Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5, Eq. (9): cost versus one-way distillation

Open
QAlgorithms.Papers.Independent.Gap48.P35.cost_distillation_tradeoff

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

prove2quantum

All systems are finite nonzero-dimensional complex quantum systems. Density states are positive semidefinite with trace one, and SSS is von Neumann entropy in bits. Write EfE_fEf​ for formation, I←I^{\leftarrow}I← for maximal Holevo correlation measured on the second party, ECE_CEC​ for asymptotic singlet preparation cost, EDE_DED​ for EPR distillation rate, CDC_DCD​ for net common randomness, and KKK for secret-key rate. A left arrow means communication from the second legitimate party to the first. Key secrecy includes all public transcripts. Asymptotic tasks use independent copies, vanishing error and no extra initially shared resource.

Let ρABC\rho_{ABC}ρABC​ be any density state. Compute its ABABAB, ACACAC, and AAA marginals from that state. The distillation term permits communication C→AC\to AC→A.

EC(ρAB)+ED←(ρAC)≤S(ρA).E_C(\rho_{AB})+E_D^{\leftarrow}(\rho_{AC})\le S(\rho_A).EC​(ρAB​)+ED←​(ρAC​)≤S(ρA​).

This retains unrestricted preparation cost and one-way distillation; the state need not be pure.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P35
Formal statement
namespace QAlgorithms.Papers.Independent.Gap48.P35

open scoped Matrix ComplexConjugate ComplexOrder

open Filter Topology

open Information100.P21 (Op)

open Information100.P08 (IsPure formation measuredCorrelation)

open Dependency50.P19 (marginalRA marginalRB)

noncomputable section

/-- PDF p.4, Theorem 5, Eq.(9): unrestricted formation cost and one-way distillation from the two marginals of the same arbitrary state. -/
theorem cost_distillation_tradeoff {A : Type} {B : Type} {C : Type}
    [Fintype A] [DecidableEq A] [Nonempty A] [Fintype B] [DecidableEq B] [Nonempty B] [Fintype C] [DecidableEq C] [Nonempty C]
    (rho : Op (A × (B × C))) (hrho : IsDensityMatrix rho) :
    finiteSystemPreparationCost (marginalRA rho) + distillLeft (marginalRB rho) ≤ vnEntropy (traceRight rho) := by sorry

end

end QAlgorithms.Papers.Independent.Gap48.P35
Source
Monogamy of entanglement and other correlations, arXiv:quant-ph/0310037v1, PDF p. 4, Theorem 5, Eq. (9): cost versus one-way distillation
Read-back

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

For all finite, nonempty sets A,B,CA,B,CA,B,C, each equipped with decidable equality, and every positive semidefinite complex matrix ρ\rhoρ on A⊗B⊗CA\otimes B\otimes CA⊗B⊗C with Tr⁡ρ=1\operatorname{Tr}\rho=1Trρ=1, let ρAB=Tr⁡Cρ\rho_{AB}=\operatorname{Tr}_C\rhoρAB​=TrC​ρ, ρAC=Tr⁡Bρ\rho_{AC}=\operatorname{Tr}_B\rhoρAC​=TrB​ρ, and ρA=Tr⁡BCρ\rho_A=\operatorname{Tr}_{BC}\rhoρA​=TrBC​ρ. Then

Eprep(ρAB)+DC→A(ρAC)≤S(ρA),E_{\mathrm{prep}}(\rho_{AB})+D_{C\to A}(\rho_{AC})\le S(\rho_A),Eprep​(ρAB​)+DC→A​(ρAC​)≤S(ρA​),

where S(X)=−Re⁡Tr⁡[f(X)]S(X)=-\operatorname{Re}\operatorname{Tr}[f(X)]S(X)=−ReTr[f(X)], with f(x)=xlog⁡2xf(x)=x\log_2xf(x)=xlog2​x applied by spectral functional calculus. Here Eprep(σXY)E_{\mathrm{prep}}(\sigma_{XY})Eprep​(σXY​) is the infimum over all real eee such that, for every ϵ,δ>0\epsilon,\delta>0ϵ,δ>0, there exist natural numbers m,nm,nm,n with n>0n>0n>0 and a finite local-operations-and-classical-communication map Λ\LambdaΛ satisfying ∣e−m/n∣≤δ|e-m/n|\le\delta∣e−m/n∣≤δ and 12∥Λ(s⊗m)−σ⊗n∥1≤ϵ\frac12\|\Lambda(s^{\otimes m})-\sigma^{\otimes n}\|_1\le\epsilon21​∥Λ(s⊗m)−σ⊗n∥1​≤ϵ, where sss is the projector onto (∣01⟩−∣10⟩)/2(|01\rangle-|10\rangle)/\sqrt2(∣01⟩−∣10⟩)/2​, ∥X∥1=Re⁡Tr⁡X†X\|X\|_1=\operatorname{Re}\operatorname{Tr}\sqrt{X^\dagger X}∥X∥1​=ReTrX†X​, and the systems and their tensor powers are reindexed using the fixed finite-coordinate bijections in the definitions. Such a finite local map is obtained recursively from identity maps by finitely many local Kraus measurements, each satisfying ∑iKi†Ki=I\sum_iK_i^\dagger K_i=I∑i​Ki†​Ki​=I, with a subsequent finite local map chosen separately for each outcome and the resulting branches summed. The number mmm may be zero, and the witnessing block length nnn may depend on both tolerances. The quantity DY→X(σXY)D_{Y\to X}(\sigma_{XY})DY→X​(σXY​) is the supremum over all real rrr for which there is a protocol for every natural block length nnn, including zero: the sender YYY applies a finite instrument with a positive number of classical outcomes and rectangular Kraus operators Kij:Y⊗n→CMnK_{ij}:Y^{\otimes n}\to\mathbb C^{M_n}Kij​:Y⊗n→CMn​, satisfying ∑i,jKij†Kij=I\sum_{i,j}K_{ij}^\dagger K_{ij}=I∑i,j​Kij†​Kij​=I, and, conditional on iii, the receiver XXX applies a finite Kraus channel to CMn\mathbb C^{M_n}CMn​, with Mn>0M_n>0Mn​>0. If ωn\omega_nωn​ is the sum of the resulting branches, the required limits are ⟨ΦMn∣ωn∣ΦMn⟩→1\langle\Phi_{M_n}|\omega_n|\Phi_{M_n}\rangle\to1⟨ΦMn​​∣ωn​∣ΦMn​​⟩→1 and log⁡2Mn/n→r\log_2M_n/n\to rlog2​Mn​/n→r, where ∣ΦM⟩=M−1/2∑j=0M−1∣jj⟩|\Phi_M\rangle=M^{-1/2}\sum_{j=0}^{M-1}|jj\rangle∣ΦM​⟩=M−1/2∑j=0M−1​∣jj⟩. No communication bound is imposed in this definition of DDD; the quotient at n=0n=0n=0 is defined as zero and does not affect either limit.

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