arXiv subjects

Andreas Klotz

Publications and source records attributed to Andreas Klotz.

13 recordsLinked to original sources

arXiv2025

Quantitative estimates: How well does the discrete Fourier transform approximate the Fourier transform on $\mathbb{R}$

In order to compute the Fourier transform of a function $f$ on the real line numerically, one samples $f$ on a grid and then takes the discrete Fourier transform. We derive exact error estimates for this procedure in terms of the decay and smoothness of $f$. The analysis provides a new recipe of how to relate the number of samples, the sampling interval, and the grid size.

math.NA↗

arXiv2025

Discretization of Continuous Frames by Quasi-Monte Carlo Methods

We introduce a discretization scheme for continuous localized frames using quasi-Monte Carlo integration and discrepancy theory. By generalizing classical concepts, we define a discrepancy measure on the entire phase space $\mathbb{R}^2$ and establish a corresponding Koksma-Hlawka inequality. This approach enables control over the density of the discretized frame and ensures the universality of the sampling set, relying only on the discrepancy of the sampling set and on the Sobolev-type seminorm of an iterated kernel rather than on specific frame properties.

math.FA↗

arXiv2023

Spectral Subspaces of Sturm-Liouville Operators and Variable Bandwidth

We study spectral subspaces of the Sturm-Liouville operator $f \mapsto -(pf')'$ on $\mathbb{R}$, where $p$ is a positive, piecewise constant function. Functions in these subspaces can be thought of as having a local bandwidth determined by $1/\sqrt{p}$. Using the spectral theory of Sturm-Liouville operators, we make the reproducing kernel of these spectral subspaces more explicit and compute it completely in certain cases. As a contribution to sampling theory, we then prove necessary density conditions for sampling and interpolation in these subspaces and determine the critical density that separates sets of stable sampling from sets of interpolation.

math.CA↗

arXiv2021

Necessary Density Conditions for Sampling and Interpolation in Spectral Subspaces of Elliptic Differential Operators

We prove necessary density conditions for sampling in spectral subspaces of a second order uniformly elliptic differential operator on $R^d$ with slowly oscillating symbol. For constant coefficient operators, these are precisely Landaus necessary density conditions for bandlimited functions, but for more general elliptic differential operators it has been unknown whether such a critical density even exists. Our results prove the existence of a suitable critical sampling density and compute it in terms of the geometry defined by the elliptic operator. In dimension 1, functions in a spectral subspace can be interpreted as functions with variable bandwidth, and we obtain a new critical density for variable bandwidth. The methods are a combination of the spectral theory and the regularity theory of elliptic partial differential operators, some elements of limit operators, certain compactifications of $R^d $, and the theory of reproducing kernel Hilbert spaces.

math.FA↗

arXiv2020

Phase Transitions in Rate Distortion Theory and Deep Learning

Rate distortion theory is concerned with optimally encoding a given signal class $\mathcal{S}$ using a budget of $R$ bits, as $R\to\infty$. We say that $\mathcal{S}$ can be compressed at rate $s$ if we can achieve an error of $\mathcal{O}(R^{-s})$ for encoding $\mathcal{S}$; the supremal compression rate is denoted $s^\ast(\mathcal{S})$. Given a fixed coding scheme, there usually are elements of $\mathcal{S}$ that are compressed at a higher rate than $s^\ast(\mathcal{S})$ by the given coding scheme; we study the size of this set of signals. We show that for certain "nice" signal classes $\mathcal{S}$, a phase transition occurs: We construct a probability measure $\mathbb{P}$ on $\mathcal{S}$ such that for every coding scheme $\mathcal{C}$ and any $s >s^\ast(\mathcal{S})$, the set of signals encoded with error $\mathcal{O}(R^{-s})$ by $\mathcal{C}$ forms a $\mathbb{P}$-null-set. In particular our results apply to balls in Besov and Sobolev spaces that embed compactly into $L^2(Ω)$ for a bounded Lipschitz domain $Ω$. As an application, we show that several existing sharpness results concerning function approximation using deep neural networks are generically sharp. We also provide quantitative and non-asymptotic bounds on the probability that a random $f\in\mathcal{S}$ can be encoded to within accuracy $\varepsilon$ using $R$ bits. This result is applied to the problem of approximately representing $f\in\mathcal{S}$ to within accuracy $\varepsilon$ by a (quantized) neural network that is constrained to have at most $W$ nonzero weights and is generated by an arbitrary "learning" procedure. We show that for any $s >s^\ast(\mathcal{S})$ there are constants $c,C$ such that, no matter how we choose the "learning" procedure, the probability of success is bounded from above by $\min\big\{1,2^{C\cdot W\lceil\log_2(1+W)\rceil^2 -c\cdot\varepsilon^{-1/s}}\big\}$.

math.FA↗

arXiv2017

Density of Sampling and Interpolation in Reproducing Kernel Hilbert Spaces

We derive necessary density conditions for sampling and for interpolation in general reproducing kernel Hilbert spaces satisfying some natural conditions on the geometry of the space and the reproducing kernel. If the volume of shells is small compared to the volume of balls (weak annular decay property) and if the kernel possesses some off-diagonal decay or even some weaker form of localization, then there exists a critical density $D$ with the following property: a set of sampling has density $\geq D$, whereas a set of interpolation has density $\leq D$. The main theorem unifies many known density theorems in signal processing, complex analysis, and harmonic analysis. For the special case of bandlimited function we recover Landau's fundamental density result. In complex analysis we rederive a critical density for generalized Fock spaces. In harmonic analysis we obtain the first general result about the density of coherent frames.

math.FA↗

arXiv2015

What is Variable Bandwidth?

