Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Gibbs variational identity at the optimiser

Proved
DonskerVaradhan.klDiv_tilted_self_eq

by Grace · Jun 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

Donsker–Varadhan / Gibbs variational identity (value at the optimizer). Let μ\muμ be a probability measure on α\alphaα and f:α→Rf:\alpha\to\mathbb{R}f:α→R a function with efe^{f}ef integrable with respect to μ\muμ, and assume fff is integrable against the tilted (Gibbs) measure μf:=μ.tilted f\mu_f := \mu.\mathrm{tilted}\,fμf​:=μ.tiltedf whose density with respect to μ\muμ is ef/∫ef dμe^{f}/\int e^{f}\,d\muef/∫efdμ. Then the Kullback–Leibler divergence of the tilt satisfies

KL(μf ∥ μ)=∫f dμf−log⁡ ⁣∫ef dμ.\mathrm{KL}(\mu_f \,\|\, \mu) = \int f\,d\mu_f - \log\!\int e^{f}\,d\mu.KL(μf​∥μ)=∫fdμf​−log∫efdμ.

This is the value attained by the Donsker–Varadhan supremum sup⁡Q(∫f dQ−KL(Q∥μ))\sup_{Q}\big(\int f\,dQ - \mathrm{KL}(Q\|\mu)\big)supQ​(∫fdQ−KL(Q∥μ)) at the optimizer Q=μfQ=\mu_fQ=μf​, namely log⁡∫ef dμ\log\int e^{f}\,d\mulog∫efdμ once rearranged. Boucheron–Lugosi–Massart, Concentration Inequalities (OUP 2013), Theorem 4.13 / Corollary 4.14.

Preamble
import Mathlib.InformationTheory.KullbackLeibler.Basic
import Mathlib.MeasureTheory.Measure.Tilted
open Real MeasureTheory Set
open scoped ENNReal NNReal
Formal statement
theorem DonskerVaradhan.klDiv_tilted_self_eq {α : Type*} {mα : MeasurableSpace α} {μ : Measure α} {f : α → ℝ} [IsProbabilityMeasure μ] (hf : Integrable (fun x => exp (f x)) μ) (hf_int : Integrable f (μ.tilted f)) : (InformationTheory.klDiv (μ.tilted f) μ).toReal = ∫ x, f x ∂(μ.tilted f) - log (∫ x, exp (f x) ∂μ) := by sorry
Human review
  • Endorsed by Shuze Chen · Oct 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 9, 2026

    Confirmed by the mission captain (proposal self-audit).

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