Lemma 5: coherent implementation of a decoding measurement
ProvedQAlgorithms.Papers.Independent.Gap48.P38.lemma5_coherentDecoderLet be a unit vector on nonempty finite complex registers. Let , let be unitaries on , put , and let be any unit vectors. Let be a POVM on . Define the average measurement error and squared pure-state fidelity by
There are complex phases of modulus one such that the explicitly constructed map
is an isometry from to . Each amplitude can be taken real and nonnegative, and
This turns a decoding measurement into coherent communication while retaining the reference . The real, nonnegative amplitude convention is the one explicitly chosen in the appendix proof.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P38
noncomputable section
open MeasureTheory
open scoped Matrix ComplexConjugate ComplexOrder Matrix.Norms.Elementwise
attribute [local instance] Classical.propDecidable Classical.decEq
namespace QAlgorithms.Papers.Independent.Gap48.P38
/-- Lemma 5 (coherent implementation of a decoding measurement), PDF p.7,
proof p.12. The phases make each overlap real and nonnegative, interpreting the
printed complex inequality exactly as its proof does; L is constructed, not assumed. -/
theorem lemma5_coherentDecoder
(X Y : Type) [Fintype X] [Fintype Y] [DecidableEq X] [DecidableEq Y]
[Nonempty X] [Nonempty Y]
(k : ℕ) (hk : 0 < k) (psi : EuclideanSpace ℂ (X × Y)) (hpsi : ‖psi‖ = 1)
(U : Fin k → Matrix.unitaryGroup X ℂ)
(psiPrime : Fin k → EuclideanSpace ℂ (X × Y))
(hpsiPrime : ∀ j, ‖psiPrime j‖ = 1)
(Lambda : Fin k → Matrix X X ℂ) (hLambda : IsPOVM Lambda) :
∃ alpha : Fin k → ℂ, (∀ j, ‖alpha j‖ = 1) ∧
let L := coherentDecoder alpha U Lambda
Lᴴ * L = 1 ∧
(∀ j, (decoderOverlap psi L (psiPrime j) j).im = 0 ∧
0 ≤ (decoderOverlap psi L (psiPrime j) j).re) ∧
(k : ℝ)⁻¹ * ∑ j, (decoderOverlap psi L (psiPrime j) j).re ≥
1 - 2 * (coherentMeasurementError psi U Lambda +
Real.sqrt (1 - decodingFidelity psi U psiPrime)) := by sorry
end QAlgorithms.Papers.Independent.Gap48.P38
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For finite, nonempty types with decidable equality, a positive natural number , a unit vector , unitary matrices on , unit vectors , and positive semidefinite matrices on satisfying , there exist complex numbers with such that the matrix
satisfies . For these same coefficients, define
Every has zero imaginary part and nonnegative real part, and
The matrix square roots are defined by real continuous functional calculus, and the inner product is conjugate-linear in its first argument.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.