Lemma IV.5: Ie(rho;phi1*⊗phi2*)≤Ie(rho1;phi1*)+H(rho2).
ProvedQAlgorithms.Papers.Independent.Renew46.P43.coherentWeakSubadditivityLet be any normalized state on the two full matrix systems , with marginals , and let be local CPTP operations. Coherent information is output entropy minus entropy exchange of a purification of the same input state.
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
/-- Lemma IV.5, PDF pp.48-50: weak subadditivity. The last term is H(rho2),
not the generally invalid stronger coherent-information term. -/
theorem coherentWeakSubadditivity {I J : Type*} [Fintype I] [Fintype J]
[DecidableEq I] [DecidableEq J]
(rho : Matrix (I × J) (I × J) ℂ) (hrho : IsDensityMatrix rho)
(T : Superop I I) (U : Superop J J) (hT : IsChannel T) (hU : IsChannel U) :
coherentInformation (both T U) rho ≤
coherentInformation T (traceRight rho) + vnEntropy (traceLeft rho) := 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 finite types with decidable equality, let be a positive semidefinite complex matrix indexed by with , and let and be maps on the matrices indexed by and , respectively, each admitting a finite Kraus representation: , , with and . Their joint map is . For a matrix indexed by a finite type , define the vector , where the square root is obtained by continuous functional calculus, and define
where the identity extension applies separately to each block indexed by the reference coordinates, and for , again using continuous functional calculus and the total logarithm with value at . With marginals and , the assertion is
Although empty finite types are not explicitly excluded, the trace-one hypothesis on cannot hold if either or is empty.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.