SearcharxivSearch

arXiv subjects

Morten Nielsen

Publications and source records attributed to Morten Nielsen.

At least 19 recordsLinked to original sources

Divergence of decreasing rearranged Fourier series on every infinite compact abelian group

On every infinite compact Hausdorff abelian group there is a complex-valued continuous function whose Fourier sums, taken over coefficients above a decreasing magnitude threshold, are unbounded almost everywhere. In fact, such functions form a dense \(G_\delta\) in the space of continuous functions with its uniform norm. Moreover, a divergent example may be chosen with Fourier support contained in any prescribed infinite subgroup of the dual, while avoiding any prescribed finite set of frequencies. No metrizability or zero-dimensionality is assumed. The proof combines K\"orner's finite circle and Walsh lemmas with a finite construction on products of odd cyclic groups. A trichotomy for the dual group, exact quotient transfer, and cyclic sampling give finite examples on every group under consideration.

math.FA

Holomorphic Toroidal Pseudodifferential Operators on the Polydisk

We give a necessary and sufficient triangular condition characterizing the toroidal pseudodifferential operators on $\mathbb T^d$ that preserve the positive-frequency cone $\mathbb N_0^d$ and hence act on holomorphic boundary values on the polydisk. For symbols of type $(1,0)$, we prove boundedness on holomorphic Besov and Triebel--Lizorkin spaces throughout the quasi-Banach range. For $S^m_{\rho,\delta}$, $0\leq\delta<\rho\leq1$, we obtain the critical loss $d(1-\rho)|1/p-1/2|$, together with endpoint estimates. A positive-cone oscillatory multiplier proves sharpness of the loss and necessity of the Besov endpoint condition $q\leq t$. As an application, we prove well-posedness for a first-order holomorphic differential operator on these scales.

math.AP

On a Conjecture about Schauder-Basis Properties of the Daubechies Wavelet Packets

Nielsen and Zhou (\emph{Mean size of wavelet packets}, ACHA 13 (2002), 22--34) conjectured that for every Daubechies filter of length at least four the associated wavelet packets fail to be a Schauder basis of $L^p(\mathbb{R})$ for every $p\neq2$; their own $\ell^1/\ell^\infty$ estimate only reaches extreme exponents. We prove the Schauder-basis-failure half of the conjecture in full, for every $1<p<\infty$ with $p\neq2$, in the length-four case. The proof combines convexity of the wavelet packet pressure function with the uniqueness of equilibrium states for irreducible matrix families due to Feng and K\"aenm\"aki (2011) and an exact algebraic separation of two periodic spectral growth rates of the high-pass transition matrices.

math.FA

Matrix-weighted Anisotropic Smoothness Spaces

Given a quasi-norm on $\mathbb{R}^d$ induced by a one-parameter dilation group, we consider matrix weights $W$ in an adapted Muckenhoupt class $\mathbf{A}_p$, $0 < p < \infty$, and use these weights to introduce and study anisotropic matrix-weighted smoothness spaces in both continuous and discrete settings. The spaces are constructed by means of a decomposition method in the frequency domain. We prove the equivalence of the continuous and discrete spaces using suitably adapted tight frames. Compatible notions of molecules and almost diagonal matrices are also introduced, and applications to the study of Fourier multipliers and pseudo-differential operators on vector-valued smoothness spaces are given.

math.FA

Matrix $A_p$-weights relative to a pseudo-metric

Matrix weights satisfying a Muckenhoupt $A_p$-condition relative to a family of anisotropic balls in $\mathbb{R}^d$ defined by a pseudo-metric are studied. It is shown that such matrix weights satisfy a doubling condition and a reverse H\"older inequality. In the special case, where the pseudo-metric is homogeneous with respect to a one-parameter dilation group, the corresponding Muckenhoupt class is shows to satisfy an invariance property under composition with affine transformations generated by the dilation group. A general sampling theorem is derived for the matrix-weighted space $L^p(W)$ for Muckenhoupt $A_p$ weights $W$ along with a corresponding multiplier result for $L^p(W)$. An application of the results to the study of anisotropic matrix-weighed Besov spaces is considered.

math.FA

Painless Construction of Unconditional Bases for Anisotropic Modulation and Triebel-Lizorkin Type Spaces

We construct smooth localised orthonormal bases compatible with anisotropic Triebel-Lizorkin and Besov type spaces on $\mathbb{R}^d$. The construction is based on tensor products of so-called univariate brushlet functions that are based on local trigonometric bases in the frequency domain, and the construction is painless in the sense that all parameters for the construction are explicitly specified. It is shown that the associated decomposition system form unconditional bases for the full family of Triebel-Lizorkin and Besov type spaces, including for the so-called $α$-modulation and $α$-Triebel-Lizorkin spaces. In the second part of the paper we study nonlinear $m$-term approximation with the constructed bases, where direct Jackson and Bernstein inequalities for $m$-term approximation with the tensor brushlet system in $α$-modulation and $α$-Triebel-Lizorkin spaces are derived. The inverse Bernstein estimates rely heavily on the fact that the constructed system is non-redundant.

