Sections 2.7 and 7.3 / (108): finite-error capacities approach Q
OpenQAlgorithms.Papers.Independent.Gap20.P07.finiteErrorCapacity_continuityLet be any memoryless CPTP channel between nonzero complex Hilbert spaces. Define exactly as Definition 2.1, but require eventual optimal full-cb error at most . For , where ,
Capacities and scalar multiplication are extended nonnegative. The strict positivity of avoids the distinct zero-error condition.
import Definitions.Def_QAlgorithms_Papers_Independent_Gap20_P07
noncomputable section
open Filter MeasureTheory
open scoped Topology ENNReal NNReal Matrix ComplexOrder ComplexConjugate
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
attribute [local instance] IsStarNormal.instContinuousFunctionalCalculus
namespace QAlgorithms.Papers.Independent.Gap20.P07
open QAlgorithms
open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33
(State IsTraceClass IsTraceClassChannel traceSize rankOne tensorApply)
open QAlgorithms.Papers.Independent.Frontier30.P14
(Operation isometryCoeff conjugateKernel)
open QAlgorithms.Papers.Independent.Claimed50.P35
(stabilizedTraceClassNorm traceClassTensorMap traceClassChannelPower
transpose transposeBound TraceLinear tracePair MeasurePrepare)
open QAlgorithms.Papers.Independent.Gap48.P29 (fe FiniteCPMap)
/-- PDF pp.6,29, (108): full-cb finite-error capacities converge from the right;
positive epsilon and positive scalar factor exclude the zero-error endpoint. -/
theorem finiteErrorCapacity_continuity {I J : Type} [Nonempty I] [Nonempty J]
(T : Operation I J) (hT : IsTraceClassChannel T) :
(∀ epsilon : ℝ, 0 < epsilon → epsilon < 1 / (4 * Real.exp 1) →
qubitReferenceCapacity T ≤ finiteErrorCapacity T epsilon ∧
finiteErrorCapacity T epsilon * ENNReal.ofReal (1 - 4 * Real.exp 1 * epsilon) ≤
qubitReferenceCapacity T) ∧
Tendsto (finiteErrorCapacity T) (𝓝[>] (0 : ℝ)) (𝓝 (qubitReferenceCapacity T)) := by sorry
end QAlgorithms.Papers.Independent.Gap20.P07
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be nonempty types and . The following definitions specify the quantities and hypothesis. For any index type , put . A matrix is trace class if it has an expansion , with and ; write for the infimum of these sums over all such expansions, with infimum when none exists. A density is a trace-class matrix whose finite principal restrictions are positive semidefinite, whose real diagonal is summable, and whose real diagonal sum is . For a matrix map , define its ancillary extension by , where , and define
The hypothesis on says that it preserves trace-class matrices, is additive and complex-homogeneous on them, and that every finite ancillary extension preserves densities. Call any map satisfying these requirements a channel. Let mean the coordinatewise -fold power: the zeroth power evaluates a matrix at the unique pair of empty tuples, and the next power applies to the first-coordinate block and the preceding power to the remaining coordinates. For positive integers , let be the infimum of over channels from matrices on to matrices on and channels from matrices on to matrices on . Define as the supremum in of all real such that, for every pair of positive-integer sequences with and , one has . Define by replacing that convergence requirement with the requirement that eventually . These suprema are if there are no eligible positive rates; they may be infinite, and the error infima need not be attained. Then, for every real ,
and as through positive real values, with convergence in the extended nonnegative reals. Neither endpoint is included in the displayed bounds.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.