Remark A.15: For measurable cq phi*, input P and CPTP psi*, I(P;psi*∘phi*)≤I(P;phi*), with I(P;phi*)=H(Pphi*)−H(phi*|P).
OpenQAlgorithms.Papers.Independent.Renew46.P43.cqProcessingLet be a measurable map from an arbitrary measurable input space into the normalized states of a finite intrinsic system, let be any input probability measure, and let be a CPTP output operation. The average and entropy averages are Bochner integrals of bounded measurable state functions. Define cq information by the mean divergence .
This is the source result with its stated operational or information-theoretic domain.
import Definitions.Def_QAlgorithms_Papers_Independent_Renew46_P43
noncomputable section
namespace QAlgorithms.Papers.Independent.Renew46.P43
open scoped BigOperators Matrix ComplexConjugate ComplexOrder ENNReal Matrix.Norms.L2Operator
open Filter MeasureTheory
open QAlgorithms.Papers.Independent.Information100.P08 (IsPure)
open QAlgorithms.Papers.Independent.Information100.P06 (Below)
open QAlgorithms.Papers.Independent.Information100.P60 (copies relativeBits)
open QAlgorithms.Papers.Independent.Dependency50.P19 (count both)
open QAlgorithms.Papers.Independent.Information100.P22 (coherentInformation)
attribute [local instance] Classical.propDecidable Classical.decEq
open QAlgorithms.Papers.Independent.Claimed50.P40 (InBlocks)
attribute [local instance] QAlgorithms.Papers.Independent.Renew46.P43.matrixMeasurableSpace
/-- Remark A.15, PDF p.73: the entropy-difference identity and genuine channel
processing for arbitrary measurable input laws. Finite-dimensional normalized state
integrands are bounded, so their Bochner integrals are not totalized junk values. -/
theorem cqProcessing {X : Type*} [MeasurableSpace X] {Y Z : IntrinsicSystem}
(mu : Measure X) [IsProbabilityMeasure mu]
(W : X → Matrix (Fin Y.dim) (Fin Y.dim) ℂ) (hW : Measurable W)
(hstates : ∀ x, Y.IsState (W x)) (T : IntrinsicOperation Z Y) :
cqInformation mu W = ENNReal.ofReal
(vnEntropy (cqAverage mu W) - ∫ x, vnEntropy (W x) ∂mu) ∧
cqInformation mu (fun x => intrinsicRestriction Z T.map (W x)) ≤ cqInformation mu W := by sorry
end QAlgorithms.Papers.Independent.Renew46.P43
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be any measurable space, a probability measure on , and systems with positive integer dimensions and block-label functions . Let be measurable, with every positive semidefinite, of trace , and zero between unequal -block labels. Let be complex-linear, send identity to identity, and send matrices supported within -blocks to matrices supported within -blocks. Assume also that, for every integer , its amplification sends positive semidefinite matrices satisfying whenever to positive semidefinite matrices. Define
where is a matrix unit and the matrix integrals are Bochner integrals. For matrices , define to be if either matrix is not positive semidefinite of trace ; if both are such matrices, define it to be when , and otherwise
with the scalar functions applied by continuous functional calculus and the total convention . Write , where . Then both
and
hold, with the nonnegative extended-real integrals understood as lower Lebesgue integrals.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.