Concrete growing-dimension witnesses
OpenQAlgorithms.Papers.Independent.Intakenext100.P70.concreteParametersLet and be integers, and set
Construct from the normalized Choi state (S18) and from (S15). The parameters are admissible and satisfy (S19), is PPT-binding, and is completely positive and trace preserving. Moreover , , and
More precisely, there is a perfect private bit on the first-share registers for which, writing for the computed marginal (S20), and , one has , and
Here is binary entropy in bits and is the full trace norm. The scalar bounds are consequences of the specified choices, not additional assumptions.
These parameters make the arbitrary-block-length phenomenon concrete. The source's intermediate logarithmic sufficiency inference is corrected; the final values of and are preserved.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P70
namespace QAlgorithms.Papers.Independent.Intakenext100.P70
noncomputable section
open scoped Matrix ComplexConjugate ComplexOrder
attribute [local instance] Classical.propDecidable Classical.decEq
open QAlgorithms.Papers.Independent.Information100.P22
(Op coherentInformation singleUseInformation quantumCapacity mixed)
/-- Unnumbered concrete parameter choice, PDF p.11, using S46--S54 on p.10.
SOURCE CORRECTION: the paragraph's upper bound on log(1+x) alone does not justify a
sufficient N. The final N=2*n*4^n is retained: log(1+x)>=x/(1+x) verifies it.
The ceiling in r is retained from Lemma 2. The raw bound below retains the factor N+1. -/
theorem concreteParameters (n m : ℕ) (hn : 0 < n) (hm : 68 ≤ m) :
let P := cubittExplicitParameters n m
let p := cubittErasureThreshold n (1 / 4)
CubittValid P ∧ CubittPPTCondition P ∧ CubittPPTBinding (cubittGammaChannel P) ∧
IsChannel (cubittPaperChannel P (1 / 4) p) ∧
n < P.N + 1 ∧ p ^ P.N < (1 / 3 : ℝ) ∧
cubittBlockInformation (cubittPaperChannel P (1 / 4) p) n = 0 ∧
0 < cubittBlockInformation (cubittPaperChannel P (1 / 4) p) (P.N + 1) ∧
0 < quantumCapacity (cubittPaperChannel P (1 / 4) p) ∧
∃ hN : 0 < P.N,
∃ rho : Op ((Bool × CubittShare P.r P.m P.d) × (Bool × CubittShare P.r P.m P.d)),
CubittPerfectPbit rho ∧
let tau := traceNorm (cubittFirstShareState P hN - rho)
0 ≤ tau ∧ tau < 1 ∧ tau ≤ cubittApproximationBound m ∧
cubittContinuityLoss tau ≤ 72 * (2 : ℝ) ^ (-(m : ℝ) / 4) * (m : ℝ) ^ ((3 : ℝ) / 2) ∧
cubittContinuityLoss tau < 1 / 3 ∧
(1 - (1 / 4 : ℝ)) * (1 - p ^ P.N - cubittContinuityLoss tau) - 1 / 4 ≤
((P.N + 1 : ℕ) : ℝ) * cubittBlockInformation (cubittPaperChannel P (1 / 4) p) (P.N + 1) := by sorry
end
end QAlgorithms.Papers.Independent.Intakenext100.P70
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For all natural numbers with and , set
so that . Define , , , and . Put
and, for , put
Let be the matrix on with entries
where the factors use , and define
For a real , let have false diagonal block , true diagonal block , and zero off-diagonal blocks. Let , on input , discard off-diagonal switch blocks, apply to the true block and to the false block, and retain the switch label in its output . Define
Here square roots and entropy use continuous functional calculus, the identity tensor map applies to each second-register block, and tensor powers apply independently in each coordinate, as determined by expansion in matrix units. Then all of the following hold:
is a channel and the output partial transpose of its normalized Choi matrix is positive semidefinite, meaning that the matrix with entries at row , column , is positive semidefinite; and is a channel, where being a channel means having a finite Kraus representation with . Moreover,
There also exist a witness of and a matrix on having the following properties. The matrix is positive semidefinite with trace one, and there exist a positive semidefinite, trace-one matrix on and unitaries on , indexed by every pair of Booleans, such that
To define the matrix compared with , let insert at index and agree with an array on all other indices, and set
For
where , with the zero terms interpreted as zero, the same satisfies
and
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.