Remark II.6: average monotonicity under public local instruments
OpenQAlgorithms.Papers.Independent.Intakenext100.P47.publicInstruments_average_monotonicityLet be a density matrix on arbitrary finite nonzero complex spaces. Consider any finite adaptive protocol in which, at each node, either Alice or Bob applies a finite rectangular quantum instrument and publicly announces its outcome. Arbitrary finite local dimensions, private Kraus multiplicities, and choices depending on earlier messages are allowed. Write for the probability of a full transcript and for its normalized final state when . Then
For one instrument this is , with the analogous statement when Bob acts. Zero branches contribute zero without normalization. Thus the theorem retains both the one-step inequality and its finite iteration, counting correlations conditional on the public transcript.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P47
noncomputable section
namespace QAlgorithms.Papers.Independent.Intakenext100.P47
open QAlgorithms Filter
open QAlgorithms.Papers.Independent.Gap48.P20 (Copies localProduct)
open QAlgorithms.Papers.Independent.Dependency50.P19 (mutualInformation first)
open QAlgorithms.Papers.Independent.Dependency50.P45 (entropyExchange)
open QAlgorithms.Papers.Independent.Information100.P27 (shannon Config)
open QAlgorithms.Papers.Independent.Information100.P71 (channelPower)
open QAlgorithms.Papers.Independent.Advance50.P05 (RectangularInstrument)
open scoped Matrix ComplexConjugate
attribute [local instance] Classical.propDecidable Classical.decEq
/-- PDF pp.5--6, Remark II.6: public local instruments decrease total correlation on
average, including every finite iteration. A one-node tree is the inequality (13). -/
theorem publicInstruments_average_monotonicity (a b : ℕ) (ha : 0 < a) (hb : 0 < b)
(rho : Matrix (Fin a × Fin b) (Fin a × Fin b) ℂ) (hrho : IsDensityMatrix rho)
(P : PublicProtocol a b) : P.average rho ≤ mutualInformation rho := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P47
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For all positive natural dimensions , every positive semidefinite trace-one complex matrix on , and every finite protocol tree built from stop nodes and local rectangular instruments with outcome-dependent continuations, the recursively defined value is at most
Here , using continuous functional calculus. At each instrument node, its finitely many Kraus matrices have a common positive local output dimension and satisfy ; the branch matrix is for Alice, or the analogous expression with for Bob. The recursive value sums over branches with , assigns zero to the other branches, and uses at every stop node. The assertion includes the protocol that stops immediately and places no upper bound on the tree depth or number of outcomes.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.