Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition II.1: finite-block decorrelation converse

Open
QAlgorithms.Papers.Independent.Intakenext100.P47.decorrelation_converse

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

prove2quantum

Let ρAB\rho_{AB}ρAB​ be a normalized bipartite density matrix on finite nonzero spaces of dimensions dA,dBd_A,d_BdA​,dB​, put d=dAdBd=d_Ad_Bd=dA​dB​, and let n≥1n\ge1n≥1. Suppose a coordinated local product-unitary ensemble RRR, with probabilities pip_ipi​ and N≥1N\ge1N≥1 terms, ϵ\epsilonϵ-decorrelates ρ⊗n\rho^{\otimes n}ρ⊗n for ϵ≥0\epsilon\ge0ϵ≥0: its output is within full trace norm ϵ\epsilonϵ of some normalized product state. Define η(x)=−xlog⁡2x\eta(x)=-x\log_2xη(x)=−xlog2​x for 0≤x≤1/e0\le x\le1/e0≤x≤1/e, with η(0)=0\eta(0)=0η(0)=0, and η(x)=e−1log⁡2e\eta(x)=e^{-1}\log_2eη(x)=e−1log2​e for x≥1/ex\ge1/ex≥1/e. Then

log⁡2N≥H(p)≥Se(R,ρ⊗n)≥n(I(A:B)ρ−3ϵlog⁡2d−η(3ϵ)).\log_2N\ge H(p)\ge S_e(R,\rho^{\otimes n})\ge n\bigl(I(A:B)_\rho-3\epsilon\log_2d-\eta(3\epsilon)\bigr).log2​N≥H(p)≥Se​(R,ρ⊗n)≥n(I(A:B)ρ​−3ϵlog2​d−η(3ϵ)).

Here e=exp⁡(1)e=\exp(1)e=exp(1) is the natural exponential constant. The factor nnn multiplies the entire parenthesis. This is the finite-block converse for total correlation erasure, including large error tolerances.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P47
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P47

open QAlgorithms Filter

open QAlgorithms.Papers.Independent.Gap48.P20 (Copies localProduct)

open QAlgorithms.Papers.Independent.Dependency50.P19 (mutualInformation first)

open QAlgorithms.Papers.Independent.Dependency50.P45 (entropyExchange)

open QAlgorithms.Papers.Independent.Information100.P27 (shannon Config)

open QAlgorithms.Papers.Independent.Information100.P71 (channelPower)

open QAlgorithms.Papers.Independent.Advance50.P05 (RectangularInstrument)

open scoped Matrix ComplexConjugate

attribute [local instance] Classical.propDecidable Classical.decEq

/-- PDF p.3, Proposition II.1: the outer n multiplies erasureEta(3 eps) as printed. -/
theorem decorrelation_converse {A B : Type} [Fintype A] [Fintype B]
    [Nonempty A] [Nonempty B] (rho : Matrix (A × B) (A × B) ℂ)
    (hrho : IsDensityMatrix rho) (n : ℕ) (hn : 0 < n) (eps : ℝ) (heps : 0 ≤ eps)
    (R : COLUR (Copies A n) (Copies B n))
    (hR : Decorrelates (R.operation (sourcePower rho n)) eps) :
    let lower := (n : ℝ) * (mutualInformation rho - 3 * eps * logDimension A B - erasureEta (3*eps))
    lower ≤ entropyExchange R.operation (sourcePower rho n) ∧
    lower ≤ shannon R.weight ∧ lower ≤ Real.logb 2 R.count := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P47

end
Source
On the quantum, classical and total amount of correlations in a quantum state, arXiv:quant-ph/0410091v2, PDF p. 3, Proposition II.1: finite-block decorrelation converse
Read-back

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

For any finite nonempty types A,BA,BA,B, any positive semidefinite, trace-one matrix ρ\rhoρ on A×BA\times BA×B, any positive natural number nnn, any real ε≥0\varepsilon\ge0ε≥0, and any positive-length list of ccc pairs of local unitaries on An,BnA^n,B^nAn,Bn, with nonnegative weights pip_ipi​ summing to 111, define T(X)=∑i<cpi(Ui⊗Vi)X(Ui⊗Vi)†T(X)=\sum_{i<c}p_i(U_i\otimes V_i)X(U_i\otimes V_i)^\daggerT(X)=∑i<c​pi​(Ui​⊗Vi​)X(Ui​⊗Vi​)† and (Xn)(a,b),(a′,b′)=∏t<nρ(at,bt),(at′,bt′)(X_n)_{(a,b),(a',b')}=\prod_{t<n}\rho_{(a_t,b_t),(a'_t,b'_t)}(Xn​)(a,b),(a′,b′)​=∏t<n​ρ(at​,bt​),(at′​,bt′​)​. Assume that there exist positive semidefinite, trace-one matrices a,ba,ba,b such that ∥T(Xn)−a⊗b∥1≤ε\|T(X_n)-a\otimes b\|_1\le\varepsilon∥T(Xn​)−a⊗b∥1​≤ε, where ∥M∥1=Re⁡tr⁡M†M\|M\|_1=\operatorname{Re}\operatorname{tr}\sqrt{M^\dagger M}∥M∥1​=RetrM†M​. Put H(Y)=−Re⁡tr⁡[ x↦xlog⁡2x ](Y)H(Y)=-\operatorname{Re}\operatorname{tr}[\,x\mapsto x\log_2x\,](Y)H(Y)=−Retr[x↦xlog2​x](Y), I(ρ)=H(ρA)+H(ρB)−H(ρ)I(\rho)=H(\rho_A)+H(\rho_B)-H(\rho)I(ρ)=H(ρA​)+H(ρB​)−H(ρ), where (ρA)a,a′=∑bρ(a,b),(a′,b)(\rho_A)_{a,a'}=\sum_b\rho_{(a,b),(a',b)}(ρA​)a,a′​=∑b​ρ(a,b),(a′,b)​ and (ρB)b,b′=∑aρ(a,b),(a,b′)(\rho_B)_{b,b'}=\sum_a\rho_{(a,b),(a,b')}(ρB​)b,b′​=∑a​ρ(a,b),(a,b′)​, and define η(x)=−xlog⁡2x\eta(x)=-x\log_2xη(x)=−xlog2​x for x≤e−1x\le e^{-1}x≤e−1 and η(x)=e−1log⁡2e\eta(x)=e^{-1}\log_2eη(x)=e−1log2​e otherwise. Then each of

H ⁣((id⊗T)(∣ξXn⟩⟨ξXn∣)),−∑i<cpilog⁡2pi,log⁡2cH\!\left((\mathrm{id}\otimes T)(|\xi_{X_n}\rangle\langle\xi_{X_n}|)\right), \qquad -\sum_{i<c}p_i\log_2p_i, \qquad \log_2cH((id⊗T)(∣ξXn​​⟩⟨ξXn​​∣)),−i<c∑​pi​log2​pi​,log2​c

is at least n(I(ρ)−3εlog⁡2(∣A∣∣B∣)−η(3ε))n\bigl(I(\rho)-3\varepsilon\log_2(|A||B|)-\eta(3\varepsilon)\bigr)n(I(ρ)−3εlog2​(∣A∣∣B∣)−η(3ε)), where ξX(u,v)=(X)v,u\xi_X(u,v)=(\sqrt X)_{v,u}ξX​(u,v)=(X​)v,u​ and matrix functions use continuous functional calculus. The tolerance may be 000 or arbitrarily large, the displayed lower bound need not be positive, and zero weights are allowed.

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