Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§4.3: generalized-ensemble bounds and supremum

Open
QAlgorithms.Papers.Independent.Gap48.P18.generalized_ensemble_capacity_and_deficit

by Goku · Oct 22, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let AAA be nonempty, closed, and of finite capacity. If M(A)\mathcal M(A)M(A) is the set of Borel probability laws concentrated on AAA, then

Cˉ(A)=sup⁡μ∈M(A)χ(μ).\bar C(A)=\sup_{\mu\in\mathcal M(A)}\chi(\mu).Cˉ(A)=μ∈M(A)sup​χ(μ).

For every such law, with average ρˉ\bar\rhoρˉ​, one has ∫D(ρ∥Ω(A)) dμ≤Cˉ(A)\int D(\rho\|\Omega(A))\,d\mu\le\bar C(A)∫D(ρ∥Ω(A))dμ≤Cˉ(A) and

Cˉ(A)−χ(μ)≥D(ρˉ∥Ω(A))≥12∥ρˉ−Ω(A)∥12.\bar C(A)-\chi(\mu)\ge D(\bar\rho\|\Omega(A))\ge\tfrac12\|\bar\rho-\Omega(A)\|_1^2.Cˉ(A)−χ(μ)≥D(ρˉ​∥Ω(A))≥21​∥ρˉ​−Ω(A)∥12​.

Both quantities subtracted in the deficit are finite; average entropy need not be finite. This extends the finite-ensemble bounds to arbitrary generalized ensembles.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P18
Formal statement
noncomputable section

open MeasureTheory Filter

open scoped ENNReal NNReal Topology Matrix.Norms.L2Operator

open QAlgorithms.Papers.Independent.Dependency50.P33
  (State PureState traceSize traceClosure IsTraceClassChannel productCoeff)

open QAlgorithms.Papers.Independent.Frontier18.P07
  (GeneralizedEnsemble HasFiniteSupport HasAverage averageCoefficients
   extendedEntropyNats extendedRelativeEntropyNats
   finiteEnsembleCapacity pureStates)

open QAlgorithms.Papers.Independent.Frontier18.P01 (traceFirst traceSecond)

namespace QAlgorithms.Papers.Independent.Gap48.P18

attribute [local instance] Classical.propDecidable Classical.decEq

/-- §4.3: generalized-ensemble bounds and supremum, arXiv:quant-ph/0510073v2, PDF pp.39.
 -/
theorem generalized_ensemble_capacity_and_deficit {I : Type} [Countable I] [Nonempty I]
    (A : Set (State I)) (hA : A.Nonempty) (hclosed : IsClosed A)
    (hfinite : setChiCapacity A < ⊤) :
    setChiCapacity A = (⨆ mu : {mu : GeneralizedEnsemble I //
      MeasureConcentratedOn (mu : Measure (State I)) A}, measureHolevo (mu.val : Measure (State I))) ∧
    ∀ mu : GeneralizedEnsemble I, MeasureConcentratedOn (mu : Measure (State I)) A →
      (∫⁻ rho, stateRelativeEntropyNats rho (optimalAverage A hA) ∂(mu : Measure (State I))) ≤ setChiCapacity A ∧
      measureHolevo (mu : Measure (State I)) < ⊤ ∧
      stateRelativeEntropyNats (probabilityAverage mu) (optimalAverage A hA) < ⊤ ∧
      stateRelativeEntropyNats (probabilityAverage mu) (optimalAverage A hA) ≤
        ENNReal.ofReal ((setChiCapacity A).toReal - (measureHolevo (mu : Measure (State I))).toReal) ∧
      distanceSquared (probabilityAverage mu) (optimalAverage A hA) / 2 ≤
        stateRelativeEntropyNats (probabilityAverage mu) (optimalAverage A hA) := by sorry

end QAlgorithms.Papers.Independent.Gap48.P18

end
Source
Entropic characteristics of subsets of states, arXiv:quant-ph/0510073v2, PDF p. 39, §4.3: generalized-ensemble bounds and supremum
Read-back

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

Let III be countable and nonempty. A state is a complex matrix XXX admitting Xij=∑rar(i)br(j)‾X_{ij}=\sum_r a_r(i)\overline{b_r(j)}Xij​=∑r​ar​(i)br​(j)​, with ar,br∈ℓ2(I;C)a_r,b_r\in\ell^2(I;\mathbb C)ar​,br​∈ℓ2(I;C) and ∑r∥ar∥∥br∥<∞\sum_r\|a_r\|\|b_r\|<\infty∑r​∥ar​∥∥br​∥<∞, whose finite quadratic forms are real and nonnegative and whose real diagonal is summable with sum 111. Define T(X)T(X)T(X) as the infimum of such expansion sums, with empty infimum +∞+\infty+∞, and give states the topology generated by T(Y−X)<εT(Y-X)<\varepsilonT(Y−X)<ε for states XXX and real ε>0\varepsilon>0ε>0. Define

