§III, equations (7)–(8): equivalent max-relative entropy formulas
ProvedQAlgorithms.Papers.Independent.Support50.P21.maxRelative_equivalent_formsAll systems are nonzero finite-dimensional complex Hilbert spaces, and logarithms have base two.
Let satisfy , and let with . With the inverse square root set to zero on the kernel,
There is also a least positive for which
and . The projector includes the zero eigenspace. These are the two equivalent finite formulas for the same divergence.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P21
noncomputable section
namespace QAlgorithms.Papers.Independent.Support50.P21
open QAlgorithms QAlgorithms.Entanglement Filter
open scoped Matrix ComplexConjugate ComplexOrder
attribute [local instance] Classical.propDecidable Classical.decEq
-- Use the accepted declarations directly; no paper-local aliases hide dimension binders.
open QAlgorithms.Papers.Independent.Claimed50.P33
(FiniteState IsAverageLocalMonotone IsAdaptiveLOCCMonotone)
open QAlgorithms.Papers.Independent.Information100.P63 (Mat)
open QAlgorithms.Papers.Independent.Dependency50.P35 (tr elog maxRelative)
open QAlgorithms.Papers.Independent.Information100.P67 (IsSubnormalized)
open QAlgorithms.Papers.Independent.Information100.P07 (support)
open QAlgorithms.Papers.Independent.Information100.P71 (supportPower SupportIncluded)
variable {I J : Type} [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
section Statements
variable [Nonempty I] [Nonempty J]
/-- PDF p.3, (7)–(8): nonzero finite domain, generalized inverse square roots,
and an attained LEAST positive threshold for the weak projector. -/
theorem maxRelative_equivalent_forms (rho sigma : Mat I)
(hrho : IsSubnormalized rho) (hne : 0 < tr rho) (hsigma : sigma.PosSemidef)
(hsupp : SupportIncluded rho sigma) :
maxRelative rho sigma = elog (largestEigenvalue
(supportPower sigma (-(1/2 : ℝ)) * rho * supportPower sigma (-(1/2 : ℝ)))) ∧
∃ t : ℝ, IsLeast (thresholdSet rho sigma) t ∧ maxRelative rho sigma = elog t := by sorry
end Statements
end QAlgorithms.Papers.Independent.Support50.P21
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For any finite index type with decidable equality, let be complex matrices indexed by , with positive semidefinite, , positive semidefinite, and , where . Define
with the infimum taken in the extended real numbers, so an empty feasible set has infimum . Let act by on every positive eigenvalue of and by zero on its kernel. For Hermitian , let be its spectral projection onto all nonnegative eigenvalues, including zero, and set
Then
and there exists a real such that for every and . The positive-trace hypothesis excludes and an empty ; no normalization or upper trace bound is imposed on .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.