
The present paper is concerned with the L2-approximation properties of exponential splines with n knots for functions defined on an interval ohm = (a, b). We first specify the Banach native space for these splines and show that it continuously embeds in L2(ohm). We derive sharp bounds on the approximation error rate for functions in this native space by exponential splines with n knots. To prove the optimality of these rates, we show that these native spaces are equivalent (as Banach spaces) to spaces of higher-order bounded variation on ohm. (c) 2026 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
In 2023, G & uuml;nt & uuml;rk and Thao proved that the sequence (x((n)))(n is an element of N) generated by random (relaxed) projections drawn from a finite innately regular collection of closed subspaces in a real Hilbert space satisfies & sum;(n is an element of N) parallel to x((n))-x(n+1)parallel to(gamma) < +infinity for all gamma > 0. We show that the G & uuml;nt & uuml;rk-Thao result is sharp in the sense that it fails for two closed subspaces that is not innately regular. We also extend their result to a finite collection of polyhedral cones. Moreover, we construct examples showing the tightness of our extension: indeed, the result fails for a line and a convex set in R-2, and for a plane and a nonpolyhedral cone in R-3.
Quasi-uniformity is a fundamental geometric property of point sets, crucial for applications such as kernel interpolation, Gaussian process regression, and space-filling experimental designs. While quasi-Monte Carlo methods are widely recognized for their low-discrepancy characteristics, understanding their quasi-uniformity remains important for practical applications. For the two-dimensional Sobol' sequence, Sobol' and Shukhman (2007) conjectured that the separation radius of the first N points achieves the optimal rate N-1/2, which would imply quasi-uniformity. This conjecture was disproved by Goda (2024), who computed exact values of the 2-separation pound radius for a sparse subsequence N = 22v-1. However, the general behavior of the Sobol' sequence for arbitrary N remained unclear. In this paper, we derive exact expressions for the pound infinity-separation radius of the first N = 2m points of the two-dimensional Sobol' sequence for all m is an element of N. As an immediate consequence, we show that the separation radius of Sobol' points is O(N-3/4), which is strictly worse than the optimal rate N-1/2, revealing that the two-dimensional Sobol' sequence has a suboptimal mesh ratio that grows at least as N1/4. (c) 2026 Elsevier Inc. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
This paper initiates a theoretical investigation of 41-oscillatory chain sequences {an}, generalizing Szwarc's classical framework for non-oscillatory chains (Szwarc, 1994, 1998, 2002, 2003) to sequences fluctuating around 41. We prove the existence of a fixed point for the critical map f (x) = 1-41x and establish convergence properties linking oscillatory behavior to parameter sequences {gn}. A complete characterization is provided via a necessary and sufficient condition, exemplified by explicit solutions an = 1 (1 + (-1)n epsilon n ). Crucially, we construct oscillatory chain sequences for which the series & sum;infinity ( ) 4 an-1 n=1 diverges, demonstrating fundamentally different behavior outside the hypothesis an >= 1 4 4 required by Chihara's bound. (c) 2026 Published by Elsevier Inc.
This paper studies binary classification using ReLU neural networks within the statistical learning framework. Unlike existing approaches that impose sparsity constraints on hidden neurons and boundedness assumptions on output functions, we consider a hypothesis space generated by a fixed hard-tanh layer and a trainable shallow ReLU layer. Using recent advances in approximation theory and refined complexity control for this hypothesis space, we derive optimal learning rates, up to logarithmic factors, for the excess misclassification error under smoothness assumptions on the regression function and the Tsybakov noise condition. (c) 2026 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).