math.FA

On bandlimited multipliers on matrix-weighted $L^p$-spaces

We extend a classical result by Triebel on boundedness of bandlimited multipliers on $L^p(\mathbb{R}^n)$, $0<p\leq 1$, to a vector-valued and matrix-weighted setting with boundedness of the bandlimited multipliers obtained on $L^p(W)$, $0<p\leq 1$, for matrix-weights $W:\mathbb{R}^n\rightarrow \mathbb{C}^{N\times N}$ that satisfies a matrix Muckenhoupt $A_p$-condition.

math.FA

AntiFold: Improved antibody structure-based design using inverse folding

The design and optimization of antibodies requires an intricate balance across multiple properties. Protein inverse folding models, capable of generating diverse sequences folding into the same structure, are promising tools for maintaining structural integrity during antibody design. Here, we present AntiFold, an antibody-specific inverse folding model, fine-tuned from ESM-IF1 on solved and predicted antibody structures. AntiFold outperforms existing inverse folding tools on sequence recovery across complementarity-determining regions, with designed sequences showing high structural similarity to their solved counterpart. It additionally achieves stronger correlations when predicting antibody-antigen binding affinity in a zero-shot manner, while performance is augmented further when including antigen information. AntiFold assigns low probabilities to mutations that disrupt antigen binding, synergizing with protein language model residue probabilities, and demonstrates promise for guiding antibody optimization while retaining structure-related properties. AntiFold is freely available under the BSD 3-Clause as a web server at https://opig.stats.ox.ac.uk/webapps/antifold/ and and pip installable package at https://github.com/oxpig/AntiFold

q-bio.BM

Matrix weighted modulation spaces

Given a matrix-weight $W$ in the Muckenhoupt class $\mathbf{A}_p(\mathbb{R}^n)$, $1\leq p<\infty$, we introduce corresponding vector-valued continuous and discrete $α$-modulation spaces $M^{s,α}_{p,q}(W)$ and $m^{s,α}_{p,q}(W)$ and prove their equivalence through the use of adapted tight frames. Compatible notions of molecules and almost diagonal matrices are also introduced, and an application to the study of pseudo-differential operators on vector valued spaces is given.

math.FA

Discrete Decomposition of Mixed-Norm $α$-modulation spaces

We study a class of almost diagonal matrices compatible with the mixed-norm $α$-modulation spaces $M_{\vec{p},q}^{s,α}(\mathbb{R}^n)$, $α\in [0,1]$, introduced recently by Cleanthous and Georgiadis [Trans.\ Amer.\ Math.\ Soc.\ 373 (2020), no. 5, 3323-3356]. The class of almost diagonal matrices is shown to be closed under matrix multiplication and we connect the theory to the continuous case by identifying a suitable notion of molecules for the mixed-norm $α$-modulation spaces. We show that the "change of frame" matrix for a pair of time-frequency frames for the mixed-norm $α$-modulation consisting of suitable molecules is almost diagonal. As examples of applications, we use the almost diagonal matrices to construct compactly supported frames for the mixed-norm $α$-modulation spaces, and to obtain a straightforward boundedness result for Fourier multipliers on the mixed-norm $α$-modulation spaces.

math.FA

Stable decomposition of homogeneous Mixed-norm Triebel-Lizorkin spaces

We construct smooth localized orthonormal bases compatible with homogeneous mixed-norm Triebel-Lizorkin spaces in an anisotropic setting on $\bR^d$. The construction is based on tensor products of so-called univariate brushlet functions that are constructed using local trigonometric bases in the frequency domain. It is shown that the associated decomposition system form unconditional bases for the homogeneous mixed-norm Triebel-Lizorkin spaces. In the second part of the paper we study nonlinear $m$-term nonlinear approximation with the constructed basis in the mixed-norm setting, where the behaviour, in general, for $d\geq 2$, is shown to be fundamentally different from the unmixed case. However, Jackson and Bernstein inequalities for $m$-term approximation can still be derived.

