Supplementary Note 2, Lemma 2: entropy bounds measured mutual information
OpenQAlgorithms.Papers.Independent.Intakenext100.P87.lemma_2Let be a normalized pure state on arbitrary finite-dimensional registers. Let and be arbitrary local generalized measurements, with arbitrary measurable outcome spaces. Let denote their classical Shannon mutual information in bits, defined as the supremum over finite measurable coarse grainings. If and are the reduced density operators, then
This bounds measured classical information by either marginal entropy. The supremum takes values in the extended nonnegative reals; the bound proves it finite.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P87
noncomputable section
namespace QAlgorithms.Papers.Independent.Intakenext100.P87
open scoped Matrix ComplexConjugate
open QAlgorithms.Measurements
/-- PDF p.19, Supplementary Note 2, Lemma 2: each marginal entropy bounds the
classical mutual information of arbitrary local generalized measurements on a pure state. -/
theorem lemma_2 {A B X Y : Type*} [Fintype A] [DecidableEq A]
[Fintype B] [DecidableEq B] [MeasurableSpace X] [MeasurableSpace Y]
(psi : EuclideanSpace ℂ (A × B)) (hpsi : IsState psi)
(P : OperatorPOVM A X) (Q : OperatorPOVM B Y) :
measuredInfo (pureDensity psi) P Q ≤ ENNReal.ofReal (vnEntropy (traceRight (pureDensity psi))) ∧
measuredInfo (pureDensity psi) P Q ≤ ENNReal.ofReal (vnEntropy (traceLeft (pureDensity psi))) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P87
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 measurable spaces, and let have Euclidean norm . Let and be operator-valued measures on and , acting on and , respectively: each assigns zero to the empty set and the identity to the whole space, assigns positive semidefinite matrices to measurable sets, and is countably additive on pairwise disjoint measurable sets, with convergence in the matrix space. Write , , and . For every pair of natural numbers and measurable maps , , define and . The supremum of , taken in over all these choices, is at most and also at most , where for , interpreted by continuous functional calculus. Zero denominators and logarithms at zero use the total conventions and . The partition sizes include zero whenever the indicated maps exist; the unit-norm hypothesis excludes empty or .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.