The completely bounded one-to-alpha norm approaches the diamond norm
OpenQAlgorithms.Papers.Independent.Intakenext100.P15.cb_norm_diamond_limitLet be a complex-linear completely positive map on nonzero finite-dimensional spaces, without a trace-preserving requirement. Its Choi matrix uses the unnormalized vector . For , let
where ranges over all density operators and is the Schatten norm. Each supremum is attained. Define the unhalved diamond norm by , where ranges over arbitrary matrices on and the reference is isomorphic to the input. Then
This relates the positive-Choi one-to-alpha quantity to the stabilized trace norm at its endpoint.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P15
noncomputable section
open scoped Matrix ComplexOrder ComplexConjugate Topology
open Filter
namespace QAlgorithms.Papers.Independent.Intakenext100.P15
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
-- These are pinned accepted definitions, not other chunks from the current run.
open QAlgorithms.Papers.Independent.Bridge50.P43 (NormalizedDensity payoff)
open QAlgorithms.Papers.Independent.Dependency50.P09 (Divergence logExt sandwichedD petzD)
open QAlgorithms.Papers.Independent.Information100.P34 (choi)
open QAlgorithms.Papers.Independent.Information100.P63 (schatten)
open QAlgorithms.Papers.Independent.Information100.P71
(supportPower SupportIncluded stateRelative channelPower diamond)
/-- PDF p.3, unnumbered observation following (2.2): CB 1-to-alpha tends to
the diamond norm for CP maps, including maps that are not trace preserving.
Attainment witnesses elaborate the maximum in the defining equation (2.1). -/
theorem cb_norm_diamond_limit
{A B : Type} [Fintype A] [Fintype B] [DecidableEq A] [DecidableEq B]
[Nonempty A] [Nonempty B] (M : Matrix A A ℂ →ₗ[ℂ] Matrix B B ℂ)
(hM : QAlgorithms.Entanglement.IsCompletelyPositive M) :
Tendsto (fun alpha : ℝ => cbOneAlpha alpha M) (𝓝[>] (1 : ℝ))
(𝓝 ((diamond M : ℝ) : EReal)) ∧
cbOneAlpha 1 M = (diamond M : EReal) ∧
(∀ alpha : ℝ, 1 ≤ alpha → ∃ rho : NormalizedDensity A,
let S := Matrix.kronecker (supportPower rho.1 (1 / (2 * alpha))) (1 : Matrix B B ℂ)
cbOneAlpha alpha M = (schatten alpha (S * choi M * S) : EReal)) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P15
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be finite nonempty index sets with decidable equality, and let be a complex-linear map from matrices on to matrices on . Assume complete positivity in the explicit sense that, for every natural number , including , and every positive semidefinite matrix on , the matrix is positive semidefinite. Let have entries . For a positive semidefinite matrix , define by applying to nonzero eigenvalues and assigning zero at zero eigenvalues, and, for , define
where the eigenvalues are listed with multiplicity and the supremum is taken in the extended real numbers. Define and
where this supremum is real-valued and the test matrices need not be positive or Hermitian. The assertion is
with the real diamond value embedded into the extended real numbers, and, for every real , there exists a density matrix attaining the displayed supremum defining . No trace-preservation assumption is imposed on .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.