Theorem A.1: Krivine preprocessing and projection,
OpenGrothendieckLB.Supplement.theorem_A_1gaussian-analysisgrothendieck-constantkrivine-roundingp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be odd measurable functions with arcsine-normalized correlation function
where are independent standard Gaussian vectors in . Suppose that has a local inverse at the origin, , which is holomorphic on a neighborhood of , and that
Then
This is Krivine's preprocessing-and-projection argument for one ordinary scheme; Krivine's bound is its special case for the hyperplane scheme.
Formalization Note is the published KrivineScheme k and its correlationFunction (the pair has coordinatewise correlation ). "Holomorphic on a neighborhood of " is for some . is the published grothendieckConst.
Differences from the source:
- encoding: γ D with γ > 0 implicit (a disk of radius γ) → explicit hypothesis 0 < γ; reason: the page's γD is a disk only for γ > 0; at γ = 0 the hypotheses hold trivially and π/0 = 0 in Lean, so the condition only excludes the undefined case
- encoding: H⁻¹ holomorphic on a neighborhood of γD → ∑ |a_r| R^r converges for some R > γ; reason: a neighborhood of the compact disk γD contains a disk of radius R > γ, and conversely; same instances
Preamble
import Mathlib import Definitions.Def_GrothendieckConstantDefs import Definitions.Def_KrivineSchemeDefs import Definitions.Def_GrothendieckLB_IsLocalInverseBranch import Definitions.Def_GrothendieckLB_taylorCoeff
Formal statement
namespace GrothendieckLB
namespace Supplement
/-- Theorem A.1, Krivine preprocessing and projection (Saha et al., *New Lower and Upper Bounds for the Grothendieck Constant*, arXiv:2608.11158 (held version, sha256 e7eb5d65…), Theorem A.1, p.47): let
`f, g : ℝ^k → {−1, 1}` be odd measurable functions with arcsine-normalized correlation function
`H(t) = (π/2) E[f(G₁) g(tG₁ + √(1 − t²) G₂)]`, `−1 < t < 1`. Suppose that `H` has a local inverse
at the origin, `H⁻¹(ζ) = ∑_{r ≥ 1} a_r ζ^r`, holomorphic on a neighborhood of `γD`, and
`∑_{r ≥ 1} |a_r| γ^r ≤ 1`. Then `K_G ≤ π/(2γ)`.
**Formalization Note** `(f, g)` is the published `GrothendieckConstant.KrivineScheme k` (oddness
almost everywhere) and `H` its `correlationFunction`: `(G₁, tG₁ + √(1 − t²)G₂)` is a pair of standard
Gaussian vectors with coordinatewise correlation `t`, whose density is the published
`gaussianPairDensity k t`. `G` is a local inverse branch (base `IsLocalInverseBranch`),
`a_r = taylorCoeff r G`. "Holomorphic on a neighborhood of `γD`" is `∑ |a_r| R^r < ∞` for some
`R > γ`. `γ > 0` is implicit in "`γD`" and is a hypothesis. -/
theorem theorem_A_1 {k : ℕ} (S : GrothendieckConstant.KrivineScheme k) (G : ℝ → ℝ)
(hG : IsLocalInverseBranch (GrothendieckConstant.correlationFunction S) G) (γ : ℝ)
(hγ : 0 < γ) (hhol : ∃ R : ℝ, γ < R ∧ Summable (fun r => |taylorCoeff r G| * R ^ r))
(hM : ∑' r : ℕ, |taylorCoeff (r + 1) G| * γ ^ (r + 1) ≤ 1) :
GrothendieckConstant.grothendieckConst ≤ Real.pi / (2 * γ) := by sorry
end Supplement
end GrothendieckLB
Source
Saha et al., New Lower and Upper Bounds for the Grothendieck Constant, arXiv:2608.11158 (held version, sha256 e7eb5d65…), Theorem A.1, p.47