Definitions for General teleportation channel, singlet fraction and quasi-distillation
DefinitionQAlgorithms_Papers_Independent_Intakenext100_P18The objects the formalized statements of General teleportation channel, singlet fraction and quasi-distillation (arXiv:quant-ph/9807091) are stated in. Each was checked against the paper by a blind read-back and a moderator.
choiInverse: PDF pp.3--4, (8)--(9): computed inverse on matrix units. Equivalentlyd Tr_reference ((X^T tensor I) rho); the transpose belongs to the input X.noisySinglet: PDF p.5, (13): the original convex-mixture noisy-singlet family. Statements about this family impose 0 <= p <= 1.retentionChannel: PDF p.6, (18): source retention p. The frozen map uses noise weight, so it is called with 1-p. It includes Tr(X) on arbitrary operators.stateTwirl: PDF p.5, (17): U tensor conjugate U, averaged with probability Haar law. The second factor is entrywise conjugation, not adjoint.standardTeleport: PDF pp.4--6, (11)--(12),(20): actual complete Bell measurement followed by the outcome-conditioned correction. All d^2 branches are retained.basisClassical: PDF p.8: explicit classical achiever, measuring in the computational basis and preparing the basis state reported by Alice.IsLocalFilter: PDF pp.9--10: a local successful filter is a contraction. Its failure can be completed by sqrt(I-A†A), so it is a physical two-outcome operation.ExactlyDistillable: PDF p.9, (33): exact noncollective m-by-m distillation. Rectangular local maps include output-frame changes. Rescaling unrestricted A,B to contractions preserves the conditional state and positive success.IsRankProjection: PDF p.10: orthogonal rank-m projection on a local space.coordinateProjection: PDF p.10: Bob's projection onto his first N coordinates.rectangularPlus: PDF p.10: N-dimensional plus vector embedded in N tensor M.rectangularResource: PDF p.10: convex resource family for exact rectangular extraction.
import Mathlib
import Definitions.Def_QAlgorithms_Defs_QuditChannelFidelities
import Definitions.Def_QAlgorithms_Defs_BipartiteSeparability
import Definitions.Def_QAlgorithms_Papers_Independent_Claimed50_P28
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P11
import Definitions.Def_QAlgorithms_Papers_Independent_Renew46_P40
/-!
Statements from Horodecki, Horodecki and Horodecki, quant-ph/9807091v2.
All page citations below are PDF pages. Compilation is pending the batch handoff.
The general adaptive-LQCC nodes are explicitly deferred in the unit's HARD.md;
no separable-operation or fixed Bell-protocol surrogate is used for them.
-/
noncomputable section
open QAlgorithms MeasureTheory
open scoped Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator
namespace QAlgorithms.Papers.Independent.Intakenext100.P18
-- Pinned accepted imports, allowed by the independent-paper override.
open QAlgorithms.Papers.Independent.Claimed50.P28 (maximallyEntangledDensity)
open QAlgorithms.Papers.Independent.Support50.P11 (choiState teleportWeyl bellKet)
open QAlgorithms.Papers.Independent.Renew46.P40 (initialJoint)
/-- PDF pp.3--4, (8)--(9): computed inverse on matrix units. Equivalently
`d Tr_reference ((X^T tensor I) rho)`; the transpose belongs to the input X. -/
def choiInverse (d : ℕ)
(rho : Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ) : Superop (Fin d) (Fin d) :=
fun X a b => (d : ℂ) * ∑ i, ∑ j, X i j * rho (i, a) (j, b)
/-- PDF p.5, (13): the original convex-mixture noisy-singlet family.
Statements about this family impose 0 <= p <= 1. -/
def noisySinglet (d : ℕ) (hd : 0 < d) (p : ℝ) :
Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ :=
(p : ℂ) • maximallyEntangledDensity d + (((1 - p) / (d : ℝ)^2 : ℝ) : ℂ) • 1
/-- PDF p.6, (18): source retention p. The frozen map uses noise weight,
so it is called with 1-p. It includes Tr(X) on arbitrary operators. -/
def retentionChannel (d : ℕ) (hd : 0 < d) (p : ℝ) : Superop (Fin d) (Fin d) :=
QuditFidelity.depolarizingMap d hd (1 - p)
/-- PDF p.5, (17): U tensor conjugate U, averaged with probability Haar law.
