Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Two-copy entropy inequality for a finite linear supremum process

Proved
TalagrandCore.two_copy_entropy_Z

by Harry_Xu · Aug 13, 2026 · Mathlib c5ea003 (Lean v4.30.0)

concentration-inequalitiesempirical-processesprobabilitytalagrand

For a centered finite Bernoulli linear supremum ZZZ with coefficient envelope one and variance proxy σ2\sigma^2σ2, define the random variance process Σ2\Sigma^2Σ2. For every λ≥0\lambda\ge0λ≥0,

λE[ZeλZ]−E[eλZ]log⁡E[eλZ]≤λ2e2λE ⁣[eλZ(Σ2+σ2)].\lambda\mathbb E[Ze^{\lambda Z}]-\mathbb E[e^{\lambda Z}]\log\mathbb E[e^{\lambda Z}] \le \lambda^2e^{2\lambda}\mathbb E\!\left[e^{\lambda Z}(\Sigma^2+\sigma^2)\right].λE[ZeλZ]−E[eλZ]logE[eλZ]≤λ2e2λE[eλZ(Σ2+σ2)].

This is the two-copy entropy estimate feeding the Gaussian upper-tail differential inequality.

Formalization Note Suprema are finite maxima, so measurable-selection issues are replaced by a finite argmax construction inside the proof.

Preamble
import Definitions.Def_talagrand_finite_bool_core
open MeasureTheory
open scoped Classical BigOperators
Formal statement
namespace TalagrandCore

variable {κ ι : Type} [DecidableEq κ] [Fintype κ]
variable [Fintype ι] [Nonempty ι]

theorem two_copy_entropy_Z (p : NNReal) (hp : p ≤ 1)
    (coeff : ι → κ → ℝ) (hB : ∀ a x, |coeff a x| ≤ 1)
    (sigmaSq : ℝ)
    (hVar : ∀ a : ι, ∑ x : κ, (p : ℝ) * (1 - (p : ℝ)) * coeff a x ^ 2 ≤ sigmaSq)
    (lam : ℝ) (hlam : 0 ≤ lam) :
    lam * Ex (p : ℝ) (fun ω => Zproc coeff (p : ℝ) ω *
        Real.exp (lam * Zproc coeff (p : ℝ) ω)) -
      Ex (p : ℝ) (fun ω => Real.exp (lam * Zproc coeff (p : ℝ) ω)) *
        Real.log (Ex (p : ℝ) (fun ω => Real.exp (lam * Zproc coeff (p : ℝ) ω))) ≤
      lam ^ 2 * Real.exp (2 * lam) *
        Ex (p : ℝ) (fun ω => Real.exp (lam * Zproc coeff (p : ℝ) ω) *
          (Sigma2 coeff (p : ℝ) ω + sigmaSq)) := by sorry

end TalagrandCore
Source
Michel Ledoux, On Talagrand’s deviation inequalities for product measures, ESAIM Probability and Statistics 1 (1996), Theorems 2.4–2.5, pp. 63–87. Formal Lean proof extracted from Prove2Me accepted submission bf106d23-ff42-48f5-a837-a8528101c849 by tianyipeng.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me