Conditional attained thermal coherent-information maximum
OpenQAlgorithms.Papers.Independent.Support50.P06.singleLetter_thermalOptimizerAssume global minimum output entropy conjecture 2. Let and be finite. For arbitrary normalized single-mode input states with mean photons at most , define using the two marginals of the fixed vacuum dilation. Entropies are in nats. Then
Both receiver entropies are finite for every feasible input. The normalized thermal input of mean exactly attains this maximum, Bob's output is thermal of mean , and the input is the isotropic Gaussian coherent mixture of variance (vacuum at zero). The optimization has no Gaussian restriction.
This identifies as a constrained coherent-information optimization; it does not define an operational quantum transmission capacity.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P06
noncomputable section
namespace QAlgorithms.Papers.Independent.Support50.P06
open MeasureTheory Filter
open scoped ENNReal BigOperators Topology
open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33 (State IsDensity rankOne)
open QAlgorithms.Papers.Independent.Next46.P32
(NumberString FockKernel identicalThermalKernel vacuum productCoherentKernel
jointMix bobMarginal charlieMarginal pureLoss broadcast mixedOutput
spectralEntropyNats FockPOVM realFockTracePair modeEnergy complexNormal
StrongConjecture2)
open QAlgorithms.Papers.Independent.Claimed50.P39 (g)
open QAlgorithms.Papers.Independent.Claimed50.P11 (thermalEntropyInverse)
attribute [local instance] Classical.propDecidable Classical.decEq
/-- PDF p.3, (12): attained, unrestricted coherent-information optimum under
mean photons <=Nbar, conditional on global MOE2. Printed Gaussian variance K
reuses Bob's entropy variable; changed to Alice variance Nbar as fixed by (10)
and the theorem's own prior. Bob's thermal mean is eta*Nbar, not K. Finite
receiver entropy is concluded for every feasible input; it is not assumed. -/
theorem singleLetter_thermalOptimizer (hmoe : StrongConjecture2)
(eta Nbar : ℝ) (heta : 1 / 2 < eta) (heta_one : eta ≤ 1) (hNbar : 0 ≤ Nbar) :
coherentInformationMaximum 1 eta Nbar =
((g (eta * Nbar) - g ((1 - eta) * Nbar) : ℝ) : EReal) ∧
holevoMessagePrivacyCapacityNats 1 eta Nbar = ENNReal.ofReal (g (eta * Nbar) - g ((1 - eta) * Nbar)) ∧
IsDensity (identicalThermalKernel 1 Nbar) ∧
totalEnergy (identicalThermalKernel 1 Nbar) = ENNReal.ofReal Nbar ∧
pureLoss 1 eta (identicalThermalKernel 1 Nbar) = identicalThermalKernel 1 (eta * Nbar) ∧
coherentMixture Nbar = identicalThermalKernel 1 Nbar ∧
vacuumLossCoherentInformationNats eta (identicalThermalKernel 1 Nbar) =
g (eta * Nbar) - g ((1 - eta) * Nbar) ∧
(∀ rho : State (NumberString 1), totalEnergy rho.coefficients ≤ ENNReal.ofReal Nbar →
spectralEntropyNats (pureLoss 1 eta rho.coefficients) ≠ ⊤ ∧
spectralEntropyNats ((charlieMarginal (broadcast 1 eta rho.coefficients))) ≠ ⊤ ∧
vacuumLossCoherentInformationNats eta rho.coefficients ≤ g (eta * Nbar) - g ((1 - eta) * Nbar)) := by sorry
end QAlgorithms.Papers.Independent.Support50.P06
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Assume that, for every positive integer , every , and every , the kernels
satisfy and , and every density kernel with satisfies . Here and in the extended nonnegative reals. For any and , put . The theorem asserts
It further asserts that every one-mode density kernel with has , , and . All the quantities here have the following meanings. A density kernel on has an expansion with and , has real nonnegative quadratic forms on every finite set of indices, and has summable real diagonal entries with sum . The entropy is the infimum of over expansions , with nonnegative summable weights of sum , unit vectors at positive-weight indices, and orthogonality between distinct positive-weight vectors; zero-weight vectors are unrestricted, and no such expansion gives entropy . The maps are
where in the joint-kernel sum range over pairs of number strings, and real conversion sends to . Also, entry by entry, with , , and the point mass at zero for , otherwise the measure with density . To define , a length- private code has messages, nonnegative summable randomization weights of sum , density kernels , and effects with real nonnegative finite quadratic forms and . Its budget is ; its leakage is , where and is extended nonnegative subtraction. A rate is achievable if, for every and , all sufficiently large positive integers admit such a code with budget at most , leakage at most , , and for every , including zero-weight . The capacity is the supremum of over these rates. The cases and are included.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.