OctoInformBeta

OpenAI preprint, September 2026 · §1. Introduction

Theorem 1.1: Integrability on symmetric annuli and measurability almost everywhere for F,G∈L3(ℝ2)F, G \in L^3(\mathbb{R}^2), with cyclic-coordinate L3/2L^{3/2} bounds

An L3/2L^{3/2} estimate for the maximal triangular Hilbert transform.

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

Take complex functions F,G∈L3(ℝ2)F, G \in L^3(\mathbb{R}^2), a point (x,y)∈ℝ2(x, y) \in \mathbb{R}^2, and integrate F(x+t,y)G(x,y+t)/tF(x + t, y)\,G(x, y + t)/t over symmetric annuli ε<|t|<R\varepsilon < |t| < R with 0<ε<R0 < \varepsilon < R. The question is whether these truncations are locally integrable on symmetric annuli almost everywhere with a maximal operator measurable almost everywhere, and whether that maximal operator satisfies an L3(ℝ2)×L3(ℝ2)→L3/2(ℝ2)L^3(\mathbb{R}^2) \times L^3(\mathbb{R}^2) \to L^{3/2}(\mathbb{R}^2) bound.

Truncations and Maximal Operators

For ε,R∈ℝ\varepsilon, R \in \mathbb{R}, the symmetric annulus Aε,R⊂ℝA_{\varepsilon, R} \subset \mathbb{R} is the set {t∈ℝ:ε<|t|<R}\{t \in \mathbb{R} : \varepsilon < |t| < R\}.✦ A pair (ε,R)∈ℝ2(\varepsilon, R) \in \mathbb{R}^2 of real truncation endpoints is a pair in ℝ2\mathbb{R}^2 satisfying 0<ε<R0 < \varepsilon < R.✦

Given functions F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C} and a point z=(x,y)∈ℝ2z = (x, y) \in \mathbb{R}^2, the triangular Hilbert integrand UF,G(z,⋅):ℝ→ℂU_{F, G}(z, \cdot) : \mathbb{R} \to \mathbb{C} is the function defined, with value 00 at t=0t = 0, by UF,G(z,t)=F(x+t,y)G(x,y+t)t.U_{F, G}(z, t) = \frac{F(x + t, y)\,G(x, y + t)}{t}.(1)✦ For real truncation endpoints (ε,R)(\varepsilon, R), the truncated triangular Hilbert transform Tε,R(F,G)(z)∈ℂT_{\varepsilon, R}(F, G)(z) \in \mathbb{C} of F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C} at z∈ℝ2z \in \mathbb{R}^2 is the integral of (1) over Aε,RA_{\varepsilon, R}, given by (2): Tε,R(F,G)(z)=∫Aε,RUF,G(z,t)dt.T_{\varepsilon, R}(F, G)(z) = \int_{A_{\varepsilon, R}} U_{F, G}(z, t)\,dt.(2)✦ A locally integrable point z∈ℝ2z \in \mathbb{R}^2 for F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C} is a point in ℝ2\mathbb{R}^2 such that for every integer n≥0n \ge 0, the integrand UF,G(z,⋅)U_{F, G}(z, \cdot) is integrable on A(n+2)−1,n+2A_{(n+2)^{-1},\, n+2}.✦

For complex functions F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C} and a point z∈ℝ2z \in \mathbb{R}^2, the triangular maximal operator B*(F,G)(z)∈[0,∞]B_*(F, G)(z) \in [0, \infty] is the supremum of the modulus of the truncated triangular Hilbert transform Tε,R(F,G)(z)T_{\varepsilon, R}(F, G)(z) over finite hard-truncation annuli Aε,RA_{\varepsilon, R} with 0<ε<R0 < \varepsilon < R when zz is a locally integrable point for FF and GG, and 00 elsewhere.✦ The maximal L3/2L^{3/2} norm ∥B*(F,G)∥L3/2(ℝ2)∈[0,∞]\|B_*(F, G)\|_{L^{3/2}(\mathbb{R}^2)} \in [0, \infty] of F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C} is the value in [0,∞][0, \infty] defined by (3): ∥B*(F,G)∥L3/2(ℝ2)=(∫ℝ2B*(F,G)(z)3/2dz)2/3.\|B_*(F, G)\|_{L^{3/2}(\mathbb{R}^2)} = \left(\int_{\mathbb{R}^2} B_*(F, G)(z)^{3/2}\,dz\right)^{2/3}.(3)✦

Cyclic Coordinates and the Bound Constant

The cyclic coordinate charts ϕB,ϕC:ℝ2→ℝ2\phi_B, \phi_C : \mathbb{R}^2 \to \mathbb{R}^2 are the linear maps ϕB(a,b)=(a+b,−a)\phi_B(a, b) = (a + b, -a)✦ and ϕC(a,b)=(−b,a+b)\phi_C(a, b) = (-b, a + b)✦. For complex functions F,G:ℝ2→ℂF, G : \mathbb{R}^2 \to \mathbb{C}, the cyclic coordinate pullbacks BF,CG:ℝ2→ℂB_F, C_G : \mathbb{R}^2 \to \mathbb{C} are the compositions BF=F∘ϕBB_F = F \circ \phi_B✦ and CG=G∘ϕCC_G = G \circ \phi_C✦.

