§4.8, quantum minimizing condition (38)–(40)
OpenQAlgorithms.Papers.Independent.Intakenext100.P45.quantumHellingerFor invertible finite-dimensional densities , let and . Then
A minimizer exists on ; a parameter minimizes exactly when the two traces agree, equivalently when . Here can be non-self-adjoint; its logarithm is transported by similarity from the positive definite sandwich, and . No equality of these values with the Chernoff distance is asserted.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P45
noncomputable section
namespace QAlgorithms.Papers.Independent.Intakenext100.P45
open QAlgorithms Filter
open scoped Matrix ComplexConjugate ComplexOrder ENNReal BigOperators Topology
open QAlgorithms.Papers.Independent.Information100.P07
(support positivePart positiveProjection IsProjection)
open QAlgorithms.Papers.Independent.Information100.P56 (IsLaw)
open QAlgorithms.Papers.Independent.Information100.P60 (Mat copies)
open QAlgorithms.Papers.Independent.Dependency50.P28 (logNonnegativeExtended relativeEntropy)
open QAlgorithms.Papers.Independent.Claimed50.P15 (positiveMatrixRpow)
attribute [local instance] Classical.propDecidable Classical.decEq
variable {I J : Type} [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
/-- arXiv:0708.4282v1, PDF pp.19,20: §4.8, quantum minimizing condition (38)–(40). -/
theorem quantumHellinger (rho sigma : Mat I)
(hrho : IsDensityMatrix rho) (hsigma : IsDensityMatrix sigma)
(hir : IsUnit rho) (his : IsUnit sigma) :
(∀ s ∈ Set.Icc (0 : ℝ) 1,
deriv (chernoffOverlap rho sigma) s =
-(positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log rho).trace.re +
(positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log sigma).trace.re ∧
((∀ t ∈ Set.Icc (0 : ℝ) 1, chernoffOverlap rho sigma s ≤ chernoffOverlap rho sigma t) ↔
(positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log rho).trace =
(positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log sigma).trace) ∧
((positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log rho).trace =
(positiveMatrixRpow rho (1-s)*positiveMatrixRpow sigma s*cfc Real.log sigma).trace ↔
arcRelative rho sigma rho s = arcRelative rho sigma sigma s)) ∧
(∃ s ∈ Set.Icc (0 : ℝ) 1,
(∀ t ∈ Set.Icc (0 : ℝ) 1, chernoffOverlap rho sigma s ≤ chernoffOverlap rho sigma t) ∧
arcRelative rho sigma rho s = arcRelative rho sigma sigma s) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P45
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every finite index type with decidable equality, let be invertible positive semidefinite complex matrices indexed by , each with trace . Define matrix powers and matrix logarithms by continuous functional calculus, and put
and . Then, for every , the ordinary derivative of the function , also at either endpoint, satisfies
For each such , for every if and only if the complex traces and are equal; this complex-trace equality holds if and only if . Finally, there exists minimizing on this interval and satisfying . The hypotheses exclude singular density matrices, and no uniqueness of is asserted.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.