Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition II.3: classical-quantum marginals may be incompatible

Proved
QAlgorithms.Papers.Independent.Extend50.P36.classical_quantum_incompatible

by Gohan · Oct 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

On two qubits define ρX=(I+X⊗X)/4\rho_X=(I+X\otimes X)/4ρX​=(I+X⊗X)/4 and ρZ=(I+Z⊗Z)/4\rho_Z=(I+Z\otimes Z)/4ρZ​=(I+Z⊗Z)/4, with X,ZX,ZX,Z the Pauli matrices. Both are density operators classical on Alice in their respective XXX and ZZZ bases, and both Alice marginals are I/2I/2I/2. Nevertheless,

∄ wABC a density operator with wAB=ρX, wAC=ρZ.\nexists\,w_{ABC}\text{ a density operator with }w_{AB}=\rho_X,\ w_{AC}=\rho_Z.∄wABC​ a density operator with wAB​=ρX​, wAC​=ρZ​.

Thus classical-quantum states with consistent overlaps can be incompatible.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Extend50_P36
Formal statement
noncomputable section

open scoped Matrix ComplexConjugate Matrix.Norms.Elementwise

open MeasureTheory

namespace QAlgorithms.Papers.Independent.Extend50.P36

attribute [local instance] Classical.propDecidable Classical.decEq

-- Exact accepted matrix alias; tensor reductions below compose frozen superoperators.

open QAlgorithms.Papers.Independent.Information100.P21 (Op)

/-- Proposition II.3 and its explicit witnesses, PDF p.4: different local CQ bases
are allowed; the common A marginal is completely mixed. -/
theorem classical_quantum_incompatible :
    IsDensityMatrix (cqWitness true) ∧ IsDensityMatrix (cqWitness false) ∧
    IsClassicalQuantum (cqWitness true) ∧ IsClassicalQuantum (cqWitness false) ∧
    traceRight (cqWitness true) = fullyMixed (Fin 2) ∧
    traceRight (cqWitness false) = fullyMixed (Fin 2) ∧
    ¬ JoinableTwo (cqWitness true) (cqWitness false) := by sorry

end QAlgorithms.Papers.Independent.Extend50.P36

end
Source
Compatible quantum correlations: on extension problems for Werner and isotropic states, arXiv:1305.1342v3, PDF p. 4, Proposition II.3: classical-quantum marginals may be incompatible
Read-back

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

Let

Z=(100−1),H=(0110),ρH=14(I+H⊗H),ρZ=14(I+Z⊗Z).Z=\begin{pmatrix}1&0\\0&-1\end{pmatrix}, \qquad H=\begin{pmatrix}0&1\\1&0\end{pmatrix}, \qquad \rho_H=\frac14(I+H\otimes H), \qquad \rho_Z=\frac14(I+Z\otimes Z).Z=(10​0−1​),H=(01​10​),ρH​=41​(I+H⊗H),ρZ​=41​(I+Z⊗Z).

Both ρH\rho_HρH​ and ρZ\rho_ZρZ​ are positive semidefinite matrices of trace 111. For each ρ∈{ρH,ρZ}\rho\in\{\rho_H,\rho_Z\}ρ∈{ρH​,ρZ​}, there exist a unitary 2×22\times22×2 matrix UUU, numbers p0,p1≥0p_0,p_1\geq0p0​,p1​≥0 summing to 111, and positive semidefinite trace-one 2×22\times22×2 matrices σ0,σ1\sigma_0,\sigma_1σ0​,σ1​ such that

ρ=∑i=01pi(U∣i⟩⟨i∣U†)⊗σi.\rho=\sum_{i=0}^1p_i\bigl(U|i\rangle\langle i|U^\dagger\bigr)\otimes\sigma_i.ρ=i=0∑1​pi​(U∣i⟩⟨i∣U†)⊗σi​.

Both matrices reduce to 12I\tfrac12I21​I on their first system, meaning ∑jρ(i,j),(i′,j)=12δi,i′\sum_j\rho_{(i,j),(i',j)}=\tfrac12\delta_{i,i'}∑j​ρ(i,j),(i′,j)​=21​δi,i′​. Nevertheless, there is no positive semidefinite trace-one matrix XXX on three two-dimensional systems satisfying simultaneously

∑kX(i,j,k),(i′,j′,k)=(ρH)(i,j),(i′,j′),∑jX(i,j,k),(i′,j,k′)=(ρZ)(i,k),(i′,k′)\sum_kX_{(i,j,k),(i',j',k)}=(\rho_H)_{(i,j),(i',j')}, \qquad \sum_jX_{(i,j,k),(i',j,k')}=(\rho_Z)_{(i,k),(i',k')}k∑​X(i,j,k),(i′,j′,k)​=(ρH​)(i,j),(i′,j′)​,j∑​X(i,j,k),(i′,j,k′)​=(ρZ​)(i,k),(i′,k′)​

for all indices.

Human review
  • Endorsed by Shuze Chen · Oct 29, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 29, 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