math.FA

Pseudodifferential operators on Mixed-Norm $α$-modulation spaces

Mixed-norm $α$-modulation spaces were introduced recently by Cleanthous and Georgiadis [Trans.\ Amer.\ Math.\ Soc.\ 373 (2020), no. 5, 3323-3356]. The mixed-norm spaces $M^{s,α}_{\vec{p},q}(\mathbb{R}^n)$, $α\in [0,1]$, form a family of smoothness spaces that contain the mixed-norm Besov spaces as special cases. In this paper we prove that a pseudodifferential operator $σ(x,D)$ with symbol in the Hörmander class $S^b_ρ$ extends to a bounded operator $σ(x,D)\colon M^{s,α}_{\vec{p},q}(\mathbb{R}^n) \rightarrow M^{s-b,α}_{\vec{p},q}(\mathbb{R}^n)$ provided $0<α\leq ρ\leq 1$, $\vec{p}\in (0,\infty)^n$, and $0<q<\infty$. The result extends the known result that pseudodifferential operators with symbol in the class $S^b_{1}$ maps the mixed-norm Besov space $B^s_{\vec{p},q}(\mathbb{R}^n)$ into $B^{s-b}_{\vec{p},q}(\mathbb{R}^n)$.

math.FA

Unconditional bases for homogeneous $α$-modulation type spaces

In this article we construct orthonormal bases compatible with bi-variate homogeneous $α$-modulation spaces and the associated spaces of Triebel-Lizorkin type. The construction is based on generating a separable $α$-covering and using carefully selected tensor products of univariate brushlet functions with regards to this covering. We show that the associated systems form an unconditional bases for the homogeneous $α$-spaces of Triebel-Lizorkin type.

math.FA

Approximation spaces of deep neural networks

We study the expressivity of deep neural networks. Measuring a network's complexity by its number of connections or by its number of neurons, we consider the class of functions for which the error of best approximation with networks of a given complexity decays at a certain rate when increasing the complexity budget. Using results from classical approximation theory, we show that this class can be endowed with a (quasi)-norm that makes it a linear function space, called approximation space. We establish that allowing the networks to have certain types of "skip connections" does not change the resulting approximation spaces. We also discuss the role of the network's nonlinearity (also known as activation function) on the resulting spaces, as well as the role of depth. For the popular ReLU nonlinearity and its powers, we relate the newly constructed spaces to classical Besov spaces. The established embeddings highlight that some functions of very low Besov smoothness can nevertheless be well approximated by neural networks, if these networks are sufficiently deep.

math.FA

On a discrete transform of homogeneous decomposition spaces

We introduce almost diagonal matrices in the setting of (anisotropic) discrete homogeneous Triebel-Lizorkin type spaces and homogeneous modulation spaces, and it is shown that the class of almost diagonal matrices is closed under matrix multiplication. We then connect the results to the continuous setting and show that the "change of frame" matrix for a pair of time-frequency frames, with suitable decay properties, is almost diagonal. As an application of this result, we consider a construction of compactly supported frame expansions for homogeneous decomposition spaces of Triebel-Lizorkin type and for the associated modulation spaces.

math.FA

Fourier Multipliers on Anisotropic Mixed-Norm Spaces of Distributions

A new general Hormander type condition involving anisotropies and mixed norms is introduced, and boundedness results for Fourier multi- pliers on anisotropic Besov and Triebel-Lizorkin spaces of distributions with mixed Lebesgue norms are obtained. As an application, the continuity of such operators is established on mixed Sobolev and Lebesgue spaces too. Some lifting properties and equivalent norms are obtained as well.

math.FA

Nonlinear approximation with nonstationary Gabor frames

We consider sparseness properties of adaptive time-frequency representations obtained using nonstationary Gabor frames (NSGFs). NSGFs generalize classical Gabor frames by allowing for adaptivity in either time or frequency. It is known that the concept of painless nonorthogonal expansions generalizes to the nonstationary case, providing perfect reconstruction and an FFT based implementation for compactly supported window functions sampled at a certain density. It is also known that for some signal classes, NSGFs with flexible time resolution tend to provide sparser expansions than can be obtained with classical Gabor frames. In this article we show, for the continuous case, that sparseness of a nonstationary Gabor expansion is equivalent to smoothness in an associated decomposition space. In this way we characterize signals with sparse expansions relative to NSGFs with flexible time resolution. Based on this characterization we prove an upper bound on the approximation error occurring when thresholding the coefficients of the corresponding frame expansions. We complement the theoretical results with numerical experiments, estimating the rate of approximation obtained from thresholding the coefficients of both stationary and nonstationary Gabor expansions.

math.FA

On Homogeneous Decomposition Spaces and Associated Decompositions of Distribution Spaces

A new construction of decomposition smoothness spaces of homogeneous type is considered. The smoothness spaces are based on structured and flexible decompositions of the frequency space $\mathbb{R}^d\backslash\{0\}$. We construct simple adapted tight frames for $L_2(\mathbb{R}^d)$ that can be used to fully characterise the smoothness norm in terms of a sparseness condition imposed on the frame coefficients. Moreover, it is proved that the frames provide a universal decomposition of tempered distributions with convergence in the tempered distributions modulo polynomials. As an application of the general theory, the notion of homogeneous $α$-modulation spaces is introduced.

math.FA