Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Choi and root-fidelity primal semidefinite programs

Open
QAlgorithms.Papers.Independent.Intakenext100.P75.recovery_primal_sdp

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

prove2quantum

Let ρAB\rho_{AB}ρAB​ and σAC\sigma_{AC}σAC​ be density operators on finite-dimensional nonzero complex systems, and let σACD\sigma_{ACD}σACD​ purify σAC\sigma_{AC}σAC​. Set s=σAD⊗IBs=\sqrt{\sigma_{AD}}\otimes I_Bs=σAD​​⊗IB​, reindexed onto ABDABDABD, and T(τ)=tr⁡D(sτs)T(\tau)=\operatorname{tr}_D(s\tau s)T(τ)=trD​(sτs). Then both following maxima are attained:

FC→B(ρ∥σ)=max⁡τABD≥0tr⁡Bτ=IADF(ρAB,T(τ)),F_{C\to B}(\rho\|\sigma)=\max_{\substack{\tau_{ABD}\ge0\\\operatorname{tr}_B\tau=I_{AD}}}F(\rho_{AB},T(\tau)),FC→B​(ρ∥σ)=τABD​≥0trB​τ=IAD​​max​F(ρAB​,T(τ)), FC→B(ρ∥σ)=max⁡τ≥0, tr⁡Bτ=IADZ∈L(AB), (ρZZ†T(τ))≥012tr⁡(Z+Z†).\sqrt{F_{C\to B}(\rho\|\sigma)} =\max_{\substack{\tau\ge0,\ \operatorname{tr}_B\tau=I_{AD}\\Z\in\mathcal L(AB),\ \left(\begin{smallmatrix}\rho&Z\\Z^\dagger&T(\tau)\end{smallmatrix}\right)\ge0}} \frac12\operatorname{tr}(Z+Z^\dagger).FC→B​(ρ∥σ)​=τ≥0, trB​τ=IAD​Z∈L(AB), (ρZ†​ZT(τ)​)≥0​max​21​tr(Z+Z†).

Here FFF is squared Uhlmann fidelity. The Choi variable τ\tauτ is positive but unnormalized; the second program optimizes root fidelity. These are the primal optimization formulations used in the duality argument, including for singular purifications.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P75
Formal statement
namespace QAlgorithms.Papers.Independent.Intakenext100.P75

open scoped Matrix ComplexOrder

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

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

open QAlgorithms.Papers.Independent.Information100.P08 (IsPure)

open QAlgorithms.Papers.Independent.Dependency50.P18 (bipartiteTensor)

variable {A B C D Ap Bp Cp R I J K L : Type}

variable [Fintype A] [DecidableEq A] [Nonempty A]

variable [Fintype B] [DecidableEq B] [Nonempty B]

variable [Fintype C] [DecidableEq C] [Nonempty C]

variable [Fintype D] [DecidableEq D] [Nonempty D]

variable [Fintype Ap] [DecidableEq Ap] [Nonempty Ap]

variable [Fintype Bp] [DecidableEq Bp] [Nonempty Bp]

variable [Fintype Cp] [DecidableEq Cp] [Nonempty Cp]

variable [Fintype R] [DecidableEq R] [Nonempty R]