A pair (r,R)∈ℚ2(r, R) \in \mathbb{Q}^2 of rational truncation endpoints is a pair in ℚ2\mathbb{Q}^2 satisfying 0<r<R0 < r < R.✦ For r,R∈ℝr, R \in \mathbb{R}, the hard-truncated Hilbert kernel Kr,R:ℝ→ℝK_{r, R} : \mathbb{R} \to \mathbb{R} is the function given, with value 00 at t=0t = 0, by Kr,R(t)=𝟏r<|t|−𝟏R<|t|t.K_{r, R}(t) = \frac{\mathbf{1}_{r < |t|} - \mathbf{1}_{R < |t|}}{t}.(4)✦

Given B,C:ℝ2→ℂB, C : \mathbb{R}^2 \to \mathbb{C} and a kernel K:ℝ→ℝK : \mathbb{R} \to \mathbb{R}, the cyclic kernel operator 𝒯K(B,C):ℝ2→ℂ\mathcal{T}_K(B, C) : \mathbb{R}^2 \to \mathbb{C} is the function evaluated at z=(z1,z2)∈ℝ2z = (z_1, z_2) \in \mathbb{R}^2 by 𝒯K(B,C)(z)=∫ℝB(z2,c)C(c,z1)K(z1+z2+c)dc.\mathcal{T}_K(B, C)(z) = \int_{\mathbb{R}} B(z_2, c)\,C(c, z_1)\,K(z_1 + z_2 + c)\,dc.(5)✦ The cyclic maximal operator Mcyc(B,C):ℝ2→[0,∞]M_{\mathrm{cyc}}(B, C) : \mathbb{R}^2 \to [0, \infty] of B,C:ℝ2→ℂB, C : \mathbb{R}^2 \to \mathbb{C} is the supremum of (5) with (4) over rational truncation endpoints (r,R)∈ℚ2(r, R) \in \mathbb{Q}^2 satisfying 0<r<R0 < r < R: Mcyc(B,C)(z)=sup(r,R)∈ℚ20<r<R|𝒯Kr,R(B,C)(z)|.M_{\mathrm{cyc}}(B, C)(z) = \sup_{\substack{(r, R) \in \mathbb{Q}^2 \\ 0 < r < R}} \left|\mathcal{T}_{K_{r, R}}(B, C)(z)\right|.(6)✦ For functions B,C:ℝ2→ℂB, C : \mathbb{R}^2 \to \mathbb{C}, the cyclic L3/2L^{3/2} norm ∥Mcyc(B,C)∥L3/2(ℝ2)∈[0,∞]\|M_{\mathrm{cyc}}(B, C)\|_{L^{3/2}(\mathbb{R}^2)} \in [0, \infty] is the value in [0,∞][0, \infty] defined from (6) by (7): ∥Mcyc(B,C)∥L3/2(ℝ2)=(∫ℝ2Mcyc(B,C)(z)3/2dz)2/3.\|M_{\mathrm{cyc}}(B, C)\|_{L^{3/2}(\mathbb{R}^2)} = \left(\int_{\mathbb{R}^2} M_{\mathrm{cyc}}(B, C)(z)^{3/2}\,dz\right)^{2/3}.(7)✦

For σ>0\sigma > 0, the variance-σ\sigma Gaussian density gσ:ℝ→ℝg_\sigma : \mathbb{R} \to \mathbb{R} is the function defined by (8): gσ(t)=(2πσ)−1/2exp⁡(−t22σ).g_\sigma(t) = (2\pi\sigma)^{-1/2} \exp\!\left(-\frac{t^2}{2\sigma}\right).(8)✦ Using (8), the heat kernel constant κ\kappa is the real number −23g3(0)-\frac{2}{3} g_3(0)✦, and the smooth kernel profile k:ℝ→ℝk : \mathbb{R} \to \mathbb{R} is the function defined by (9): k(v)=κ−2(v29−13)g3(v).k(v) = \kappa - 2\!\left(\frac{v^2}{9} - \frac{1}{3}\right) g_3(v).(9)✦ The kernel error profile E:ℝ→ℝE : \mathbb{R} \to \mathbb{R} is the function given from (9) by E(v)=k(v)−κ𝟏|v|>1E(v) = k(v) - \kappa\,\mathbf{1}_{|v| > 1}.✦ There exists a constant K>0K > 0 such that |E(v)|≤K|v|g6(v)(v∈ℝ).|E(v)| \le K |v|\, g_6(v) \qquad (v \in \mathbb{R}).(10)✦ An error envelope constant Kenv>0K_{\mathrm{env}} > 0 is a positive constant satisfying (10).✦