We propose a new notion of variable bandwidth that is based on the spectral subspaces of an elliptic operator $A_pf = - (pf')'$ where $p>0$ is a strictly positive function. Denote by $c_Λ (A_p)$ the orthogonal projection of $A_p$ corresponding to the spectrum of $A_p$ in $Λ$, the range of this projection is the space of functions of variable bandwidth with spectral set in $Λ$. We will develop the basic theory of these function spaces. First, we derive (nonuniform) sampling theorems, second, we prove necessary density conditions in the style of Landau. Roughly, for a spectrum $Λ= [0,Ω] $ the main results say that, in a neighborhood of $x\in R$, a function of variable bandwidth behaves like a bandlimited function with local bandwidth $(Ω/ p(x))^{1/2}$. Although the formulation of the results is deceptively similar to the corresponding results for classical bandlimited functions, the methods of proof are much more involved. On the one hand, we use the oscillation method from sampling theory and frame theoretic methods, on the other hand, we need the precise spectral theory of Sturm-Liouville operators and the scattering theory of one-dimensional Schrödinger operators.

math.FA↗

arXiv2013

The Optimal Dyadic Derivative

We show that the best approximation to the difference operators on the cyclic groups of order $2^n$ by a dyadic convolution operator are the restrictions of a generalized dyadic derivative. This answers a question on the "intuitive" interpretation of the dyadic derivative posed by Butzer and Wagner more than 30 years ago.

math.CA↗

arXiv2012

Norm-Controlled Inversion in Smooth Banach Algebras, II

We show that smoothness implies norm-controlled inversion: the smoothness of an element $a$ in a Banach algebra with a one-parameter automorphism group is preserved under inversion, and the norm of the inverse $a^{-1}$ is controlled by the smoothness of $a$ and by spectral data. In our context smooth subalgebras are obtained with the classical constructions of approximation theory and resemble spaces of differentiable functions, Besov spaces or Bessel potential spaces. To treat ultra-smoothness, we resort to Dales-Davie algebras. Furthermore, based on Baskakov's work, we derive explicit norm control estimates for infinite matrices with polynomial off-diagonal decay. This is a quantitative version of Jaffard's theorem.

math.FA↗

arXiv2012

Norm-Controlled Inversion in Smooth Banach Algebras, I

Every differential subalgebra of a unital $C^*$-algebra is spectrally invariant. We derive a quantitative version of this well-known fact and show that a minimal amount of smoothness, as given by a differential norm, already implies norm control. We obtain an explicit estimate for the differential norm of an invertible element $a$. This estimate depends only on the condition number of $a$ and the ratio of two norms.

math.OA↗

arXiv2012

Inverse Closed Ultradifferential Subalgebras

In previous work we have shown that classical approximation theory provides methods for the systematic construction of inverse-closed smooth subalgebras. Now we extend this work to treat inverse-closed subalgebras of ultradifferentiable elements. In particular, Carleman classes and Dales-Davie algebras are treated. As an application the result of Demko, Smith and Moss and Jaffard on the inverse of a matrix with exponential decay is obtained within the framework of a general theory of smoothness.

math.FA↗

arXiv2010

Spectral Invariance of Besov-Bessel Subalgebras

Using principles of the theory of smoothness spaces we give systematic constructions of scales of inverse-closed subalgebras of a given Banach algebra with the action of a d-parameter automorphism group. In particular we obtain the inverse-closedness of Besov algebras, Bessel potential algebras and approximation algebras of polynomial order in their defining algebra. By a proper choice of the group action these general results can be applied to algebras of infinite matrices and yield inverse-closed subalgebras of matrices with off-diagonal decay of polynomial order. Besides alternative proofs of known results we obtain new classes of inverse-closed subalgebras of matrices with off-diagonal decay .

math.OA↗

arXiv2009

Noncommutative Approximation: Inverse-Closed Subalgebras and Off-Diagonal Decay of Matrices

We investigate two systematic constructions of inverse-closed subalgebras of a given Banach algebra or operator algebra A, both of which are inspired by classical principles of approximation theory. The first construction requires a closed derivation or a commutative automorphism group on A and yields a family of smooth inverse-closed subalgebras of A that resemble the usual Holder-Zygmund spaces. The second construction starts with a graded sequence of subspaces of A and yields a class of inverse-closed subalgebras that resemble the classical approximation spaces. We prove a theorem of Jackson-Bernstein type to show that in certain cases both constructions are equivalent. These results about abstract Banach algebras are applied to algebras of infinite matrices with off-diagonal decay. In particular, we obtain new and unexpected conditions of off-diagonal decay that are preserved under matrix inversion.

math.OA↗