Effective-dimension bounds in the central limit theorem for centered Euclidean balls
Abstract
We establish a nonasymptotic Gaussian approximation bound over centered Euclidean balls for sums of independent, not necessarily identically distributed, random vectors in ℝp. Let Sn denote the sum, let Σ=Var(Sn), and let G∼ N(0,Σ). Define the effective rank and the normalized squared spectral mass by r(Σ)=trΣ/|Σ|, v(Σ)=tr(Σ2)/|Σ|2. Under a uniform directional fourth-moment condition, covariance-balance assumptions, and the spectral conditions r(Σ)≤ CΣv(Σ) and v(Σ)≥6, we prove supz≥0 | ℙ{|Sn|2≤ z} − ℙ{|G|2≤ z} | ≤ C r(Σ)3/2/n. The constant depends only on the fixed parameters in the assumptions and is independent of the ambient dimension and individual small eigenvalues of the covariance matrix. The bound is uniform at the endpoint v(Σ)=6, without a logarithmic loss. Its main feature is the combination of the n−1 rate with dependence on effective rank, including this endpoint. Consequently, the bound can vanish when r(Σ)=o(n2/3), even when p is much larger than n. The same estimate extends to random elements of a separable Hilbert space with trace class covariance. The proof builds on the Fourier and symmetrization methods of \citet{GoetzeZaitsev2014}, combined with moment matching, truncation, and comparison estimates that retain the covariance structure.
Related
- Bellman Search in Arbitrary Finite Dimension: A Self-Similar Cell Theorem and Effective Computability of Planar Shoreline Search
- Superdiffusive central limit theorem for a class of driven diffusive systems at the critical dimension
- Exact cost of the Euclidean Maximum Traveling Salesman Problem in one dimension
- Full-radius dimension-free maximal inequalities for discrete Euclidean balls
- A Frankel type theorem in Euclidean and hyperbolic spaces
BibTeX
@misc{naumov2026effectivedimension,
title = {{Effective-dimension bounds in the central limit theorem for centered Euclidean balls}},
author = {Alexey Naumov and Sergey Samsonov and Vladimir Spokoiny and Vladimir Ulyanov},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12722},
eprint = {2610.12722},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12722},
}
RIS
TY - UNPB
TI - Effective-dimension bounds in the central limit theorem for centered Euclidean balls
AU - Alexey Naumov
AU - Sergey Samsonov
AU - Vladimir Spokoiny
AU - Vladimir Ulyanov
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.12722
UR - https://arxiv.org/abs/2610.12722
AB - We establish a nonasymptotic Gaussian approximation bound over centered Euclidean balls for sums of independent, not necessarily identically distributed, random vectors in ℝ^p. Let S_n denote the sum, let Σ=Var(S_n), and let G∼ N(0,Σ). Define the effective rank and the normalized squared spectral mass by r(Σ)=trΣ/|Σ|, v(Σ)=tr(Σ^2)/|Σ|^2. Under a uniform directional fourth-moment condition, covariance-balance assumptions, and the spectral conditions r(Σ)≤ C_Σv(Σ) and v(Σ)≥6, we prove sup_z≥0 | ℙ{|S_n|^2≤ z} − ℙ{|G|^2≤ z} | ≤ C r(Σ)^3/2/n. The constant depends only on the fixed parameters in the assumptions and is independent of the ambient dimension and individual small eigenvalues of the covariance matrix. The bound is uniform at the endpoint v(Σ)=6, without a logarithmic loss. Its main feature is the combination of the n^−1 rate with dependence on effective rank, including this endpoint. Consequently, the bound can vanish when r(Σ)=o(n^2/3), even when p is much larger than n. The same estimate extends to random elements of a separable Hilbert space with trace class covariance. The proof builds on the Fourier and symmetrization methods of \citet{GoetzeZaitsev2014}, combined with moment matching, truncation, and comparison estimates that retain the covariance structure.
ER -