§V (45)–(48): comparison with the Matthews–Wehner bound
ProvedQAlgorithms.Papers.Independent.Support50.P11.nonSignalling_le_matthewsWehnerLet be finite CPTP and . Write for its unnormalized Choi matrix, and define as the supremum of over density matrices and operators satisfying
Then
The relaxation is an operator inequality on the marginal, retaining the same full-transpose objective. This compares the NS bound with the Matthews–Wehner bound in the form used here.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P11
noncomputable section
open scoped Matrix ComplexConjugate ComplexOrder ENNReal
namespace QAlgorithms.Papers.Independent.Support50.P11
open QAlgorithms.Papers.Independent.Information100.P34 (choi dephasing permutationMatrix)
open QAlgorithms.Papers.Independent.Information100.P21 (pt)
open QAlgorithms.Papers.Independent.Dependency50.P09 (tensorAction)
open QAlgorithms.Papers.Independent.Information100.P76 (swapMatrix)
open QAlgorithms.Papers.Independent.Information100.P69 (replacer)
open QAlgorithms.Papers.Independent.Gap20.P19 (wernerIsometry)
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
section Statements
variable {I O : Type} [Fintype I] [Nonempty I] [Fintype O] [Nonempty O]
/-- PDF p.8, (45)--(48): the NS optimum lies below the relaxed SDP whose
marginal equality is replaced by a Loewner inequality. The (47) class typo is EA. -/
theorem nonSignalling_le_matthewsWehner
(N : Superop I O) (hN : IsChannel N) (K : ℕ) (hK : 0 < K) :
optimalCodeFidelity .EA N K ≤ optimalCodeFidelity .NS N K ∧ optimalCodeFidelity .NS N K ≤ relaxedBound N (K * K) := by sorry
end Statements
end QAlgorithms.Papers.Independent.Support50.P11
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For finite input and output index types , every channel and positive integer satisfy
Here , and is the real supremum of over codes in class , with the projector onto and
Every code is a channel and has a channel with for all . The non-signalling class also requires a channel with . The entanglement-assisted class instead requires to arise by adjoining a positive semidefinite trace-one state on finite positive-dimensional registers , then applying local channels from to and from to . Channels have finite Kraus representations with , and denotes positive semidefinite difference. The right-hand supremum uses the inequality , and is included.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.