Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3: Haar overlap concentration

Open
QAlgorithms.Papers.Independent.Intakenext100.P54.haar_overlap_concentration

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let D,K≥1D,K\ge1D,K≥1, let 0≤p≤D0\le p\le D0≤p≤D be an integer, let ϕ\phiϕ be a normalized pure state on CD\mathbb C^DCD, and let PPP be an orthogonal projector of rank ppp. Suppose U1,…,UKU_1,\ldots,U_KU1​,…,UK​ are independent measurable random unitaries, each having the normalized Haar law. For 0<ε≤10<\varepsilon\le10<ε≤1,

Pr⁡{∣1K∑k=1KTr⁡(UkϕUk∗P)−pD∣≥εpD}≤2 2−Kpε2/6.\Pr\left\{\left|\frac1K\sum_{k=1}^K\operatorname{Tr}(U_k\phi U_k^*P)-\frac pD\right|\ge\frac{\varepsilon p}{D}\right\}\le 2\,2^{-Kp\varepsilon^2/6}.Pr{​K1​k=1∑K​Tr(Uk​ϕUk∗​P)−Dp​​≥Dεp​}≤22−Kpε2/6.

Here a vector and its rank-one density operator are distinguished in the formalization. The result controls Haar overlap fluctuations; at rank zero the event has probability one and the upper bound is two.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P54
Formal statement
noncomputable section

open scoped BigOperators Matrix ComplexConjugate ComplexOrder NNReal ENNReal

open MeasureTheory ProbabilityTheory

namespace QAlgorithms.Papers.Independent.Intakenext100.P54

open QAlgorithms.Papers.Independent.Information100.P03 (PureEnsemble)

open QAlgorithms.Papers.Independent.Information100.P53 (fidelity)

open QAlgorithms.Papers.Independent.Dependency50.P07 (wordProb)

open QAlgorithms.Papers.Independent.Claimed50.P08 (classicalEntropy)

open QAlgorithms.Papers.Independent.Claimed50.P38 (tensorSuperop)

/-- Lemma 3, PDF pp.3,14–15: two-sided concentration of independent Haar overlaps.
Rank zero is included; its right-hand side is two. The paper's exp has base two. -/
theorem haar_overlap_concentration {Omega : Type*} [MeasurableSpace Omega]
    (mu : Measure Omega) [IsProbabilityMeasure mu] (D K p : ℕ)
    (hD : 0 < D) (hK : 0 < K) (hp : p ≤ D)
    (phi : RSPVector (Fin D)) (P : Matrix (Fin D) (Fin D) ℂ)
    (hHerm : P.IsHermitian) (hProj : P * P = P) (hRank : P.trace.re = p)
    (U : Fin K → Omega → Matrix.unitaryGroup (Fin D) ℂ)
    (hMeas : ∀ k, Measurable (U k)) (hInd : iIndepFun U mu)
    (hHaar : ∀ k, Measure.map (U k) mu = QAlgorithms.Schur.unitaryHaar D)
    (eps : ℝ) (heps : 0 < eps) (hepsOne : eps ≤ 1) :
    mu {w | |(∑ k, ((U k w : Matrix (Fin D) (Fin D) ℂ) * pureDensity phi.val *
      (U k w : Matrix (Fin D) (Fin D) ℂ)ᴴ * P).trace.re) /
      (K : ℝ) - (p : ℝ) / D| ≥ eps * p / D} ≤
      ENNReal.ofReal (2 * (2 : ℝ) ^ (-(K : ℝ) * p * eps ^ 2 / 6)) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P54

end
Source
Remote preparation of quantum states, arXiv:quant-ph/0307100v3, PDF p. 3, Lemma 3: Haar overlap concentration
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

Let Ω\OmegaΩ be any measurable space with a probability measure μ\muμ; let D,K,pD,K,pD,K,p be natural numbers with D>0D>0D>0, K>0K>0K>0, and p≤Dp\le Dp≤D; let ϕ∈CD\phi\in\mathbb C^Dϕ∈CD have Euclidean norm one; and let PPP be a complex D×DD\times DD×D matrix satisfying P†=PP^\dagger=PP†=P, P2=PP^2=PP2=P, and Re⁡tr⁡P=p\operatorname{Re}\operatorname{tr}P=pRetrP=p. For every family of measurable, jointly independent random unitary matrices Uk:Ω→U(D)U_k:\Omega\to U(D)Uk​:Ω→U(D), 0≤k<K0\le k<K0≤k<K, each having the normalized Haar distribution on U(D)U(D)U(D), and every real 0<ε≤10<\varepsilon\le10<ε≤1, the following probability bound holds:

μ ⁣{w:∣1K∑k=0K−1Re⁡tr⁡(Uk(w)ϕϕ†Uk(w)†P)−pD∣≥εpD}≤2 2−Kpε2/6.\mu\!\left\{ w: \left| \frac1K\sum_{k=0}^{K-1} \operatorname{Re}\operatorname{tr} \bigl(U_k(w)\phi\phi^\dagger U_k(w)^\dagger P\bigr) -\frac pD \right| \ge\frac{\varepsilon p}{D} \right\} \le 2\,2^{-Kp\varepsilon^2/6}.μ{w:​K1​k=0∑K−1​Retr(Uk​(w)ϕϕ†Uk​(w)†P)−Dp​​≥Dεp​}≤22−Kpε2/6.

The right side is embedded into the extended nonnegative reals used for measure values. The assumptions allow p=0p=0p=0; in that case the event is all of Ω\OmegaΩ, because its threshold is zero, and the displayed upper bound is 222.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 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