynomial in $N$ number of points. In the first way we weaken the discretization inequality by allowing a bound of the uniform norm by the discrete norm multiplied by an extra factor, which may depend on $N$. In the second way we impose restrictions on the finite dimensional subspace under consideration. In particular, we prove a general result, which connects the upper bound on the number of sampling points in the discretization theorem for the uniform norm with the best $m$-term bilinear approximation of the Dirichlet kernel associated with the given subspace.
△ Less
Submitted 7 May, 2024; v1 submitted 25 June, 2023;
originally announced June 2023.
-
Regularity of solutions to Kolmogorov equations with perturbed drifts
Authors:
Vladimir I. Bogachev,
Egor D. Kosov,
Alexander V. Shaposhnikov
Abstract:
We prove that a probability solution of the stationary Kolmogorov equation generated by a first order perturbation $v$ of the Ornstein--Uhlenbeck operator $L$ possesses a highly integrable density with respect to the Gaussian measure satisfying the non-perturbed equation provided that $v$ is sufficiently integrable. More generally, a similar estimate is proved for solutions to inequalities connect…
▽ More
We prove that a probability solution of the stationary Kolmogorov equation generated by a first order perturbation $v$ of the Ornstein--Uhlenbeck operator $L$ possesses a highly integrable density with respect to the Gaussian measure satisfying the non-perturbed equation provided that $v$ is sufficiently integrable. More generally, a similar estimate is proved for solutions to inequalities connected with Markov semigroup generators under the curvature condition $CD(θ,\infty)$. For perturbations from $L^p$ an analog of the Log-Sobolev inequality is obtained. It is also proved in the Gaussian case that the gradient of the density is integrable to all powers. We obtain dimension-free bounds on the density and its gradient, which also covers the infinite-dimensional case.
△ Less
Submitted 9 April, 2021;
originally announced April 2021.
-
Regularity of linear and polynomial images of Skorohod differentiable measures
Authors:
Egor D. Kosov
Abstract:
In this paper we study the regularity properties of linear and polynomial images of Skorohod differentiable measures. Firstly, we obtain estimates for the Skorohod derivative norm of a projection of a product of Scorohod differentiable measures. In the second part of the paper we prove Nikolskii--Besov regularity of a polynomial image of a Skorohod differentiable measure on $\mathbb{R}^n$.
In this paper we study the regularity properties of linear and polynomial images of Skorohod differentiable measures. Firstly, we obtain estimates for the Skorohod derivative norm of a projection of a product of Scorohod differentiable measures. In the second part of the paper we prove Nikolskii--Besov regularity of a polynomial image of a Skorohod differentiable measure on $\mathbb{R}^n$.
△ Less
Submitted 1 July, 2019;
originally announced July 2019.
-
Moment measures and stability for Gaussian inequalities
Authors:
Alexander V. Kolesnikov,
Egor D. Kosov
Abstract:
Let $γ$ be the standard Gaussian measure on $\mathbb{R}^n$ and let $\mathcal{P}_γ$ be the space of probability measures that are absolutely continuous with respect to $γ$. We study lower bounds for the functional $\mathcal{F}_γ(μ) = {\rm Ent}(μ) - \frac{1}{2} W^2_2(μ, ν)$, where $μ\in \mathcal{P}_γ, ν\in \mathcal{P}_γ$, ${\rm Ent}(μ) = \int \log\bigl( \fracμγ\bigr) d μ$ is the relative Gaussian en…
▽ More
Let $γ$ be the standard Gaussian measure on $\mathbb{R}^n$ and let $\mathcal{P}_γ$ be the space of probability measures that are absolutely continuous with respect to $γ$. We study lower bounds for the functional $\mathcal{F}_γ(μ) = {\rm Ent}(μ) - \frac{1}{2} W^2_2(μ, ν)$, where $μ\in \mathcal{P}_γ, ν\in \mathcal{P}_γ$, ${\rm Ent}(μ) = \int \log\bigl( \fracμγ\bigr) d μ$ is the relative Gaussian entropy, and $W_2$ is the quadratic Kantorovich distance. The minimizers of $\mathcal{F}_γ$ are solutions to a dimension-free Gaussian analog of the (real) Kähler-Einstein equation. We show that $\mathcal{F}_γ(μ) $ is bounded from below under the assumption that the Gaussian Fisher information of $ν$ is finite and prove a priori estimates for the minimizers. Our approach relies on certain stability estimates for the Gaussian log-Sobolev and Talagrand transportation inequalities.
△ Less
Submitted 30 December, 2017;
originally announced January 2018.
-
Besov classes on finite- and infinite-dimensional spaces and embedding theorems
Authors:
Egor D. Kosov
Abstract:
We give a new description of classical Besov spaces in terms of a new modulus of continuity. Then a similar approach is used to introduce Besov classes on an infinite-dimensional space endowed with a Gaussian measure.
We give a new description of classical Besov spaces in terms of a new modulus of continuity. Then a similar approach is used to introduce Besov classes on an infinite-dimensional space endowed with a Gaussian measure.
△ Less
Submitted 5 November, 2017;
originally announced November 2017.
