Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem C.1 — Independent-row comparison

Open
GinibrePermanent977.Anticoncentration.Supplement.theorem_C_1

by mikedeng1 · Nov 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

anticoncentrationp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1probabilityrandom-matrix-theory

Let K∈{R,C,H}\mathbb K\in\{\mathbb R,\mathbb C,\mathbb H\}K∈{R,C,H}, β=dim⁡RK\beta=\dim_{\mathbb R}\mathbb Kβ=dimR​K, and n≥1n\ge1n≥1. Let AAA have mutually independent rows RiR_iRi​, with arbitrary dependence among the coordinates of each row. Suppose σi>0\sigma_i>0σi​>0 and

(36)ERi∗Ri=σi2In,γ1,β≤Lt∣Ric∣2σi2∥c∥22(c∈Kn∖{0}).(36)\qquad ER_i^*R_i=\sigma_i^2I_n,\qquad \gamma_{1,\beta}\le_{\mathrm{Lt}}\frac{|R_ic|^2}{\sigma_i^2\|c\|_2^2}\quad(c\in\mathbb K^n\setminus\{0\}).(36)ERi∗​Ri​=σi2​In​,γ1,β​≤Lt​σi2​∥c∥22​∣Ri​c∣2​(c∈Kn∖{0}).

Here Ric=∑jRi,jcjR_ic=\sum_jR_{i,j}c_jRi​c=∑j​Ri,j​cj​ uses right coefficients without conjugating the row, and γ1,β\gamma_{1,\beta}γ1,β​ has gamma shape β/2\beta/2β/2 and scale 2/β2/\beta2/β. Put Xn=∣per⁡KG∣2/n!X_n=|\operatorname{per}_{\mathbb K}G|^2/n!Xn​=∣perK​G∣2/n!, and let Δn\Delta_nΔn​ be the normalized squared determinant of GGG over the real and complex fields or normalized squared Study determinant over quaternions. Then both links hold:

(37)Δn≤LtXn≤Lt∣per⁡KA∣2n!∏iσi2.(37)\qquad\Delta_n\le_{\mathrm{Lt}}X_n\le_{\mathrm{Lt}}\frac{|\operatorname{per}_{\mathbb K}A|^2}{n!\prod_i\sigma_i^2}.(37)Δn​≤Lt​Xn​≤Lt​n!∏i​σi2​∣perK​A∣2​.

The exact finite second moment is

E∣per⁡KA∣2=n!∏iσi2.E|\operatorname{per}_{\mathbb K}A|^2=n!\prod_i\sigma_i^2.E∣perK​A∣2=n!i∏​σi2​.

For every ε>0\varepsilon>0ε>0,

(38)P(∣per⁡KA∣≤εn!∏iσi)≤min⁡{1,(2πeβ)β/2κn,βεβ}.(38)\qquad P\left(|\operatorname{per}_{\mathbb K}A|\le\varepsilon\sqrt{n!}\prod_i\sigma_i\right)\le\min\left\{1,\left(\frac{2\pi e}{\beta}\right)^{\beta/2}\kappa_{n,\beta}\varepsilon^\beta\right\}.(38)P(∣perK​A∣≤εn!​i∏​σi​)≤min{1,(β2πe​)β/2κn,β​εβ}.

This is a comparison at the exact second-moment scale for non-Gaussian independent rows, with no coordinate independence or common-law requirement.

Formalization Note Covariance entry (j,l)(j,l)(j,l) is E(Ri,j‾Ri,l)E(\overline{R_{i,j}}R_{i,l})E(Ri,j​​Ri,l​); its integrability is part of the finite covariance equality. The projection denominator is used only for nonzero ccc.

Differences from the source:

  • encoding: K ∈ {ℝ,ℂ,ℍ} and β = dim_ℝ K → published Scalar carriers and beta; reason: the three concrete fields and real-coordinate normalizations are identical
  • encoding: independent random inputs → coordinate functions under their probability product laws; reason: these are exactly the joint laws of independent Borel random variables; no moment or density restriction is added
  • encoding: Laplace comparison for s ≥ 0 → published LaplaceOrder for s > 0; reason: at s = 0 both bounded integrands equal 1 under probability laws
  • encoding: the finite covariance matrix ER_i*R_i = σ_i² I → entrywise Integrable conjugate products and their Bochner integrals equal σ_i² δ_jl; reason: finite diagonal second moments imply all cross products are integrable by Cauchy–Schwarz; this excludes the Bochner default value without adding a moment assumption
  • generalization: the labelled Theorem C.1 gives (37) and (38) → also Integrable |per A|² and its exact expectation n! ∏ σ_i²; reason: the paragraph immediately following (38), p.19, states this identity; it is expressly required by the supplement brief and dropping it leaves the full labelled theorem
Preamble
import Mathlib
import Definitions.Def_GinibrePermanent977_Anticoncentration_Supplement_Model
open MeasureTheory ProbabilityTheory
open scoped BigOperators Nat ENNReal
Formal statement
namespace GinibrePermanent977.Anticoncentration.Supplement

/-- Theorem C.1 (36)–(38), p.19, including the exact finite second moment stated after (38).
Each μ_i is a row law, with arbitrary coordinate dependence; their product is the matrix law. -/
theorem theorem_C_1 (F : Scalar) (n : ℕ) (hn : 1 ≤ n)
    (μ : Fin n → Measure (Fin n → F.K)) [∀ i, IsProbabilityMeasure (μ i)]
    (σ : Fin n → ℝ) (hσ : ∀ i, 0 < σ i)
    (hiso : ∀ i, RowIsotropic (μ i) (σ i))
    (hproj : ∀ i (c : Fin n → F.K), c ≠ 0 →
      StochasticOrders.LaplaceTransform.LaplaceOrder (gammaLaw F 1) (μ i)
        id (fun r => ‖rowAction r c‖ ^ 2 / (σ i ^ 2 * norm2 c ^ 2))) :
    StochasticOrders.LaplaceTransform.LaplaceOrder (ginibre F n n) (ginibre F n n)
      (deltaPower F) (fun G => ‖normPer G‖ ^ 2) ∧
    StochasticOrders.LaplaceTransform.LaplaceOrder (ginibre F n n) (Measure.pi μ)
      (fun G => ‖normPer G‖ ^ 2)
      (fun A => ‖perK A‖ ^ 2 / ((n ! : ℝ) * ∏ i, σ i ^ 2)) ∧
    Integrable (fun A => ‖perK A‖ ^ 2) (Measure.pi μ) ∧
    (∫ A, ‖perK A‖ ^ 2 ∂Measure.pi μ) = (n ! : ℝ) * ∏ i, σ i ^ 2 ∧
    ∀ ε : ℝ, 0 < ε →
      (Measure.pi μ) {A | ‖perK A‖ ≤ ε * Real.sqrt (n ! : ℝ) * ∏ i, σ i} ≤
        ENNReal.ofReal (min 1 (((2 * Real.pi * Real.exp 1) / (F.beta : ℝ)) ^
          ((F.beta : ℝ) / 2) * kappa F n * ε ^ F.beta)) := by sorry

end GinibrePermanent977.Anticoncentration.Supplement
Source
Frederic Koehler, Pui Kuen Leung, STOC 2027 submission #977 (October 2026), Theorem C.1, p.19 (PDF SHA256 e25737f1d5e3492905f8eeaa41d92c55d95f73622a3b5846cc87aac6a346d4ec)

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