Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 6a

Open
QAlgorithms.Papers.Independent.Gap48.P18.layer_entropy_and_relative_entropy

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

prove2quantum

Choose an eigenbasis of a density state σ\sigmaσ and let L(σ)L(\sigma)L(σ) consist of all density states with the same diagonal as σ\sigmaσ in that basis. The layer is compact and convex, and

sup⁡ρ∈L(σ)H(ρ)=H(σ).\sup_{\rho\in L(\sigma)}H(\rho)=H(\sigma).ρ∈L(σ)sup​H(ρ)=H(σ).

Entropy is finite and continuous on the layer exactly when H(σ)<∞H(\sigma)<\inftyH(σ)<∞. In that case every ρ∈L(σ)\rho\in L(\sigma)ρ∈L(σ) has finite entropy and relative entropy, with D(ρ∥σ)=H(σ)−H(ρ)D(\rho\|\sigma)=H(\sigma)-H(\rho)D(ρ∥σ)=H(σ)−H(ρ). If H(σ)=∞H(\sigma)=\inftyH(σ)=∞, every pure state in the layer has infinite relative entropy to σ\sigmaσ. This describes entropy within a fixed diagonal layer.

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

/-- Proposition 6a, arXiv:quant-ph/0510073v2, PDF pp.22,23.
Chosen eigenbasis is the coordinate basis; hdiag enforces it. Repeated eigenvalues do not erase the choice. The finite difference is written additively after both finiteness conclusions. -/
theorem layer_entropy_and_relative_entropy {I : Type} [Countable I] [Nonempty I]
    (sigma : State I) (hdiag : Diagonal sigma) :
    IsCompact (layer sigma) ∧ CoefficientConvexStates (layer sigma) ∧
      entropySup (layer sigma) = stateEntropyNats sigma ∧
      (FiniteContinuousOn stateEntropyNats (layer sigma) ↔ stateEntropyNats sigma < ⊤) ∧
      (stateEntropyNats sigma < ⊤ → ∀ rho ∈ layer sigma,
        stateEntropyNats rho < ⊤ ∧ stateRelativeEntropyNats rho sigma < ⊤ ∧
          stateRelativeEntropyNats rho sigma + stateEntropyNats rho = stateEntropyNats sigma) ∧
      (stateEntropyNats sigma = ⊤ → ∀ rho : PureState I,
        rho.val ∈ layer sigma → stateRelativeEntropyNats rho.val sigma = ⊤) := by sorry

end QAlgorithms.Papers.Independent.Gap48.P18

end
Source
Entropic characteristics of subsets of states, arXiv:quant-ph/0510073v2, PDF p. 22, Proposition 6a
Read-back

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

Let III be an at most countable, nonempty index set. A state RRR is a complex matrix admitting an expansion Rij=∑mum,ivm,j‾R_{ij}=\sum_m u_{m,i}\overline{v_{m,j}}Rij​=∑m​um,i​vm,j​​, with um,vm∈ℓ2(I;C)u_m,v_m\in\ell^2(I;\mathbb C)um​,vm​∈ℓ2(I;C) and ∑m∥um∥∥vm∥<∞\sum_m\|u_m\|\|v_m\|<\infty∑m​∥um​∥∥vm​∥<∞, whose finite quadratic forms are real and nonnegative and whose real diagonal is summable with sum 111. Let T(X)T(X)T(X) be the infimum of the expansion costs [∑m∥um∥∥vm∥]+[\sum_m\|u_m\|\|v_m\|]_+[∑m​∥um​∥∥vm​∥]+​, with empty infimum ∞\infty∞, and use the state topology generated by T(R−R0)<εT(R-R_0)<\varepsilonT(R−R0​)<ε for ε>0\varepsilon>0ε>0. Define S(X)S(X)S(X) as the supremum, over all finite principal compressions XsX_sXs​, including the empty one, of

[Re⁡tr⁡(Xs)log⁡ ⁣(Re⁡tr⁡(Xs))−Re⁡tr⁡(Xslog⁡cfcXs)]+,\left[ \operatorname{Re}\operatorname{tr}(X_s)\log\!\bigl(\operatorname{Re}\operatorname{tr}(X_s)\bigr) -\operatorname{Re}\operatorname{tr}(X_s\log_{\mathrm{cfc}}X_s) \right]_+,[Retr(Xs​)log(Retr(Xs​))−Retr(Xs​logcfc​Xs​)]+​,

and define D(X∥Y)=sup⁡sd(Xs∥Ys)D(X\|Y)=\sup_s d(X_s\|Y_s)D(X∥Y)=sups​d(Xs​∥Ys​), where d(B∥C)d(B\|C)d(B∥C) is [Re⁡tr⁡(Blog⁡cfcB−Blog⁡cfcC+C−B)]+[\operatorname{Re}\operatorname{tr}(B\log_{\mathrm{cfc}}B-B\log_{\mathrm{cfc}}C+C-B)]_+[Retr(Blogcfc​B−Blogcfc​C+C−B)]+​ if ker⁡C⊆ker⁡B\ker C\subseteq\ker BkerC⊆kerB, and is ∞\infty∞ otherwise. Here [r]+[r]_+[r]+​ denotes max⁡(r,0)\max(r,0)max(r,0) in [0,∞][0,\infty][0,∞], and the functional calculus uses the total real logarithm with log⁡0=0\log0=0log0=0. For every state σ\sigmaσ with σij=0\sigma_{ij}=0σij​=0 whenever i≠ji\ne ji=j, let Lσ={R:R is a state and Rii=σii for every i}L_\sigma=\{R:R\text{ is a state and }R_{ii}=\sigma_{ii}\text{ for every }i\}Lσ​={R:R is a state and Rii​=σii​ for every i}. Then LσL_\sigmaLσ​ is compact, its coefficient matrices form a real convex set, and

sup⁡R∈LσS(R)=S(σ).\sup_{R\in L_\sigma}S(R)=S(\sigma).R∈Lσ​sup​S(R)=S(σ).

Moreover, SSS is finite everywhere and continuous on LσL_\sigmaLσ​ if and only if S(σ)<∞S(\sigma)<\inftyS(σ)<∞. If S(σ)<∞S(\sigma)<\inftyS(σ)<∞, then every R∈LσR\in L_\sigmaR∈Lσ​ satisfies S(R)<∞S(R)<\inftyS(R)<∞, D(R∥σ)<∞D(R\|\sigma)<\inftyD(R∥σ)<∞, and D(R∥σ)+S(R)=S(σ)D(R\|\sigma)+S(R)=S(\sigma)D(R∥σ)+S(R)=S(σ). If S(σ)=∞S(\sigma)=\inftyS(σ)=∞, then every pure state R∈LσR\in L_\sigmaR∈Lσ​—meaning a state with Rij=uiuj‾R_{ij}=u_i\overline{u_j}Rij​=ui​uj​​ for some u∈ℓ2(I;C)u\in\ell^2(I;\mathbb C)u∈ℓ2(I;C) of norm 111—satisfies D(R∥σ)=∞D(R\|\sigma)=\inftyD(R∥σ)=∞.

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