The second factor is entrywise conjugation, not adjoint. -/
def stateTwirl (d : ℕ)
(rho : Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ) :
Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ :=
∫ U : Matrix.unitaryGroup (Fin d) ℂ,
let W := Matrix.kronecker (U : Matrix (Fin d) (Fin d) ℂ)
(fun i j => conj ((U : Matrix (Fin d) (Fin d) ℂ) i j))
W * rho * Wᴴ ∂Schur.unitaryHaar d
/-- PDF pp.4--6, (11)--(12),(20): actual complete Bell measurement followed
by the outcome-conditioned correction. All d^2 branches are retained. -/
def standardTeleport (d : ℕ)
(rho : Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ) : Superop (Fin d) (Fin d) :=
fun X => ∑ c : Fin (d * d), conjSuperop (teleportWeyl d c)
(traceLeft (conjSuperop
(Matrix.kronecker (pureDensity (bellKet d c)) (1 : Matrix (Fin d) (Fin d) ℂ))
(initialJoint X rho)))
/-- PDF p.8: explicit classical achiever, measuring in the computational
basis and preparing the basis state reported by Alice. -/
def basisClassical (d : ℕ) : Superop (Fin d) (Fin d) :=
fun X i j => if i = j then X i i else 0
/-- PDF pp.9--10: a local successful filter is a contraction. Its failure
can be completed by sqrt(I-A†A), so it is a physical two-outcome operation. -/
def IsLocalFilter {N m : ℕ} (A : Matrix (Fin m) (Fin N) ℂ) : Prop :=
(1 - Aᴴ * A).PosSemidef
/-- PDF p.9, (33): exact noncollective m-by-m distillation. Rectangular
local maps include output-frame changes. Rescaling unrestricted A,B to
contractions preserves the conditional state and positive success. -/
def ExactlyDistillable {N M : ℕ}
(rho : Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ) (m : ℕ) (hm : 0 < m) : Prop :=
∃ A : Matrix (Fin m) (Fin N) ℂ, ∃ B : Matrix (Fin m) (Fin M) ℂ,
IsLocalFilter A ∧ IsLocalFilter B ∧
(let omega := conjSuperop (Matrix.kronecker A B) rho;
0 < omega.trace.re ∧ ((omega.trace.re)⁻¹ : ℂ) • omega = maximallyEntangledDensity m)
/-- PDF p.10: orthogonal rank-m projection on a local space. -/
def IsRankProjection {N : ℕ} (m : ℕ) (P : Matrix (Fin N) (Fin N) ℂ) : Prop :=
Pᴴ = P ∧ P * P = P ∧ P.rank = m
/-- PDF p.10: Bob's projection onto his first N coordinates. -/
def coordinateProjection (N M : ℕ) : Matrix (Fin M) (Fin M) ℂ :=
Matrix.diagonal (fun i : Fin M => if i.val < N then 1 else 0)
/-- PDF p.10: N-dimensional plus vector embedded in N tensor M. -/
def rectangularPlus (N M : ℕ) : Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ :=
pureDensity (WithLp.toLp 2 fun z : Fin N × Fin M =>
if z.1.val = z.2.val then ((Real.sqrt N)⁻¹ : ℂ) else 0)
/-- PDF p.10: convex resource family for exact rectangular extraction. -/
def rectangularResource (N M : ℕ) (p : ℝ)
(tau : Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ) :
Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ :=
(p : ℂ) • rectangularPlus N M + ((1 - p : ℝ) : ℂ) • tau
/-- Sanity: the computed matrix-unit inverse is homogeneous at zero. -/
example (d : ℕ) : choiInverse d 0 = fun _ => 0 := by
funext X a b
simp [choiInverse]
/-- Sanity: retention one means the identity channel. -/
example (d : ℕ) (hd : 0 < d) (X : Matrix (Fin d) (Fin d) ℂ) :
retentionChannel d hd 1 X = X := by
simp [retentionChannel, QuditFidelity.depolarizingMap]
/-- Sanity: mixture weight one selects the imported normalized plus projector. -/
example (d : ℕ) (hd : 0 < d) :
noisySinglet d hd 1 = maximallyEntangledDensity d := by
simp [noisySinglet]
/-- Sanity: a zero twirl has no affine constant. -/
example (d : ℕ) : stateTwirl d 0 = 0 := by
simp [stateTwirl]
/-- Sanity: every retained Bell branch vanishes on the zero input operator. -/
example (d : ℕ) (rho : Matrix (Fin d × Fin d) (Fin d × Fin d) ℂ) :
standardTeleport d rho 0 = 0 := by
have hzero : initialJoint (0 : Matrix (Fin d) (Fin d) ℂ) rho = 0 := by
ext a b
simp [initialJoint]
have hmat : Matrix.of (fun (_ _ : Fin d) => (0 : ℂ)) =
(0 : Matrix (Fin d) (Fin d) ℂ) := rfl
simp [standardTeleport, hzero, conjSuperop, traceLeft, hmat]
/-- Sanity: the classical achiever preserves each diagonal coefficient. -/
example (d : ℕ) (X : Matrix (Fin d) (Fin d) ℂ) (i : Fin d) :
basisClassical d X i i = X i i := by simp [basisClassical]
/-- Sanity: identity is a physical local filter. -/
example (N : ℕ) : IsLocalFilter (1 : Matrix (Fin N) (Fin N) ℂ) := by
simp [IsLocalFilter, Matrix.PosSemidef.zero]
/-- Sanity: identity is a full-rank orthogonal projection. -/
example (N : ℕ) : IsRankProjection N (1 : Matrix (Fin N) (Fin N) ℂ) := by
simp [IsRankProjection]
/-- Sanity: Bob retains the first coordinate in the rectangular example. -/
example : coordinateProjection 1 2 0 0 = 1 := by
norm_num [coordinateProjection]
/-- Sanity: the embedded one-dimensional target is |00><00|. -/
example : rectangularPlus 1 2 (0, 0) (0, 0) = 1 := by
norm_num [rectangularPlus, pureDensity]
