Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(19), proof of Theorem 2, p. 34 β€” 𝔼(Sβ€²)^Ξ“(A,β€’) βˆ’ 𝔼(Sβ€²)^Ξ“(L,β€’) = ∫ (Ξ± βˆ’ f_Y βˆ— w)(πŸ™_{Aβˆ–L} βˆ’ πŸ™_{Lβˆ–A}) h dz

Proved
MultivarScores.LevelSet.eq_19

by mikedeng1 Β· Oct 14, 2026 Β· Mathlib 0df444a (Lean v4.33.1)

consistent-scoring-functionfubinilevel-setsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1scoring-rules

Let PYP_YPY​ be a probability measure on Rd\mathbb R^dRd (the distribution of the observation YYY). Let w,h:Rdβ†’Rw,h:\mathbb R^d\to\mathbb Rw,h:Rdβ†’R be measurable with wβ‰₯0w\ge0wβ‰₯0, hβ‰₯0h\ge0hβ‰₯0, hhh integrable, and (s,z)↦w(zβˆ’s)h(z)(s,z)\mapsto w(z-s)h(z)(s,z)↦w(zβˆ’s)h(z) integrable with respect to PYβŠ—LdP_Y\otimes\mathcal L^dPYβ€‹βŠ—Ld. Let Ξ±>0\alpha>0Ξ±>0, let AβŠ†RdA\subseteq\mathbb R^dAβŠ†Rd be Borel, and write L=L(fYβˆ—w;Ξ±)={z:(fYβˆ—w)(z)β‰₯Ξ±}L=L(f_Y*w;\alpha)=\{z: (f_Y*w)(z)\ge\alpha\}L=L(fYβ€‹βˆ—w;Ξ±)={z:(fYβ€‹βˆ—w)(z)β‰₯Ξ±}. Then

EPY[(Sβ€²)Ξ“(A,βˆ™;w,h,Ξ±)]βˆ’EPY[(Sβ€²)Ξ“(L,βˆ™;w,h,Ξ±)]=∫Rd(Ξ±βˆ’(fYβˆ—w)(z))[1Aβˆ–L(z)βˆ’1Lβˆ–A(z)]h(z) dz.\mathbb E_{P_Y}\bigl[(S')^\Gamma(A,\bullet;w,h,\alpha)\bigr]-\mathbb E_{P_Y}\bigl[(S')^\Gamma(L,\bullet;w,h,\alpha)\bigr]=\int_{\mathbb R^d}\bigl(\alpha-(f_Y*w)(z)\bigr)\bigl[\mathbb 1_{A\setminus L}(z)-\mathbb 1_{L\setminus A}(z)\bigr]h(z)\,dz.EPY​​[(Sβ€²)Ξ“(A,βˆ™;w,h,Ξ±)]βˆ’EPY​​[(Sβ€²)Ξ“(L,βˆ™;w,h,Ξ±)]=∫Rd​(Ξ±βˆ’(fYβ€‹βˆ—w)(z))[1Aβˆ–L​(z)βˆ’1Lβˆ–A​(z)]h(z)dz.

This is display (19), Ξ”β€²=EPY[Ξ”β€²β€²]\Delta'=\mathbb E_{P_Y}[\Delta'']Ξ”β€²=EPY​​[Ξ”β€²β€²]. Since the integrand is nonnegative (Ξ±βˆ’fYβˆ—w>0\alpha-f_Y*w>0Ξ±βˆ’fYβ€‹βˆ—w>0 on Aβˆ–LA\setminus LAβˆ–L and ≀0\le0≀0 on Lβˆ–AL\setminus ALβˆ–A), it yields the consistency of (Sβ€²)Ξ“(S')^\Gamma(Sβ€²)Ξ“.

Formalization Note The paper's sets Ξ£+βˆ’,Ξ£βˆ’+\Sigma_{+-},\Sigma_{-+}Ξ£+βˆ’β€‹,Ξ£βˆ’+​ are built from {Ξ[Y]>Ξ±}\{\Xi[Y]>\alpha\}{Ξ[Y]>Ξ±} (a printed slip); here Ξ£+βˆ’=Aβˆ–L\Sigma_{+-}=A\setminus LΞ£+βˆ’β€‹=Aβˆ–L and Ξ£βˆ’+=Lβˆ–A\Sigma_{-+}=L\setminus AΞ£βˆ’+​=Lβˆ–A with L={fYβˆ—wβ‰₯Ξ±}L=\{f_Y*w\ge\alpha\}L={fYβ€‹βˆ—wβ‰₯Ξ±} as in (9), which changes nothing because the integrand vanishes on {fYβˆ—w=Ξ±}\{f_Y*w=\alpha\}{fYβ€‹βˆ—w=Ξ±}. The product-integrability hypothesis pins down Assumption 1 and Theorem 2's "(Sβ€²)Ξ“(A,βˆ™)fY(βˆ™)(S')^\Gamma(A,\bullet)f_Y(\bullet)(Sβ€²)Ξ“(A,βˆ™)fY​(βˆ™) is integrable": with w,hβ‰₯0w,h\ge0w,hβ‰₯0 it makes y↦(Sβ€²)Ξ“(B,y)y\mapsto (S')^\Gamma(B,y)y↦(Sβ€²)Ξ“(B,y) PYP_YPY​-integrable for every Borel BBB at once, including LLL. The paper allows www to be a local Borel measure (Assumption 1(a)); here www is a measurable function, a disclosed specialization that excludes only the Dirac kernel w=Ξ΄0w=\delta_0w=Ξ΄0​ (the density level set score (12)).

Preamble
import Mathlib
import Definitions.Def_MultivarScores_LevelSet_Setting
Formal statement
namespace MultivarScores.LevelSet

open MeasureTheory

theorem eq_19 {d : β„•} (w h : MultivarScores.L2Rule.E d β†’ ℝ) (PY : Measure (MultivarScores.L2Rule.E d)) [IsProbabilityMeasure PY]
    (hw_meas : Measurable w) (hw_nonneg : βˆ€ z, 0 ≀ w z)
    (hh_meas : Measurable h) (hh_nonneg : βˆ€ z, 0 ≀ h z) (hh_int : Integrable h)
    (hwh_int : Integrable (fun p : MultivarScores.L2Rule.E d Γ— MultivarScores.L2Rule.E d => w (p.2 - p.1) * h p.2) (PY.prod volume))
    (Ξ± : ℝ) (hΞ± : 0 < Ξ±) (A : Set (MultivarScores.L2Rule.E d)) (hA : MeasurableSet A) :
    (∫ y, scoreΞ“ w h Ξ± A y βˆ‚PY) - ∫ y, scoreΞ“ w h Ξ± (levelSet (MultivarScores.L2Rule.smooth PY w) Ξ±) y βˆ‚PY =
      ∫ z, (α - MultivarScores.L2Rule.smooth PY w z) *
        ((A \ levelSet (MultivarScores.L2Rule.smooth PY w) Ξ±).indicator 1 z
          - (levelSet (MultivarScores.L2Rule.smooth PY w) Ξ± \ A).indicator 1 z) * h z := by sorry

end MultivarScores.LevelSet
Source
Meng, Taylor, Ben Taieb, Li, Scores for Multivariate Distributions and Level Sets, arXiv:2002.09578v6, p. 34, proof of Theorem 2, display (19)
Human review
  • Endorsed by Shuze Chen Β· Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 Β· Oct 14, 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
Β© 2026 Prove2Me