D(X∥Y)=sup⁡s⊆Is finite{[Re⁡tr⁡(Xslog⁡Xs−Xslog⁡Ys+Ys−Xs)]+,ker⁡Ys⊆ker⁡Xs,+∞,otherwise,D(X\|Y)=\sup_{\substack{s\subseteq I\\s\text{ finite}}} \begin{cases} [\operatorname{Re}\operatorname{tr}(X_s\log X_s-X_s\log Y_s+Y_s-X_s)]_+, &\ker Y_s\subseteq\ker X_s,\\ +\infty,&\text{otherwise}, \end{cases}D(X∥Y)=s⊆Is finite​sup​{[Retr(Xs​logXs​−Xs​logYs​+Ys​−Xs​)]+​,+∞,​kerYs​⊆kerXs​,otherwise,​

where the empty principal submatrix is included, [r]+=max⁡(r,0)[r]_+=\max(r,0)[r]+​=max(r,0), and matrix logarithms use continuous functional calculus of the natural logarithm with log⁡0=0\log0=0log0=0. For a Borel probability measure μ\muμ, let mμij=∫Xij dμ(X)m_\mu{}_{ij}=\int X_{ij}\,d\mu(X)mμ​ij​=∫Xij​dμ(X), χ(μ)=∫−D(X∥mμ) dμ(X)\chi(\mu)=\int^-D(X\|m_\mu)\,d\mu(X)χ(μ)=∫−D(X∥mμ​)dμ(X), and b(μ)b(\mu)b(μ) be its state average under the fixed chosen continuous barycenter map, whose coordinates are integrable and equal these integrals. For a state set AAA, let C(A)C(A)C(A) be the supremum of χ(ν)\chi(\nu)χ(ν) over probability measures with finite topological support contained in AAA. Suppose AAA is nonempty and closed and C(A)<+∞C(A)<+\inftyC(A)<+∞. Let ωA\omega_AωA​ be a chosen state satisfying D(X∥ωA)≤C(A)D(X\|\omega_A)\le C(A)D(X∥ωA​)≤C(A) for all X∈AX\in AX∈A, if one exists, and otherwise the member of AAA chosen from its given nonemptiness hypothesis. Then

C(A)=sup⁡μ a Borel probability measureX∈A μ-almost everywhereχ(μ).C(A)=\sup_{\substack{\mu\text{ a Borel probability measure}\\X\in A\ \mu\text{-almost everywhere}}}\chi(\mu).C(A)=μ a Borel probability measureX∈A μ-almost everywhere​sup​χ(μ).

For every Borel probability measure μ\muμ assigning full measure to AAA, all of the following hold:

∫−D(X∥ωA) dμ(X)≤C(A),χ(μ)<+∞,D(b(μ)∥ωA)<+∞,\int^-D(X\|\omega_A)\,d\mu(X)\le C(A),\qquad \chi(\mu)<+\infty,\qquad D(b(\mu)\|\omega_A)<+\infty,∫−D(X∥ωA​)dμ(X)≤C(A),χ(μ)<+∞,D(b(μ)∥ωA​)<+∞, D(b(μ)∥ωA)≤[C(A)R−χ(μ)R]+,T(b(μ)−ωA)22≤D(b(μ)∥ωA).D(b(\mu)\|\omega_A) \le \bigl[C(A)_{\mathbb R}-\chi(\mu)_{\mathbb R}\bigr]_+, \qquad \frac{T(b(\mu)-\omega_A)^2}{2}\le D(b(\mu)\|\omega_A).D(b(μ)∥ωA​)≤[C(A)R​−χ(μ)R​]+​,2T(b(μ)−ωA​)2​≤D(b(μ)∥ωA​).

Here C(A)RC(A)_{\mathbb R}C(A)R​ and χ(μ)R\chi(\mu)_{\mathbb R}χ(μ)R​ are the real values of the finite extended quantities, and the last inequality uses extended-nonnegative-real arithmetic.

Human review
  • Endorsed by Shuze Chen · Oct 22, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 22, 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