/-- Sanity: weight one selects the rectangular target. -/
example (N M : ℕ) (tau : Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ) :
rectangularResource N M 1 tau = rectangularPlus N M := by
simp [rectangularResource]
/-- Sanity: a zero resource has zero success and cannot satisfy exact distillation. -/
example {N M : ℕ} (m : ℕ) (hm : 0 < m) :
¬ ExactlyDistillable (0 : Matrix (Fin N × Fin M) (Fin N × Fin M) ℂ) m hm := by
rintro ⟨A, B, _, _, hp, _⟩
simpa [conjSuperop] using hp
end QAlgorithms.Papers.Independent.Intakenext100.P18
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
QAlgorithms.Papers.Independent.Intakenext100.P18.ExactlyDistillable
For arbitrary natural numbers , an arbitrary complex matrix indexed by , a natural number , and a proof that , this predicate means that there exist complex matrices of size and of size such that and are positive semidefinite and, writing
one has
No positivity or normalization of is assumed. If or , then , so the required strict trace inequality cannot hold.
QAlgorithms.Papers.Independent.Intakenext100.P18.IsLocalFilter
For arbitrary natural numbers and a complex matrix , this predicate means that
is positive semidefinite, where denotes conjugate transpose. Both dimensions may be zero: if , the condition is positivity of , and if , it is positivity of the empty matrix.
QAlgorithms.Papers.Independent.Intakenext100.P18.IsRankProjection
For arbitrary natural numbers and a complex matrix , this predicate means precisely
Thus may be zero, in which case the condition requires ; it cannot hold when . For , it holds exactly when .
QAlgorithms.Papers.Independent.Intakenext100.P18.basisClassical
For every natural number , this defines the function on complex matrices that retains each diagonal entry and sets every off-diagonal entry to zero:
It applies to arbitrary , and includes , when the matrices have empty index sets.
QAlgorithms.Papers.Independent.Intakenext100.P18.choiInverse
For every natural number and every complex matrix indexed by pairs from , this defines a function sending each complex matrix to the matrix
No positivity, normalization, or marginal condition is imposed on . The definition includes , when the input and output matrices have empty index sets.
QAlgorithms.Papers.Independent.Intakenext100.P18.coordinateProjection
For every pair of natural numbers , this defines the complex diagonal matrix
There is no assumption : when , this is , and when , it is zero. The case gives the empty matrix.
QAlgorithms.Papers.Independent.Intakenext100.P18.noisySinglet
For every natural number , a proof that , and every real number , this defines
where all four indices lie in and is the identity matrix on the pair index set. There is no restriction . For , the resulting matrix is for every .
QAlgorithms.Papers.Independent.Intakenext100.P18.rectangularPlus
For every pair of natural numbers , this defines the matrix on , where
Equivalently, its -entry is when and , and is zero otherwise. No relation between and is required. For , its trace is , so it has trace exactly when . If either dimension is zero, the matrix has empty index sets.
QAlgorithms.Papers.Independent.Intakenext100.P18.rectangularResource
For every pair of natural numbers , every real number , and every complex matrix indexed by , this defines
There are no restrictions on , no positivity or normalization assumptions on , and no required ordering of . If either dimension is zero, the resulting matrix has empty index sets.
QAlgorithms.Papers.Independent.Intakenext100.P18.retentionChannel
For every natural number , a proof that , and every real number , this defines the function on complex matrices given by
The parameter is unrestricted; the definition itself imposes no channel condition on this function. When , the function is the identity for every .
QAlgorithms.Papers.Independent.Intakenext100.P18.standardTeleport
For every natural number and every complex matrix indexed by , this defines a function on complex matrices by
Here corresponds to a pair under the supplied finite-product indexing equivalence,
The joint matrix has entries , and sums over the first two coordinates. No density condition is imposed on . For , the sum is empty and defines the zero function on the empty-index matrix space.
QAlgorithms.Papers.Independent.Intakenext100.P18.stateTwirl
For every natural number and every complex matrix indexed by , this defines
where denotes entrywise complex conjugation, denotes conjugate transpose, is the Kronecker product, and is Haar measure on the unitary group normalized to give the whole group measure . No positivity or trace condition is imposed on . The definition includes , when the resulting matrix has empty index sets.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.