Remark 2, (14)–(17): alternative error criteria
OpenQAlgorithms.Papers.Independent.Intakenext100.P20.alternative_error_criteriaLet be an arbitrary discrete label set and let be a nonzero finite-dimensional complex Hilbert space. The source consists of normalized pure states and probabilities with ; write . Each memory space is finite and nonzero. Blind encoders and decoders are CPTP; a visible encoder maps labels to normalized density operators. A blind encoder is fixed before its input label.
For positive block length , the word probabilities and states are the independent products of the source probabilities and states. Admissible code sequences satisfy . An error exponent is the supremum over admissible sequences of ; a success exponent is the infimum of . Logarithms are natural, and .
For either coding model define , interpreting in the blind case. Subscript means the optimized exponent formed with this loss. Then
Thus the square-root-overlap criterion preserves the optimal error exponent and gives the stated comparisons for success exponents.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P20
noncomputable section
open scoped Matrix ENNReal ComplexOrder Topology
namespace QAlgorithms.Papers.Independent.Intakenext100.P20
open QAlgorithms.Papers.Independent.Information100.P08 (IsPure)
open QAlgorithms.Papers.Independent.Information100.P47 (IsProbability)
open QAlgorithms.Papers.Independent.Information100.P53 (SpectralData)
open QAlgorithms.Papers.Independent.Gap48.P19 (shannonNats)
open QAlgorithms.Papers.Independent.Dependency50.P28 (classicalD)
attribute [local instance] Classical.propDecidable
/-- Remark 2, PDF p.3, (14)-(17), Appendix C pp.10-11: unchanged error
exponents and the stated one-sided success-exponent comparisons.
Printed blind-loss input: F(rho_i); changed to E(rho_i), because Section II
defines E as the blind quantum encoder and Appendix C uses that model. -/
theorem alternative_error_criteria {L : Type*} {d : ℕ} (S : DiscretePureSource L d) (R : ℝ) :
blindExponent S R true false = blindExponent S R false false ∧
visibleExponent S R true false = visibleExponent S R false false ∧
blindExponent S R true true ≤ blindExponent S R false true ∧
visibleExponent S R true true ≤ visibleExponent S R false true := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P20
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be any type, a natural number, and consist of , nonnegative summable weights with , and positive semidefinite trace-one matrices for unit vectors ; no finiteness or countability of is assumed. Let . For each length , source words have weights and states . Visible codes assign each word a density matrix in a positive finite memory dimension and then apply a channel decoder; blind codes apply a channel encoder to followed by a channel decoder. Each channel is required to have a finite Kraus representation with . For the reconstructed states , define two losses
For each code kind and each , restrict to sequences with positive memory dimensions satisfying . Write for the supremum of , and for the infimum of , where and otherwise, with natural logarithms. Then
The rate is unrestricted, and all extrema are in the extended reals, with empty supremum and empty infimum .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.