arXiv Science⌕ Search

arXiv · 2611.16448

Weighted STSAFT-Modulation Spaces and Approximation by Phase-Adapted Neural Networks

Abstract

We develop a transform-adapted approximation framework for shallow neural networks based on the short-time special affine Fourier transform (STSAFT). Replacing the short-time Fourier transform (STFT) in the definition of classical modulation spaces, we work with weighted STSAFT-modulation spaces, building on the existing SAFT-associated modulation-space framework, and use their chirp correspondence with classical weighted modulation spaces through the chirp factorization of the SAFT kernel. We construct an associated neural dictionary and establish an $N$-term Sobolev approximation bound of order $N^{-1/2}$ on bounded one-dimensional domains. A global approximation theorem is also obtained when the spatial translation parameter ranges over the whole real line; a compact-support argument explains why an unrestricted global conclusion cannot in general be inferred from a bounded translation range. For a fixed-scale chirp family $f_κ(x)=e^{iπκx^2}u(x)$, we combine the global estimate with a phase analysis to obtain the stronger phase-dependent upper-bound factor $\langleλ\rangle^n\langleκ-λ\rangle^{n+3/2-1/r}.$ For each fixed nonzero Schwartz envelope, matching the dictionary phase reduces the growth of this specified upper bound from order $|κ|^{n+3/2-1/r}$ to order $|κ|^n$, while preserving the $N^{-1/2}$ rate. The associated mismatch scale is stated as a comparison of upper bounds, not as a threshold for the optimal approximation error. Four numerical experiments examine model comparison, error decay with dictionary size, phase mismatch, and the obstruction caused by restricted translations. We discuss the effective number of fitted output coefficients, including the rank loss of the sine branch at zero phase, and distinguish the numerical diagnostics from the theoretical best-approximation estimates.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Ming Hu, Bei Liu, Bolin Liu, Jie Shen. 2026-11-12. Weighted STSAFT-Modulation Spaces and Approximation by Phase-Adapted Neural Networks. https://arxiv.science/abs/2611.16448

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Annihilating ideals and Agler--McCarthy spectral varieties in the bidisc

The closed unit bidisc $\overline{\mathbb{D}^2}$ is known to be a spectral set for any pair $(T_1,T_2)$ of commuting contractions. When each $T_i$ is pure and has finite defect, the pair admits a much smaller spectral set: the closure of a distinguished variety $V\subset \mathbb{D}^2$. We exhibit conditions on $(T_1,T_2)$ that guarantee that no proper closed subset of $\overline{V}$ can be a spectral set. In addition, we examine the relationship between $V$ and the annihilating ideal $Ann(T_1,T_2)$ in $H^\infty(\mathbb{D}^2)$. While $V$ is typically strictly larger than the zero set of $Ann(T_1,T_2)$, we isolate a natural constrained co-extension $(S_1,S_2)$ of $(T_1,T_2)$ whose Taylor spectrum is contained in $V$ and is closely linked to the so-called support of $Ann(T_1,T_2)$. We also find simple conditions on $Ann(T_1)$ detecting whether $Ann(T_1,T_2)$ is the ideal of functions vanishing on the joint point spectrum of $(S_1^*,S_2^*)$.

math.FA↗

On Talagrand Type

We prove that for every $1<s<2$ and every $0<δ<1$, Rademacher type $s$ coincides with Talagrand type $(s,ψ_{s,δ})$. Together with the result of Cordero-Erausquin and Eskenazis for $s=2$, this yields the equivalence for every $1<s\le2$ and every $0<δ<1$.

math.FA↗

Weighted orbits of hypercyclic vectors on F-spaces

We study $\mathcal{F}$-hypercyclicity of weighted orbits $λ_nT^nx$ for continuous linear operators $T$ on complex F-spaces, where $\mathcal{F}$ is a Furstenberg family with the cut--shift--and--paste (CuSP) property. We characterize the scalar sequences for which weighted hypercyclicity always implies ordinary hypercyclicity: their nonzero values must be bounded and bounded away from zero. This condition also guarantees weight removal for every Furstenberg family with the CuSP property. For the converse inclusion, we prove that, for any Furstenberg family $\mathcal{F}$ with the CuSP property, a bounded scalar sequence preserves all $\mathcal{F}$-hypercyclic vectors if its backward-shift $ω$-limit system is distal and does not contain the zero sequence. For chaotic operators, minimality of this system suffices. In particular, every complex nilsequence which is not identically zero preserves $\mathcal{F}$-hypercyclicity, and equality holds when its nonzero values are bounded away from zero. We also obtain simultaneous weighted return results and stability under unimodular weights with asymptotically trivial multiplicative differences.

math.FA↗