§V, unnumbered: entropy potentials and optimal costs (corrected dynamic sign)
ProvedQAlgorithms.Papers.Independent.Gap48.P38.entropyPotentialsLet be a unit vector on nonempty finite complex registers, and use bit entropies of its actual marginals. Define dynamic potentials , and static potentials , . For the numerical corner costs and from equation (2),
These identities express the source's entropy-potential interpretation of the optimal corner expressions.
Formalization Note Page 11 prints for the dynamic difference. The statement corrects this to , as forced by its own definitions and the entropy chain rule. For a Bell pair on with trivial, , , and . Independent moderation of this source correction is required.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P38
noncomputable section
open MeasureTheory
open scoped Matrix ComplexConjugate ComplexOrder Matrix.Norms.Elementwise
attribute [local instance] Classical.propDecidable Classical.decEq
namespace QAlgorithms.Papers.Independent.Gap48.P38
/-- Section V, PDF p.11: entropy potentials and the corner costs (2).
SOURCE CORRECTION: the printed K_final-K_initial has the opposite sign to
its own definitions; changed to K_initial-K_final, forced by the CMI chain
rule and the Bell(C,D) example. Static potential uses S_final-S_initial as printed.
Independent moderation of this correction is still required. -/
theorem entropyPotentials
(I : Fin 4 → Type) [∀ i, Fintype (I i)] [∀ i, Nonempty (I i)]
(psi : EuclideanSpace ℂ (∀ i, I i)) (hpsi : ‖psi‖ = 1) :
let rho := pureDensity psi
let Kinitial := registerMutualInfo I rho {0,2} {3} / 2
let Kfinal := registerMutualInfo I rho {0} {3} / 2
let Sinitial := registerMutualInfo I rho {0,2} {1} / 2
let Sfinal := registerMutualInfo I rho {0} {1,2} / 2
Kinitial - Kfinal = registerCMI I rho {2} {3} {0} / 2 ∧
Kinitial - Kfinal = cornerQ I rho ∧ Sfinal - Sinitial = cornerE I rho := by sorry
end QAlgorithms.Papers.Independent.Gap48.P38
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every family of four finite, nonempty index sets , allowing singleton sets, and every complex vector indexed by with , put . Define the reduction by summing over assignments to the complementary registers, and write , where is applied by continuous functional calculus with ; digit strings denote register sets. Set
Then all three equalities hold:
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.