§4.5, joint concavity of Q and convexity of xi
OpenQAlgorithms.Papers.Independent.Intakenext100.P45.jointConcavityFor density pairs and , write and similarly for . Then
The second inequality uses extended nonnegative values, with a zero coefficient contributing zero. These are joint concavity of and joint convexity of its negative logarithm.
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 p.18: §4.5, joint concavity of Q and convexity of xi. Source correction: the cited Lieb premise on PDF p.18 says all real s; it applies here only to 0<=s<=1. The stated conclusion is unchanged; independent moderation required. -/
theorem jointConcavity (r0 r1 s0 s1 : Mat I)
(hr0 : IsDensityMatrix r0) (hr1 : IsDensityMatrix r1)
(hs0 : IsDensityMatrix s0) (hs1 : IsDensityMatrix s1)
(t : ℝ) (ht0 : 0 ≤ t) (ht1 : t ≤ 1) :
chernoffInfimum ((t : ℂ) • r0 + ((1-t : ℝ) : ℂ) • r1)
((t : ℂ) • s0 + ((1-t : ℝ) : ℂ) • s1) ≥
t*chernoffInfimum r0 s0 + (1-t)*chernoffInfimum r1 s1 ∧
chernoffDistance ((t : ℂ) • r0 + ((1-t : ℝ) : ℂ) • r1)
((t : ℂ) • s0 + ((1-t : ℝ) : ℂ) • s1) ≤
(t : EReal)*chernoffDistance r0 s0 + ((1-t : ℝ) : EReal)*chernoffDistance r1 s1 := 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 positive semidefinite complex matrices indexed by , each with trace , and let satisfy . For such matrices define
where powers are defined by continuous functional calculus using total real powers, and for , while for . The distance takes values in the extended real numbers. With and , both inequalities
hold. Singular matrices and are included; the second inequality uses extended-real multiplication and addition, including their conventions for zero coefficients and infinite values.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.