Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2 (p. 13), (20) (p. 34) — (S′)^Γ is strictly consistent when 𝓛^d{f_Y ∗ w = α} = 0 and h ≠ 0 a.e.

Proved
MultivarScores.LevelSet.strictly_consistent

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

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

Let PYP_YPY​, www, hhh, α>0\alpha>0α>0 and the Borel set AAA be as in Theorem 2: w,hw,hw,h measurable and nonnegative, hhh integrable, and (s,z)↦w(z−s)h(z)(s,z)\mapsto w(z-s)h(z)(s,z)↦w(z−s)h(z) integrable for PY⊗LdP_Y\otimes\mathcal L^dPY​⊗Ld. Write L=L(fY∗w;α)L=L(f_Y*w;\alpha)L=L(fY​∗w;α). Assume moreover

  1. h(z)≠0h(z)\neq0h(z)=0 for Ld\mathcal L^dLd-almost every zzz;
  2. the set {z:(fY∗w)(z)=α}\{z : (f_Y*w)(z)=\alpha\}{z:(fY​∗w)(z)=α} is Ld\mathcal L^dLd-null;
  3. AAA and LLL differ by a set that is not Ld\mathcal L^dLd-null, i.e. Ld(A △ L)≠0\mathcal L^d(A\,\triangle\,L)\neq0Ld(A△L)=0.

Then

EPY[(S′)Γ(L,∙;w,h,α)]<EPY[(S′)Γ(A,∙;w,h,α)].\mathbb E_{P_Y}\bigl[(S')^\Gamma(L,\bullet;w,h,\alpha)\bigr]<\mathbb E_{P_Y}\bigl[(S')^\Gamma(A,\bullet;w,h,\alpha)\bigr].EPY​​[(S′)Γ(L,∙;w,h,α)]<EPY​​[(S′)Γ(A,∙;w,h,α)].

So L(fY∗w;α)L(f_Y*w;\alpha)L(fY​∗w;α) is the unique minimizer of the expected level set score, up to Ld\mathcal L^dLd-null sets: the strict consistency clause of Theorem 2.

Formalization Note The paper states strict consistency under the hypothesis "L(fY∗w;α)L(f_Y*w;\alpha)L(fY​∗w;α) equals the closure of {fY∗w>α}\{f_Y*w>\alpha\}{fY​∗w>α}", and its proof (p. 34) uses only the consequence that {fY∗w=α}\{f_Y*w=\alpha\}{fY​∗w=α} is Ld\mathcal L^dLd-null; that implication is false (for d=1d=1d=1, PY=δ0P_Y=\delta_0PY​=δ0​, w=α+dist(⋅,C)w=\alpha+\mathrm{dist}(\cdot,C)w=α+dist(⋅,C) with C⊂[0,1]C\subset[0,1]C⊂[0,1] a fat Cantor set and hhh a Gaussian density, L=RL=\mathbb RL=R is the closure of {w>α}\{w>\alpha\}{w>α} but {w=α}=C\{w=\alpha\}=C{w=α}=C has positive measure and A=R∖CA=\mathbb R\setminus CA=R∖C scores the same as LLL), so the hypothesis the proof actually uses, item 2, replaces the printed closure condition. The proof's "h>0h>0h>0 by Assumption 1(c)" is Theorem 2's own hypothesis "hhh is nonzero Ld\mathcal L^dLd-a.e." (item 1). "Only minimizer" is read modulo null sets, as in the proof. 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=δ0w=\delta_0w=δ0​ (the density level set score (12)).

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

open MeasureTheory

theorem strictly_consistent {d : ℕ} (w h : MultivarScores.L2Rule.E d → ℝ) (PY : Measure (MultivarScores.L2Rule.E d))
    [IsProbabilityMeasure PY]
    (hw_meas : Measurable w) (hw_nonneg : ∀ z, 0 ≤ w z)
    (hh_meas : Measurable h) (hh_nonneg : ∀ z, 0 ≤ h z) (hh_int : Integrable h)
    (hwh_int : Integrable (fun p : MultivarScores.L2Rule.E d × MultivarScores.L2Rule.E d => w (p.2 - p.1) * h p.2) (PY.prod volume))
    (α : ℝ) (hα : 0 < α) (A : Set (MultivarScores.L2Rule.E d)) (hA : MeasurableSet A)
    (hh_ne : ∀ᵐ z, h z ≠ 0)
    (hnull : volume {z | MultivarScores.L2Rule.smooth PY w z = α} = 0)
    (hdiff : volume (symmDiff A (levelSet (MultivarScores.L2Rule.smooth PY w) α)) ≠ 0) :
    ∫ y, scoreΓ w h α (levelSet (MultivarScores.L2Rule.smooth PY w) α) y ∂PY < ∫ y, scoreΓ w h α A y ∂PY := by sorry

end MultivarScores.LevelSet
Source
Meng, Taylor, Ben Taieb, Li, Scores for Multivariate Distributions and Level Sets, arXiv:2002.09578v6, p. 13, Theorem 2 (strict consistency sentence); p. 34, display (20) and the paragraph after it
Human review
  • Endorsed by Shuze Chen · Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · 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
© 2026 Prove2Me