Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

App. A, Notation, p. 13 — ∫f dP=⟨f,μP⟩H\int f\,dP=\langle f,\mu_P\rangle_{\mathcal H}∫fdP=⟨f,μP​⟩H​

Proved
KernelDRO.Duality.reproducing_identity

by mikedeng1 · Oct 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

kernel-mean-embeddingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1rkhs

Let X\mathcal XX be a compact metric space with its Borel σ\sigmaσ-algebra, H\mathcal HH a real Hilbert space and φ:X→H\varphi:\mathcal X\to\mathcal Hφ:X→H a continuous feature map. For every f∈Hf\in\mathcal Hf∈H and every Borel probability measure PPP on X\mathcal XX,

∫Xf(ξ) dP(ξ)=⟨f,μP⟩H,f(ξ)=⟨f,φ(ξ)⟩H,μP=∫φ dP.\int_{\mathcal X} f(\xi)\,dP(\xi)=\langle f,\mu_P\rangle_{\mathcal H},\qquad f(\xi)=\langle f,\varphi(\xi)\rangle_{\mathcal H},\quad \mu_P=\int\varphi\,dP .∫X​f(ξ)dP(ξ)=⟨f,μP​⟩H​,f(ξ)=⟨f,φ(ξ)⟩H​,μP​=∫φdP.

The expectation of any RKHS function is thus the inner product with the mean embedding; the paper uses this identity throughout its proofs, in particular to rewrite the Lagrangian (13).

Formalization Note Both integrals are Bochner integrals; continuity of φ\varphiφ on the compact X\mathcal XX makes them genuine.

Preamble
import Mathlib
import Definitions.Def_WassersteinDRO_Duality_erealExpectation
import Definitions.Def_KernelDRO_Duality_Setting

open MeasureTheory
Formal statement
namespace KernelDRO.Duality

open WassersteinDRO.Duality

/-- **Reproducing identity** `∫ f dP = ⟨f, µ_P⟩_ℋ` for `f ∈ ℋ` and `P ∈ 𝒫` (Zhu, Jitkrittum, Diehl &
Schölkopf, Kernel Distributionally Robust Optimization, arXiv:2006.06981v3, App. A, Notation, p. 13):
integrating the RKHS function `ξ ↦ f(ξ) = ⟪f, φ ξ⟫` against a Borel probability measure `P` on the
compact space `𝒳` gives the inner product of `f` with the mean embedding `µ_P = ∫ φ dP`.

**Formalization Note.** `𝒳` is a compact metric space with its Borel σ-algebra (a compact
`𝒳 ⊂ ℝᵈ` is an instance) and the feature map `φ` is continuous (equivalently, the kernel
`k(x, y) = ⟪φ x, φ y⟫` is continuous), so both integrals are genuine Bochner integrals. -/
theorem reproducing_identity {X : Type*} [MetricSpace X] [CompactSpace X] [MeasurableSpace X]
    [BorelSpace X] {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
    (φ : X → H) (hφ : Continuous φ) (f : H) (P : ProbabilityMeasure X) :
    ∫ ξ, rkhsEval φ f ξ ∂(P : Measure X) = inner ℝ f (meanEmbedding φ P) := by sorry

end KernelDRO.Duality
Source
Zhu, Jitkrittum, Diehl & Schölkopf, Kernel Distributionally Robust Optimization, arXiv:2006.06981v3, App. A, Notation, p. 13
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 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