Paperlane

no. 7 · Tuesday 3 November 2026

A direct obstacle proof of a dimension-free weak-type $(1,1)$ bound for the Gaussian Riesz transforms

Cristian E. Gutiérrez and Juan Manfredi

math.AP · v2 · code · ∎ lean builds · 33 pp

Abstract

We give a detailed, direct proof of a dimension-free weak-type $(1,1)$ estimate for the first-order Gaussian vector Riesz transform, using the fractional obstacle method of Ouyang, Spector, and Stockdale. For $dγ_n=π^{-n/2}e^{-|x|^2}\,dx$, let $L=-\tfrac12Δ+x\cdot\nabla$, $D=2^{-1/2}\nabla$, and $R_γ=DL^{-1/2}Π_0$, where $Π_0$ removes the mean. We prove that $R_γ$ has weak-type constant at most $2$ on real-valued inputs, independently of $n$. The main construction is the unique unshifted decomposition $F=g+L^{1/2}u$, for nonnegative $F\in L^2(γ_n)$ and $a>\int F\,dγ_n$, with $u\geq0$, $\min\{F,a\}\leq g\leq a$, preservation of mass, and $g=a$ on $\{u>0\}$. We retain the details of coercivity, the Lewy–Stampacchia estimate, the upgrade from the form domain to the operator domain, and locality. We also identify the $L^1$ extension with the principal-value transform and obtain the constant $8/π$ for complex-valued inputs.

code
github.com/na0b/gaussianriesz-lean ∎ Lean project (lakefile.lean, lean-toolchain) builds, no sorry at dc16c5d, checked 2026-11-03 build details
comments
33 pages. A Lean 4 formalization of the results of this paper is available at https://github.com/na0b/gaussianriesz-lean and at https://palomar-registry.org/entry?id=PALOMAR-2026-10-30-000009&version=1
fields
Analysis of PDEs, Functional Analysis, Probability
pdf ↗ page arXiv:2610.26478 ↗