Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 3.15 — without (strong) negative type, distance covariance fails to characterize independence

Proved
DistCov.Converse.proposition_3_15

by mikedeng1 · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

distance-covarianceindependencenegative-typep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1strong-negative-type

Let X\mathcal XX and Y\mathcal YY be separable metric spaces, and suppose Y\mathcal YY has at least two points.

  1. If X\mathcal XX is not of negative type, there is a Borel probability measure θ\thetaθ on X×Y\mathcal X\times\mathcal YX×Y whose marginals have finite first moments and such that
dcov⁡(θ)<0.\operatorname{dcov}(\theta)<0.dcov(θ)<0.
  1. If X\mathcal XX is not of strong negative type, there is a Borel probability measure θ\thetaθ on X×Y\mathcal X\times\mathcal YX×Y whose marginals have finite first moments and such that
dcov⁡(θ)=0,\operatorname{dcov}(\theta)=0,dcov(θ)=0,

yet θ\thetaθ is not a product measure.

Theorem 3.11 of the paper says that on spaces of strong negative type, dcov⁡(θ)=0\operatorname{dcov}(\theta)=0dcov(θ)=0 forces θ\thetaθ to be the product of its marginals, so that distance covariance characterizes independence. This proposition shows that the hypothesis on X\mathcal XX cannot be dropped: without negative type distance covariance can be negative, and without strong negative type it can vanish on a dependent law.

Formalization Note. Separability of X\mathcal XX and Y\mathcal YY is the standing assumption of the author's Errata (i); it makes ddd measurable for the product Borel σ\sigmaσ-field. "For every metric space Y\mathcal YY with at least two points" is rendered by taking Y\mathcal YY as a universally quantified parameter, with the hypothesis that it has two distinct points; θ\thetaθ is chosen after Y\mathcal YY. "Not a product measure" means θ\thetaθ differs from the product of its own marginals. Part 2's hypothesis is the bare negation of strong negative type and so covers spaces of negative type as well as spaces that fail it.

Preamble
import Mathlib
import Definitions.Def_DistCov_Converse_Setting

open MeasureTheory
Formal statement
namespace DistCov.Converse

theorem proposition_3_15
    {X : Type*} [MetricSpace X] [MeasurableSpace X] [BorelSpace X] [SecondCountableTopology X]
    {Y : Type*} [MetricSpace Y] [MeasurableSpace Y] [BorelSpace Y] [SecondCountableTopology Y]
    (hY : ∃ y₁ y₂ : Y, y₁ ≠ y₂) :
    (¬ DistCov.Indep.NegType X → ∃ θ : Measure (X × Y), IsProbabilityMeasure θ ∧
        DistCov.Indep.FiniteFirstMoment (θ.map Prod.fst) ∧ DistCov.Indep.FiniteFirstMoment (θ.map Prod.snd) ∧ DistCov.Indep.dcov θ < 0) ∧
    (¬ DistCov.Indep.StrongNegType X → ∃ θ : Measure (X × Y), IsProbabilityMeasure θ ∧
        DistCov.Indep.FiniteFirstMoment (θ.map Prod.fst) ∧ DistCov.Indep.FiniteFirstMoment (θ.map Prod.snd) ∧
        DistCov.Indep.dcov θ = 0 ∧ ¬ DistCov.Indep.IsProduct θ) := by sorry

end DistCov.Converse
Source
Lyons, Distance covariance in metric spaces, arXiv:1106.5758 (version of 19 Dec. 2020), p. 17, Proposition 3.15; Errata (i), p. 24
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