Theorem 7: strong purity and Holevo additivity for general Hadamard channels
OpenQAlgorithms.Papers.Independent.Gap48.P36.hadamard_strong_additivityAll Hilbert spaces have positive finite dimension. Channels are completely positive and trace preserving; conjugates use the same normalized pure-environment dilation. Write and , with density inputs, Schatten norm, and von Neumann entropy. Let have normalized Kraus operators , where the are an orthonormal output basis and the are independent of . For every channel ,
Here is the supremum of the single-use Holevo quantity over finite input ensembles. The need not be orthogonal, and the output dimension may exceed the input dimension.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P36
noncomputable section
namespace QAlgorithms.Papers.Independent.Gap48.P36
open scoped BigOperators Matrix Kronecker ComplexConjugate ComplexOrder ENNReal
open QAlgorithms.Papers.Independent.Claimed50.P48
open QAlgorithms.Papers.Independent.Information100.P67 (receiverChannel complementaryChannel)
open QAlgorithms.Papers.Independent.Information100.P56 (entropy normalizedChoi)
open QAlgorithms.Papers.Independent.Information100.P50 (schattenNorm)
open QAlgorithms.Papers.Independent.Information100.P03 (convexRoof)
open QAlgorithms.Papers.Independent.Dependency50.P45 (holevoCapacity)
attribute [local instance] Classical.propDecidable Classical.decEq
variable {A B C D E K L A2 B2 C2 : Type*}
variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]
variable [Fintype C] [DecidableEq C] [Fintype D] [DecidableEq D]
variable [Fintype E] [DecidableEq E] [Fintype K] [DecidableEq K]
variable [Fintype L] [DecidableEq L] [Fintype A2] [DecidableEq A2]
variable [Fintype B2] [DecidableEq B2] [Fintype C2] [DecidableEq C2]
/-- Theorem 7, PDF pp.11–13: arbitrary common-vector Kraus maps, with all three conclusions.
The partner and finite exponent are unrestricted in the source's domains. -/
theorem hadamard_strong_additivity [Nonempty A] [Nonempty B] [Nonempty K] [Nonempty A2] [Nonempty B2]
(v : Matrix K B ℂ) (w : K → EuclideanSpace ℂ A)
(hF : IsKrausFamily (gramKraus v w)) (Psi : Superop A2 B2) (hPsi : IsChannel Psi) :
(∀ p : ℝ, 1 ≤ p → Multiplicative (ENNReal.ofReal p) (krausApply (gramKraus v w)) Psi) ∧
Additive (krausApply (gramKraus v w)) Psi ∧
holevoCapacity (superopTensor (krausApply (gramKraus v w)) Psi) =
holevoCapacity (krausApply (gramKraus v w)) + holevoCapacity Psi := by sorry
end QAlgorithms.Papers.Independent.Gap48.P36
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be finite nonempty types with decidable equality, let be a complex matrix with rows indexed by and columns indexed by , and let be arbitrary vectors. Define matrices from to , indexed by , by , and assume . Set , and let be any channel from matrices on to matrices on , meaning it has a finite Kraus representation with the same normalization condition. Then
and
Here , with ; is the infimum of over density matrices, using natural logarithms; and is the supremum, over all finite ensembles of density matrices with nonnegative real weights summing to , of , where . All matrix functions use continuous functional calculus, and .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.