AFTD Auto-Formalizing Theoretical Domains

CommunicationComplexity.integral_exp_sub_cgf_eq_one

theoremverifiedimported Lean

Exponential tilting by the cumulant generating function normalises to one. Let $\mu$ be a probability measure on a measurable space $\Omega$, let $X : \Omega \to \bbr$ be a random variable, and let $t \in \bbr$. Write $\Lambda(t) = \log \int_\Omega e^{t X(x)}\,d\mu(x)$ for the cumulant generating function of $X$ at $t$. If the function $x \mapsto e^{t X(x)}$ is $\mu$-integrable, then \[ \int_\Omega e^{\,t X(x) - \Lambda(t)}\,d\mu(x) = 1. \]

Statement

set_option relaxedAutoImplicit false in
set_option autoImplicit false in
open MeasureTheory in
open ProbabilityTheory in
open scoped ENNReal in
theorem CommunicationComplexity.integral_exp_sub_cgf_eq_one
    {Ω : Type*} [MeasurableSpace Ω] {μ : Measure Ω} [IsProbabilityMeasure μ]
    {X : Ω → ℝ} {t : ℝ} (h_exp_int : Integrable (fun x => Real.exp (t * X x)) μ) :
    ∫ x, Real.exp (t * X x - ProbabilityTheory.cgf X μ t) ∂μ = 1
Source
TCSlib, `CommunicationComplexity.integral_exp_sub_cgf_eq_one`
Lean proof
By Lucy Horowitz, Timothe Kasriel, Mihir Singhal, from github.com/Shilun-Allan-Li/tcslib at c8a02591b1e5 (Copyright (c) 2026 Lucy Horowitz, Timothe Kasriel, and Mihir Singhal. All rights reserved, Apache-2.0). Compiled here against leanprover/lean4:v4.33.1, Mathlib v4.33.1; written for leanprover/lean4:v4.25.0.
Verified
07 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 1
Used by 1

Lean source view module on GitHub

set_option relaxedAutoImplicit false inset_option autoImplicit false inopen MeasureTheory inopen ProbabilityTheory inopen scoped ENNReal in/-- The exponentially tilted density `exp (t X − cgf_X(t))` integrates to `1` whenever `exp (t X)` is integrable. -/theorem CommunicationComplexity.integral_exp_sub_cgf_eq_one    {Ω : Type*} [MeasurableSpace Ω] {μ : Measure Ω} [IsProbabilityMeasure μ]    {X : Ω → ℝ} {t : ℝ} (h_exp_int : Integrable (fun x => Real.exp (t * X x)) μ) :    ∫ x, Real.exp (t * X x - ProbabilityTheory.cgf X μ t) ∂μ = 1 := by  rw [exp_sub_cgf_eq, integral_const_mul]  change Real.exp (-ProbabilityTheory.cgf X μ t) * ProbabilityTheory.mgf X μ t = 1  rw [← ProbabilityTheory.exp_cgf h_exp_int]  rw [Real.exp_neg, inv_mul_cancel₀ (Real.exp_ne_zero _)]
See it in the graphDiscuss this resultChallenge it