Lemma 1
OpenQAlgorithms.Papers.Independent.Gap48.P18.chi_finite_mixture_and_binary_boundLet be Borel probability measures on the trace-norm state space of a nonzero separable complex Hilbert space, with barycenters , and let sum to one. With natural logarithms and extended nonnegative integrals,
For two such measures and ,
Zero weights contribute zero, including at infinite entropy. This identity separates within-ensemble and between-ensemble information.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P18
noncomputable section
open MeasureTheory Filter
open scoped ENNReal NNReal Topology Matrix.Norms.L2Operator
open QAlgorithms.Papers.Independent.Dependency50.P33
(State PureState traceSize traceClosure IsTraceClassChannel productCoeff)
open QAlgorithms.Papers.Independent.Frontier18.P07
(GeneralizedEnsemble HasFiniteSupport HasAverage averageCoefficients
extendedEntropyNats extendedRelativeEntropyNats
finiteEnsembleCapacity pureStates)
open QAlgorithms.Papers.Independent.Frontier18.P01 (traceFirst traceSecond)
namespace QAlgorithms.Papers.Independent.Gap48.P18
attribute [local instance] Classical.propDecidable Classical.decEq
/-- Lemma 1, arXiv:quant-ph/0510073v2, PDF pp.6.
NNReal weights with t≤1 encode exactly [0,1]; finite-mixture measures are constructed explicitly. -/
theorem chi_finite_mixture_and_binary_bound {I : Type} [Countable I] [Nonempty I] :
(∀ (m : ℕ) (w : Fin m → ℝ≥0) (mu : Fin m → GeneralizedEnsemble I),
(∑ k, w k) = 1 →
measureHolevo (∑ k, (w k : ℝ≥0∞) • (mu k : Measure (State I))) =
(∑ k, (w k : ℝ≥0∞) * measureHolevo (mu k : Measure (State I))) +
measureHolevo (∑ k, (w k : ℝ≥0∞) • Measure.dirac (probabilityAverage (mu k)))) ∧
(∀ (mu nu : GeneralizedEnsemble I) (t : ℝ≥0), t ≤ 1 →
(t : ℝ≥0∞) * measureHolevo (mu : Measure (State I)) +
((1 - t : ℝ≥0) : ℝ≥0∞) * measureHolevo (nu : Measure (State I)) +
((t * (1 - t) / 2 : ℝ≥0) : ℝ≥0∞) *
distanceSquared (probabilityAverage nu) (probabilityAverage mu) ≤
measureHolevo ((t : ℝ≥0∞) • (mu : Measure (State I)) +
((1 - t : ℝ≥0) : ℝ≥0∞) • (nu : Measure (State I)))) := by sorry
end QAlgorithms.Papers.Independent.Gap48.P18
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every countable, nonempty index type , both of the following assertions hold. Write for the states: complex matrices admitting with and , having real nonnegative finite-coordinate quadratic forms, and having summable real diagonal entries of sum . For any matrix , let be the infimum of over those expansions, with empty infimum ; give the topology generated by for states and , and its Borel measurable structure. For finite matrices set when , and otherwise, with logarithms taken by continuous functional calculus and ; put . For any measure on , let and . For a Borel probability measure , let be its state average selected by a fixed chosen continuous barycenter map: every coordinate is integrable and ; continuity uses the probability-measure topology and the stated state topology. First, for every natural number , nonnegative real weights , and Borel probability measures on , if , then
where is the Dirac measure at . Second, for every pair of Borel probability measures and every nonnegative real ,
All products and sums in these two assertions are in , with , and no entropy or -finiteness assumption is imposed. The weight condition is impossible when ; and are included, and the nonnegative-real subtraction agrees with ordinary subtraction under . Coordinate integrals in the definition for arbitrary measures are total integrals, assigned zero when nonintegrable, and the finite-compression supremum includes the empty compression.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.