/-- PDF p.3, (15),(16): both primal maxima are attained; (16) optimizes
ROOT fidelity, whereas (15) optimizes squared fidelity. -/
theorem recovery_primal_sdp
    [Nonempty A] [Nonempty B] [Nonempty C] [Nonempty D] (rho : Mat (A × B)) (sigma : Mat (A × C))
    (purification : Mat ((A × C) × D))
    (hrho : IsDensityMatrix rho) (hsigma : IsDensityMatrix sigma)
    (hpure : IsPure purification) (hmarginal : traceRight purification = sigma) :
    IsGreatest {v : ℝ | ∃ t : Mat ((A × B) × D), t.PosSemidef ∧
      (Matrix.of fun x y : A × D => ∑ b, t ((x.1, b), x.2) ((y.1, b), y.2)) = 1 ∧
      v = fidelity rho (choiOutput (marginalAD purification) t)} (generalizedRecoveryFidelity rho sigma) ∧
    IsGreatest {v : ℝ | ∃ t : Mat ((A × B) × D), ∃ z : Mat (A × B),
      t.PosSemidef ∧
      (Matrix.of fun x y : A × D => ∑ b, t ((x.1, b), x.2) ((y.1, b), y.2)) = 1 ∧
      (Matrix.fromBlocks rho z z.conjTranspose
        (choiOutput (marginalAD purification) t)).PosSemidef ∧
      v = (z + z.conjTranspose).trace.re / 2} (Real.sqrt (generalizedRecoveryFidelity rho sigma)) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P75
Source
The Fidelity of Recovery is Multiplicative, arXiv:1502.07973v2, PDF p. 3, Section II, (15),(16)
Read-back

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

Let A,B,C,DA,B,C,DA,B,C,D be finite nonempty types with decidable equality, let ρ,σ\rho,\sigmaρ,σ be positive semidefinite trace-one matrices on A×B,A×CA\times B,A\times CA×B,A×C, and let P=vv†P=vv^\daggerP=vv† on (A×C)×D(A\times C)\times D(A×C)×D, where ∥v∥=1\|v\|=1∥v∥=1 and tr⁡DP=σ\operatorname{tr}_D P=\sigmatrD​P=σ. Set S(a,d),(a′,d′)=∑cP((a,c),d),((a′,c),d′)S_{(a,d),(a',d')}=\sum_cP_{((a,c),d),((a',c),d')}S(a,d),(a′,d′)​=∑c​P((a,c),d),((a′,c),d′)​, and, for a matrix ttt on (A×B)×D(A\times B)\times D(A×B)×D, define OS(t)=tr⁡D(LtL)O_S(t)=\operatorname{tr}_D(LtL)OS​(t)=trD​(LtL), where L((a,b),d),((a′,b′),d′)=(S)(a,d),(a′,d′)δb,b′L_{((a,b),d),((a',b'),d')}=(\sqrt S)_{(a,d),(a',d')}\delta_{b,b'}L((a,b),d),((a′,b′),d′)​=(S​)(a,d),(a′,d′)​δb,b′​. With F(X,Y)=(Re⁡tr⁡X Y X)2F(X,Y)=(\operatorname{Re}\operatorname{tr}\sqrt{\sqrt X\,Y\,\sqrt X})^2F(X,Y)=(RetrX​YX​​)2, let GGG be the supremum of F(ρ,(id⁡A⊗R)(σ))F(\rho,(\operatorname{id}_A\otimes\mathcal R)(\sigma))F(ρ,(idA​⊗R)(σ)) over finite Kraus channels R:C→B\mathcal R:C\to BR:C→B, whose Kraus operators obey ∑jKj†Kj=I\sum_jK_j^\dagger K_j=I∑j​Kj†​Kj​=I. Both of the following maxima are attained and have the stated values:

max⁡t≥0∑bt((a,b),d),((a′,b),d′)=δa,a′δd,d′F(ρ,OS(t))=G,\max_{\substack{t\ge0\\ \sum_b t_{((a,b),d),((a',b),d')}=\delta_{a,a'}\delta_{d,d'}}} F(\rho,O_S(t))=G,t≥0∑b​t((a,b),d),((a′,b),d′)​=δa,a′​δd,d′​​max​F(ρ,OS​(t))=G, \max_{\substack{t\ge0,\ z\text{ any matrix on }A\times B\\ \sum_b t_{((a,b),d),((a',b),d')}=\delta_{a,a'}\delta_{d,d'}\\ \begin{psmallmatrix}\rho&z\\z^\dagger&O_S(t)\end{psmallmatrix}\ge0}} \frac{\operatorname{Re}\operatorname{tr}(z+z^\dagger)}2=\sqrt G.
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