Proposition 5.1 / (70),(71): normalized CP encoding replacement
OpenQAlgorithms.Papers.Independent.Gap20.P07.normalizedCP_encodingReplacementLet have finite dimensions , let be a density matrix on , let be completely positive with , and let be CPTP. No trace condition on at other inputs is assumed. There exists a CPTP replacement such that
If , it has the form for an isometry . If , it has the form
for an isometry . Fidelity is the unnormalized squared CP-map formula. Both dimensional branches belong to the replacement result.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap20_P07
noncomputable section
open Filter MeasureTheory
open scoped Topology ENNReal NNReal Matrix ComplexOrder ComplexConjugate
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
attribute [local instance] IsStarNormal.instContinuousFunctionalCalculus
namespace QAlgorithms.Papers.Independent.Gap20.P07
open QAlgorithms
open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33
(State IsTraceClass IsTraceClassChannel traceSize rankOne tensorApply)
open QAlgorithms.Papers.Independent.Frontier30.P14
(Operation isometryCoeff conjugateKernel)
open QAlgorithms.Papers.Independent.Claimed50.P35
(stabilizedTraceClassNorm traceClassTensorMap traceClassChannelPower
transpose transposeBound TraceLinear tracePair MeasurePrepare)
open QAlgorithms.Papers.Independent.Gap48.P29 (fe FiniteCPMap)
/-- PDF pp.20--22, Proposition 5.1, (70)--(71): E is CP and normalized only on
rho. Both dimension branches and their isometry directions are retained. -/
theorem normalizedCP_encodingReplacement (eta kappa : ℕ) (heta : 0 < eta) (hkappa : 0 < kappa)
(rho : Matrix (Fin eta) (Fin eta) ℂ) (hrho : IsDensityMatrix rho)
(E : Superop (Fin eta) (Fin kappa)) (hE : FiniteCPMap E) (htrace : (E rho).trace = 1)
(T : Superop (Fin kappa) (Fin eta)) (hT : IsChannel T) :
∃ replacement : Superop (Fin eta) (Fin kappa), IsChannel replacement ∧
(fe rho (T ∘ E))^2 ≤ fe rho (T ∘ replacement) ∧
(eta ≤ kappa → ∃ V : Matrix (Fin kappa) (Fin eta) ℂ,
V.conjTranspose * V = 1 ∧ ∀ X, replacement X = V * X * V.conjTranspose) ∧
(kappa < eta → ∃ W : Matrix (Fin eta) (Fin kappa) ℂ,
W.conjTranspose * W = 1 ∧ ∀ X,
replacement X = W.conjTranspose * X * W +
((((1 - W * W.conjTranspose) * X).trace / (kappa : ℂ))) • 1) := by sorry
end QAlgorithms.Papers.Independent.Gap20.P07
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For all positive integers , every positive semidefinite with , every map equal on all matrices to a complex linear map whose blockwise amplification preserves positive semidefiniteness for every , and every map admitting a finite Kraus representation with , suppose that the complex trace satisfies . There then exists a map admitting a finite Kraus representation with identity Kraus sum such that
where denotes a matrix unit. Moreover, if , there is with and for every ; if , there is with and, for every ,
The hypotheses impose normalization of at and do not impose a trace nonincrease condition on . Equality of the dimensions falls under the first alternative.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.