Lemma 2: product fidelity from a near-pure marginal
ProvedQAlgorithms.Papers.Independent.Support50.P02.product_fidelityLet be finite, positive-dimensional complex registers, a unit vector, a density operator, and a density operator. Define its marginals by and . Let satisfy , where is squared Uhlmann fidelity. Then
The trace norm is not halved. This relates a near-pure marginal to a joint product target. No positive-overlap assumption is imposed; the zero-overlap case is included.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P02
noncomputable section
open scoped BigOperators Matrix ComplexConjugate
namespace QAlgorithms.Papers.Independent.Support50.P02
open QAlgorithms.Papers.Independent.Information100.P67
(squaredFidelity receiverChannel complementaryChannel IsDegradable)
open QAlgorithms.Papers.Independent.Information100.P21 (Op)
open QAlgorithms.Papers.Independent.Information100.P03 (PureEnsemble)
open QAlgorithms.Papers.Independent.Dependency50.P19 (epr)
attribute [local instance] Classical.propDecidable Classical.decEq
variable {A B C E R S I J : Type*}
variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]
variable [Fintype C] [DecidableEq C] [Fintype E] [DecidableEq E]
variable [Fintype R] [DecidableEq R] [Fintype S] [DecidableEq S]
variable [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
/-- Lemma 2, PDF pp.10–11: both marginals are computed from Omega.
No f>0 hypothesis is imposed: the zero-overlap case is part of the theorem.
The proof's claim of general fidelity linearity is not used as a definition. -/
theorem product_fidelity [Nonempty A] [Nonempty B]
(phi : EuclideanSpace ℂ A) (hp : IsState phi)
(rho : Op B) (hr : IsDensityMatrix rho) (Omega : Op (A × B))
(hO : IsDensityMatrix Omega) (epsilon : ℝ)
(hf : 1 - epsilon ≤ squaredFidelity (pureDensity phi) (traceRight Omega)) :
1 - traceNorm (rho - traceLeft Omega) - 3 * epsilon ≤
squaredFidelity (Matrix.kronecker (pureDensity phi) rho) Omega := by sorry
end QAlgorithms.Papers.Independent.Support50.P02
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every pair of nonempty finite index types with decidable equality, every norm-one vector , every positive semidefinite matrix on of trace , every positive semidefinite matrix on of trace , and every real number , define ,
with square roots given by continuous functional calculus. If , then
The partial traces mean and , and the tensor-product matrix has entries . No separate sign or range restriction on is assumed.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.