-
A new approach to Nikolskii-Besov classes
Authors:
Vladimir I. Bogachev,
Egor D. Kosov,
Svetlana N. Popova
Abstract:
We give a new characterization of Nikolskii-Besov classes of functions of fractional smoothness by means of a nonlinear integration by parts formula in the form of a nonlinear inequality. A similar characterization is obtained for Nikolskii-Besov classes with respect to Gaussian measures on finite- and infinite-dimensional spaces.
We give a new characterization of Nikolskii-Besov classes of functions of fractional smoothness by means of a nonlinear integration by parts formula in the form of a nonlinear inequality. A similar characterization is obtained for Nikolskii-Besov classes with respect to Gaussian measures on finite- and infinite-dimensional spaces.
△ Less
Submitted 20 July, 2017;
originally announced July 2017.
-
Fractional smoothness of images of logarithmically concave measures under polynomials
Authors:
Egor D. Kosov
Abstract:
We show that a measure on the real line that is the image of a log-concave measure under a polynomial of degree $d$ possesses a density from the Nikol'skii--Besov class of fractional order $1/d$. This result is used to prove an estimate of the total variation distance between such measures in terms of the Fortet--Mourier distance.
We show that a measure on the real line that is the image of a log-concave measure under a polynomial of degree $d$ possesses a density from the Nikol'skii--Besov class of fractional order $1/d$. This result is used to prove an estimate of the total variation distance between such measures in terms of the Fortet--Mourier distance.
△ Less
Submitted 30 April, 2016;
originally announced May 2016.
-
Deviation of polynomials from their expectations and isoperimetry
Authors:
Lavrentin M. Arutyunyan,
Egor D. Kosov
Abstract:
In the first part we study deviation of a polynomial from its mathematical expectation. This deviation can be estimated from above by Carbery--Wright inequality, so we investigate estimates of the deviation from below. We obtain such estimates in two different cases: for Gaussian measures and a polynomial of an arbitrary degree and for an arbitrary log-concave measure but only for polynomials of t…
▽ More
In the first part we study deviation of a polynomial from its mathematical expectation. This deviation can be estimated from above by Carbery--Wright inequality, so we investigate estimates of the deviation from below. We obtain such estimates in two different cases: for Gaussian measures and a polynomial of an arbitrary degree and for an arbitrary log-concave measure but only for polynomials of the second degree. In the second part we deals with isoperimetric inequality and the Poincaré inequality for probability measures on the real line that are images of the uniform distributions on convex compact sets in $\mathbb{R}^n$ under polynomial mappings.
△ Less
Submitted 16 March, 2016;
originally announced March 2016.
-
Fractional smoothness of distributions of polynomials and a fractional analog of the Hardy--Landau--Littlewood inequality
Authors:
Vladimir I. Bogachev,
Egor D. Kosov,
Georgii I. Zelenov
Abstract:
We prove that the distribution density of any non-constant polynomial $f(ξ_1,ξ_2,\ldots)$ of degree $d$ in independent standard Gaussian random variables $ξ$ (possibly, in infinitely many variables) always belongs to the Nikol'skii--Besov space $B^{1/d}(\mathbb{R}^1)$ of fractional order $1/d$ (and this order is best possible), and an analogous result holds for polynomial mappings with values in…
▽ More
We prove that the distribution density of any non-constant polynomial $f(ξ_1,ξ_2,\ldots)$ of degree $d$ in independent standard Gaussian random variables $ξ$ (possibly, in infinitely many variables) always belongs to the Nikol'skii--Besov space $B^{1/d}(\mathbb{R}^1)$ of fractional order $1/d$ (and this order is best possible), and an analogous result holds for polynomial mappings with values in $\mathbb{R}^k$.
Our second main result is an upper bound on the total variation distance between two probability measures on $\mathbb{R}^k$ via the Kantorovich distance between them and a suitable Nikol'skii--Besov norm of their difference.
As an application we consider the total variation distance between the distributions of two random $k$-dimensional vectors composed of polynomials of degree $d$ in Gaussian random variables and show that this distance is estimated by a fractional power of the Kantorovich distance with an exponent depending only on $d$ and $k$, but not on the number of variables of the considered polynomials.
△ Less
Submitted 30 April, 2016; v1 submitted 16 February, 2016;
originally announced February 2016.
-
Two properties of vectors of quadratic forms in Gaussian random variables
Authors:
Vladimir I. Bogachev,
Egor D. Kosov,
Ivan Nourdin,
Guillaume Poly
Abstract:
We study distributions of random vectors whose components are second order polynomials in Gaussian random variables. Assuming that the law of such a vector is not absolutely continuous with respect to Lebesgue measure, we derive some interesting consequences. Our second result gives a characterization of limits in law for sequences of such vectors.
We study distributions of random vectors whose components are second order polynomials in Gaussian random variables. Assuming that the law of such a vector is not absolutely continuous with respect to Lebesgue measure, we derive some interesting consequences. Our second result gives a characterization of limits in law for sequences of such vectors.
△ Less
Submitted 26 May, 2013;
originally announced May 2013.
