Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3 (§VIII.B): smooth min-relative entropy gives spectral inf-divergence

Proved
QAlgorithms.Papers.Independent.Support50.P21.spectralInf_smoothMin

by Goku · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

All systems are nonzero finite-dimensional complex Hilbert spaces, and logarithms have base two.

Fix a nonzero finite-dimensional system HHH. Let ρn\rho_nρn​ be density operators and σn≥0\sigma_n\ge0σn​≥0 on H⊗nH^{\otimes n}H⊗n for n≥1n\ge1n≥1. Define Πn(γ)=ρn−2nγσn\Pi_n(\gamma)=\rho_n-2^{n\gamma}\sigma_nΠn​(γ)=ρn​−2nγσn​ and

D‾(ρ^∥σ^)=sup⁡{γ∈R:lim inf⁡n→∞Tr⁡[{Πn(γ)≥0}Πn(γ)]=1}.\underline D(\hat\rho\Vert\hat\sigma)=\sup\{\gamma\in\mathbb R:\liminf_{n\to\infty}\operatorname{Tr}[\{\Pi_n(\gamma)\ge0\}\Pi_n(\gamma)]=1\}.D​(ρ^​∥σ^)=sup{γ∈R:n→∞liminf​Tr[{Πn​(γ)≥0}Πn​(γ)]=1}.

Then, in the extended real topology,

D‾(ρ^∥σ^)=lim⁡ε↓0lim inf⁡n→∞1nDmin⁡ε(ρn∥σn).\underline D(\hat\rho\Vert\hat\sigma)=\lim_{\varepsilon\downarrow0}\liminf_{n\to\infty}\frac1nD_{\min}^{\varepsilon}(\rho_n\Vert\sigma_n).D​(ρ^​∥σ^)=ε↓0lim​n→∞liminf​n1​Dminε​(ρn​∥σn​).

The order of limits is as displayed, with no iid, support, trace-bound-on-σn\sigma_nσn​, or finite-rate assumptions. This identifies the spectral inf-divergence rate with the smoothed asymptotic quantity.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P21
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Support50.P21

open QAlgorithms QAlgorithms.Entanglement Filter

open scoped Matrix ComplexConjugate ComplexOrder

attribute [local instance] Classical.propDecidable Classical.decEq

-- Use the accepted declarations directly; no paper-local aliases hide dimension binders.

open QAlgorithms.Papers.Independent.Claimed50.P33
  (FiniteState IsAverageLocalMonotone IsAdaptiveLOCCMonotone)

open QAlgorithms.Papers.Independent.Information100.P63 (Mat)

open QAlgorithms.Papers.Independent.Dependency50.P35 (tr elog maxRelative)

open QAlgorithms.Papers.Independent.Information100.P67 (IsSubnormalized)

open QAlgorithms.Papers.Independent.Information100.P07 (support)

open QAlgorithms.Papers.Independent.Information100.P71 (supportPower SupportIncluded)

variable {I J : Type} [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]

section Statements

variable [Nonempty I] [Nonempty J]

/-- Theorem 3, PDF pp.9–10, (67): first n→infinity, then eps→0+.
Proof (78) requires arbitrarily small slack at liminf, not a new hypothesis. -/
theorem spectralInf_smoothMin (rho sigma : (k : ℕ) → Mat (PositiveTensorBasis I k))
    (hrho : ∀ k, IsDensityMatrix (rho k)) (hsigma : ∀ k, (sigma k).PosSemidef) :
    Tendsto (minAsymptotic rho sigma) (nhdsWithin 0 (Set.Ioi (0 : ℝ)))
      (nhds (spectralInf rho sigma)) := by sorry

end Statements

end QAlgorithms.Papers.Independent.Support50.P21

end
Source
Min- and Max- Relative Entropies and a New Entanglement Monotone, arXiv:0803.2770v3, PDF p. 9, Theorem 3 (§VIII.B): smooth min-relative entropy gives spectral inf-divergence
Read-back

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

Let III be a finite index type with decidable equality. For every integer k≥0k\ge0k≥0, let ρk,σk\rho_k,\sigma_kρk​,σk​ be complex matrices indexed by the functions {0,…,k}→I\{0,\ldots,k\}\to I{0,…,k}→I, with ρk\rho_kρk​ positive semidefinite of trace 111 and σk\sigma_kσk​ positive semidefinite, possibly zero and without trace normalization. Let PrP_rPr​ denote the orthogonal projection onto the range of rrr, define L(x)=log⁡2xL(x)=\log_2xL(x)=log2​x for x>0x>0x>0 and L(x)=−∞L(x)=-\inftyL(x)=−∞ otherwise, and put

Dmin⁡(r∥σk)=−L(Re⁡Tr⁡(Prσk)),Smin⁡ε(ρk∥σk)=sup⁡r positive semidefinite∥r−ρk∥1≤ε, Re⁡Tr⁡r≤1Dmin⁡(r∥σk),D_{\min}(r\Vert\sigma_k)=-L(\operatorname{Re}\operatorname{Tr}(P_r\sigma_k)), \qquad S_{\min}^{\varepsilon}(\rho_k\Vert\sigma_k) =\sup_{\substack{r\text{ positive semidefinite}\\ \|r-\rho_k\|_1\le\varepsilon,\ \operatorname{Re}\operatorname{Tr}r\le1}} D_{\min}(r\Vert\sigma_k),Dmin​(r∥σk​)=−L(ReTr(Pr​σk​)),Sminε​(ρk​∥σk​)=r positive semidefinite∥r−ρk​∥1​≤ε, ReTrr≤1​sup​Dmin​(r∥σk​),

where ∥X∥1=Re⁡Tr⁡X†X\|X\|_1=\operatorname{Re}\operatorname{Tr}\sqrt{X^\dagger X}∥X∥1​=ReTrX†X​ and the supremum is extended-real. For γ∈R\gamma\in\mathbb Rγ∈R, set Ak,γ=ρk−2(k+1)γσkA_{k,\gamma}=\rho_k-2^{(k+1)\gamma}\sigma_kAk,γ​=ρk​−2(k+1)γσk​, let P≥0(Ak,γ)P_{\ge0}(A_{k,\gamma})P≥0​(Ak,γ​) be its spectral projection onto the nonnegative eigenvalues, including zero, and set mk(γ)=Re⁡Tr⁡(P≥0(Ak,γ)Ak,γ)m_k(\gamma)=\operatorname{Re}\operatorname{Tr}(P_{\ge0}(A_{k,\gamma})A_{k,\gamma})mk​(γ)=ReTr(P≥0​(Ak,γ​)Ak,γ​). Then, in the topology of the extended reals,

lim⁡ε↓0(lim inf⁡k→∞Smin⁡ε(ρk∥σk)k+1)=sup⁡{γ∈R:lim inf⁡k→∞mk(γ)=1}.\lim_{\varepsilon\downarrow0} \left(\liminf_{k\to\infty} \frac{S_{\min}^{\varepsilon}(\rho_k\Vert\sigma_k)}{k+1}\right) = \sup\{\gamma\in\mathbb R:\liminf_{k\to\infty}m_k(\gamma)=1\}.ε↓0lim​(k→∞liminf​k+1Sminε​(ρk​∥σk​)​)=sup{γ∈R:k→∞liminf​mk​(γ)=1}.

The outer limit runs through strictly positive real ε\varepsilonε. The families need not consist of tensor powers or satisfy any consistency across kkk, and k=0k=0k=0 already has one coordinate. The final supremum is extended-real, with value −∞-\infty−∞ if its defining set is empty.

Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 24, 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