Skip to content
DraconiticResearch
Search Sources API
184,937 studies31 sourcesUpdated 47 min ago
PreprintarXiv licencearXiv

Approximation numbers of critical Besov embeddings

Abstract

We determine the approximation numbers of critical embeddings between isotropic Besov spaces on nonempty bounded Lipschitz domains in any dimension. For 0 < p_1 < 2 < p_2 <= infinity and s_1 - s_2 = d max{1/p_1, 1 - 1/p_2} > d(1/p_1 - 1/p_2), the approximation numbers have order N^{-1/2}(log(e+N))^Gamma for every 0 < q_1, q_2 <= infinity, with an explicit four-case formula for Gamma. This resolves the strictly compact critical Besov approximation-number problem identified in Vybiral's overview. For q_1 < q_2, the logarithmic exponent depends on the ordering of p_1' and p_2; it can vanish for one pair (p_1,p_2) and remain positive for another with the same fine indices. A row estimate for the approximation error gives lower bounds when the target fine index is below one. Gaussian finite-rank approximations acting on several levels give the sharp upper bounds when Gamma < 1/2. For q_1 < q_2, matching lower bounds for the two strict orderings of the inner indices follow from Malykhin's finite-window estimates and norm comparisons. The results include quasi-Banach indices and the full spaces of bounded coefficient sequences at infinite fine indices.

Related

  1. Sharp approximation rates for shallow ReLU neuralnetworks on critical Besov classes
    Yupeng Wang
    arXivPreprintarXiv licenceFull textNumerical analysis & scientific computingarXiv
  2. Sharp embeddings between quasi-Banach Besov spaces and shallow ReLU variation spaces
    Yuwen Li, Yupeng Wang
    arXivPreprintarXiv licenceFull textAnalysis & PDEsarXiv
  3. Jackson and Bernstein Inequalities with Explicit Weight Dependence for Band-Limited Wavelet Approximation in Weighted Besov Spaces
    Hung-Lung Lin, Kai-Cheng Wang
    arXivPreprintarXiv licenceFull textAnalysis & PDEsarXiv
  4. Global regularity and sharp decay rates for the one-dimensional hypo-viscous compressible Navier-Stokes equations in critical Besov spaces
    Chen Liang, Zhaonan Luo, Zhaoyang Yin
    arXivPreprintarXiv licenceFull textAnalysis & PDEsarXiv
  5. Gaussian Approximation of Besov Functions for Logistic Classification on Manifolds
    Guanhang Lei, Lei Shi, Xiaoming Yuan
    arXivPreprintCC BY 4.0Full textStatisticsarXiv
BibTeX
@misc{guo2026approximation,
  title = {{Approximation numbers of critical Besov embeddings}},
  author = {Feng Guo},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.25910},
  eprint = {2610.25910},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.25910},
}
RIS
TY  - UNPB
TI  - Approximation numbers of critical Besov embeddings
AU  - Feng Guo
PY  - 2026
DA  - 2026/10/19
PB  - arXiv
DO  - 10.48550/arxiv.2610.25910
UR  - https://arxiv.org/abs/2610.25910
AB  - We determine the approximation numbers of critical embeddings between isotropic Besov spaces on nonempty bounded Lipschitz domains in any dimension. For 0 < p_1 < 2 < p_2 <= infinity and s_1 - s_2 = d max{1/p_1, 1 - 1/p_2} > d(1/p_1 - 1/p_2), the approximation numbers have order N^{-1/2}(log(e+N))^Gamma for every 0 < q_1, q_2 <= infinity, with an explicit four-case formula for Gamma. This resolves the strictly compact critical Besov approximation-number problem identified in Vybiral's overview. For q_1 < q_2, the logarithmic exponent depends on the ordering of p_1' and p_2; it can vanish for one pair (p_1,p_2) and remain positive for another with the same fine indices. A row estimate for the approximation error gives lower bounds when the target fine index is below one. Gaussian finite-rank approximations acting on several levels give the sharp upper bounds when Gamma < 1/2. For q_1 < q_2, matching lower bounds for the two strict orderings of the inner indices follow from Malykhin's finite-window estimates and norm comparisons. The results include quasi-Banach indices and the full spaces of bounded coefficient sequences at infinite fine indices.
ER  -