OctoInformBeta

OpenAI preprint, September 2026 · §1. Introduction · §1

Theorem 1.1: For f∈L1(ℝ2)f\in L^1(\mathbb{R}^2) with integrable weak gradient gg, MfMf lies in Wloc1,1W^{1,1}_{\mathrm{loc}} with a weak gradient GG satisfying ∫|G|≤C∫|g|\int|G|\le C\int|g|

An endpoint gradient bound for the centered disk maximal function.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

On the Euclidean plane ℝ2\mathbb{R}^2, let MfMf be the centered Hardy-Littlewood maximal function of a function ff, taken over disks. For p>1p>1, Kinnunen’s theorem bounds the LpL^p norm of ∇Mf\nabla Mf by a constant multiple of that of ∇f\nabla f. At the endpoint p=1p=1 the function MfMf is not integrable unless f=0f=0, and the question is whether the weak gradient ∇Mf\nabla Mf still is integrable, whenever ff and its weak gradient ∇f\nabla f are.

Disk Averages and Weak Gradients

Vectors in ℝ2\mathbb{R}^2 carry the Euclidean norm |⋅||\cdot|, and B(x,t)={y:|y−x|<t}B(x,t)=\{y:|y-x|<t\} is the open disk of center xx and radius tt. For a real function ff on ℝ2\mathbb{R}^2, a point xx and a radius t>0t>0, the absolute disk average At|f|(x)A_t|f|(x) is the mean of |f||f| over B(x,t)B(x,t), a number in [0,∞][0,\infty]: At|f|(x)=1πt2∫B(x,t)|f(y)|dy.A_t|f|(x)=\frac{1}{\pi t^2}\int_{B(x,t)}|f(y)|\,dy.(1)✦ The centered disk maximal function of ff is the function MfMf on ℝ2\mathbb{R}^2 with values in [0,∞][0,\infty] given by Mf(x)=supt>0At|f|(x).Mf(x)=\sup_{t>0}A_t|f|(x).(2)✦ The real maximal function M̃f\widetilde{M}f of ff is the real function on ℝ2\mathbb{R}^2 equal to Mf(x)Mf(x) at the points xx where Mf(x)Mf(x) is finite and to 00 at the others.✦

A vector field gg on ℝ2\mathbb{R}^2, with components g1g_1 and g2g_2, is a weak gradient of a real function ff on ℝ2\mathbb{R}^2 if, for every smooth compactly supported real function φ\varphi on ℝ2\mathbb{R}^2 and for i=1,2i=1,2, ∫ℝ2f∂iφdx=−∫ℝ2giφdx,\int_{\mathbb{R}^2}f\,\partial_i\varphi\,dx=-\int_{\mathbb{R}^2}g_i\,\varphi\,dx,(3)✦ where ∂iφ\partial_i\varphi is the derivative of φ\varphi along the ii-th coordinate vector. The local Sobolev space Wloc1,1(ℝ2)W^{1,1}_{\mathrm{loc}}(\mathbb{R}^2) is the set of locally integrable real functions on ℝ2\mathbb{R}^2 that have a locally integrable weak gradient.✦ The L1L^1 norm of a vector field gg on ℝ2\mathbb{R}^2 is ∥g∥L1=∫ℝ2|g(x)|dx\|g\|_{L^1}=\int_{\mathbb{R}^2}|g(x)|\,dx.✦

The Claim

Theorem 1. There is a constant C≥0C\ge 0 with the following property. Let ff be an integrable real function on ℝ2\mathbb{R}^2 with an integrable weak gradient gg. Then Mf(x)<∞Mf(x)<\infty for almost every x∈ℝ2x\in\mathbb{R}^2, the real maximal function M̃f\widetilde{M}f belongs to Wloc1,1(ℝ2)W^{1,1}_{\mathrm{loc}}(\mathbb{R}^2), and M̃f\widetilde{M}f has an integrable weak gradient GG with ∥G∥L1≤C∥g∥L1.\|G\|_{L^1}\le C\,\|g\|_{L^1}.(4)✦

In the paper, the result reads: there is an absolute constant C<∞C<\infty such that, for every real-valued f∈W1,1(ℝ2)f\in W^{1,1}(\mathbb{R}^2), MfMf is locally integrable, belongs to Wloc1,1(ℝ2)W^{1,1}_{\mathrm{loc}}(\mathbb{R}^2) and satisfies ∫|∇Mf|≤C∫|∇f|\int|\nabla Mf|\le C\int|\nabla f|, the distributional derivatives of MfMf being represented by globally integrable functions; the paper’s ∇f\nabla f and ∇Mf\nabla Mf are the page’s gg and GG.

What Is Checked

The claim (4) is stated over the Euclidean plane ℝ2\mathbb{R}^2, the absolute disk averages (1), the maximal function (2) and the real maximal function M̃f\widetilde{M}f, weak gradients in the sense of (3), the local Sobolev space Wloc1,1(ℝ2)W^{1,1}_{\mathrm{loc}}(\mathbb{R}^2) and the L1L^1 norm of a vector field.

For every integrable real function ff on ℝ2\mathbb{R}^2, weakly differentiable or not, the library proves that Mf(x)<∞for almost every x∈ℝ2.Mf(x)<\infty\quad\text{for almost every }x\in\mathbb{R}^2.(5)✦