The mesh mass constant mm is the real number 2e1/22e^{1/2}✦, the jump scalar constant jj is the real number 2+4822 + 48\sqrt{2}✦, and the strict positivity constant δ\delta is the real number 12min⁡(1−352,1110−122)\frac{1}{2}\min\bigl(1 - \frac{3}{5}\sqrt{2},\, \frac{11}{10} - \frac{1}{2}\sqrt{2}\bigr).✦ In terms of KenvK_{\mathrm{env}}, κ\kappa, mm, jj, and δ\delta, the masked constant M0M_0, selected scale constant S0S_0, error pairing constant ηerr\eta_{\mathrm{err}}, hard pairing constant ηhard\eta_{\mathrm{hard}}, and moment bound constant μ\mu are the real numbers M0=m(22+3j),S0=4403⋅3⋅M0δ,ηerr=Kenvmj,ηhard=S0+2ηerr|κ|,μ=4ηhard(ηhard+1).\begin{aligned} M_0 &= m(2\sqrt{2} + 3j), & S_0 &= \frac{440}{3} \cdot 3 \cdot \frac{M_0}{\delta}, & \eta_{\mathrm{err}} &= K_{\mathrm{env}}\, m\, j, \\ \eta_{\mathrm{hard}} &= \frac{S_0 + 2\eta_{\mathrm{err}}}{|\kappa|}, & \mu &= 4\eta_{\mathrm{hard}}(\eta_{\mathrm{hard}} + 1). \end{aligned}(11) From (11), the main estimate constant C0C_0 is the nonnegative real number (2μ)2/3(2\mu)^{2/3}.✦

The Claim

For any functions B,C∈L3(ℝ2)B, C \in L^3(\mathbb{R}^2), the cyclic L3/2L^{3/2} norm (7) obeys (12): ∥Mcyc(B,C)∥L3/2(ℝ2)≤C0∥B∥L3(ℝ2)∥C∥L3(ℝ2).\|M_{\mathrm{cyc}}(B, C)\|_{L^{3/2}(\mathbb{R}^2)} \le C_0\,\|B\|_{L^3(\mathbb{R}^2)}\,\|C\|_{L^3(\mathbb{R}^2)}.(12)✦ Given F,G∈L3(ℝ2)F, G \in L^3(\mathbb{R}^2), the maximal L3/2L^{3/2} norm (3) and the cyclic pullbacks BF,CGB_F, C_G satisfy (13): ∥B*(F,G)∥L3/2(ℝ2)=∥Mcyc(BF,CG)∥L3/2(ℝ2).\|B_*(F, G)\|_{L^{3/2}(\mathbb{R}^2)} = \|M_{\mathrm{cyc}}(B_F, C_G)\|_{L^{3/2}(\mathbb{R}^2)}.(13)✦ If F∈L3(ℝ2)F \in L^3(\mathbb{R}^2), the identity (14) holds for BFB_F: ∥BF∥L3(ℝ2)=∥F∥L3(ℝ2).\|B_F\|_{L^3(\mathbb{R}^2)} = \|F\|_{L^3(\mathbb{R}^2)}.(14)✦ When G∈L3(ℝ2)G \in L^3(\mathbb{R}^2), the pullback CGC_G satisfies (15): ∥CG∥L3(ℝ2)=∥G∥L3(ℝ2).\|C_G\|_{L^3(\mathbb{R}^2)} = \|G\|_{L^3(\mathbb{R}^2)}.(15)✦

Theorem 1. There exists a real constant C≥0C \ge 0 such that, for every pair of functions F,G∈L3(ℝ2)F, G \in L^3(\mathbb{R}^2), almost every z∈ℝ2z \in \mathbb{R}^2 is a locally integrable point for FF and GG, the maximal operator B*(F,G)B_*(F, G) is measurable almost everywhere on ℝ2\mathbb{R}^2, and the estimate (16) holds: ∥B*(F,G)∥L3/2(ℝ2)≤C∥F∥L3(ℝ2)∥G∥L3(ℝ2).\|B_*(F, G)\|_{L^{3/2}(\mathbb{R}^2)} \le C\,\|F\|_{L^3(\mathbb{R}^2)}\,\|G\|_{L^3(\mathbb{R}^2)}.(16)✦

In the paper, the result reads: there is an absolute constant CC such that, for every pair of complex functions F,G∈L3(ℝ2)F, G \in L^3(\mathbb{R}^2), ∥B*(F,G)∥L3/2(ℝ2)≤C∥F∥L3(ℝ2)∥G∥L3(ℝ2)\|B_*(F, G)\|_{L^{3/2}(\mathbb{R}^2)} \le C\|F\|_{L^3(\mathbb{R}^2)}\|G\|_{L^3(\mathbb{R}^2)}, where the supremum is over both finite hard-truncation endpoints.

What Is Checked

The estimates are stated over complex functions F,G,B,C∈L3(ℝ2)F, G, B, C \in L^3(\mathbb{R}^2), the locally integrable point condition on A(n+2)−1,n+2A_{(n+2)^{-1},\, n+2} for integers n≥0n \ge 0, the [0,∞][0, \infty]-valued maximal operators B*(F,G)B_*(F, G) and Mcyc(B,C)M_{\mathrm{cyc}}(B, C) with their L3/2(ℝ2)L^{3/2}(\mathbb{R}^2) norms, the cyclic pullbacks BFB_F and CGC_G, and the nonnegative constant C0C_0.