Variance Bound From Partial Dual Estimates
ProvedFunctionalInequalities.KLS.LogConcave.variance_le_sum_of_partial_dual_bounds_modularfunctional-inequalitieskls-conjecturekls-modularletwin2026log-concave
This formalizes the variance inequality of Letwin2026 (Proposition 2.8, partial relation).
Let , let be an additive Haar measure on , and let be log-concave with
Assume:
- is and convex on all of .
- is and belongs to .
- Every coordinate derivative belongs to , where .
- For each and every smooth compactly supported ,
Then
The theorem is finite-dimensional and retains the explicit convexity, , and coordinate-derivative hypotheses. It is a variance assembly result from the separate compact-test dual bounds.
Superseded upload. This snapshot was replaced by FunctionalInequalities.KLS.LogConcave.variance_le_sum_of_partial_dual_bounds_readable: open the readable replacement. The replacement preserves the frozen source and mathematical statement, with a compact declaration context. This historical node remains available for existing imports.
Preamble
import Definitions.Def_KLSModular_FunctionalInequalities_KLS_LogConcave_HminusOne noncomputable section open MeasureTheory open scoped ContDiff section LT9DHminus1VarianceAssembly section namespace FunctionalInequalities end FunctionalInequalities open FunctionalInequalities section namespace FunctionalInequalities.KLS end FunctionalInequalities.KLS open FunctionalInequalities FunctionalInequalities.KLS section namespace FunctionalInequalities.KLS.LogConcave end FunctionalInequalities.KLS.LogConcave open FunctionalInequalities FunctionalInequalities.KLS FunctionalInequalities.KLS.LogConcave open ProbabilityTheory InnerProductSpace open Analysis.BakryEmery HeatSemigroup open scoped RealInnerProductSpace
Formal statement
theorem FunctionalInequalities.KLS.LogConcave.variance_le_sum_of_partial_dual_bounds_modular {n : ℕ}
(m : Measure (EuclideanSpace ℝ (Fin n))) [m.IsAddHaarMeasure]
{ν : ProbabilityMeasure (EuclideanSpace ℝ (Fin n))} (hν : FunctionalInequalities.KLS.LogConcave ν)
{W : EuclideanSpace ℝ (Fin n) → ℝ} (hνW : (ν : Measure _) = m.tilted (-W))
(hW : ContDiff ℝ 2 W) (hWc : ConvexOn ℝ Set.univ W)
{f : EuclideanSpace ℝ (Fin n) → ℝ} (hf : ContDiff ℝ 1 f) (hf2 : MemLp f 2 (ν : Measure _))
(hdf2 : ∀ i, MemLp (fun x => fderiv ℝ f x (EuclideanSpace.single i 1)) 2 (ν : Measure _))
(K : Fin n → ℝ)
(hK : ∀ i (h : EuclideanSpace ℝ (Fin n) → ℝ), ContDiff ℝ ∞ h → HasCompactSupport h →
(∫ x, fderiv ℝ f x (EuclideanSpace.single i 1) * h x ∂(ν : Measure _)) ^ 2 ≤
K i * ∫ x, ‖fderiv ℝ h x‖ ^ 2 ∂(ν : Measure _)) :
variance f (ν : Measure _) ≤ ∑ i, K i := by sorry
Source
Source: Letwin, The KLS constant is O(log^(1/4) n) (arXiv:2607.24164v1), Proposition 2.8 (Barthe and Klartag [7, Proposition 10]). Lean source: https://github.com/human-ai-mathematics/kls-lean/blob/408915f/formalization/FunctionalInequalities/KLS/LogConcave/HminusOne.lean#L450-L631 (repository human-ai-mathematics/kls-lean, commit 408915f, module `FunctionalInequalities.KLS.LogConcave.HminusOne`, declaration `FunctionalInequalities.KLS.LogConcave.variance_le_sum_of_partial_dual_bounds`; the repository is private at upload time).