Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Low dimensional bound entanglement with one-way distillable cryptographic key

Definition
QAlgorithms_Papers_Independent_Intakenext100_P37

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

The objects the formalized statements of Low dimensional bound entanglement with one-way distillable cryptographic key (arXiv:quant-ph/0506203) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • flipAlice: PDF p.1, Eq.(1): conjugation by the Pauli X on Alice's KEY qubit only.
  • privateMixture: PDF p.1, Eq.(1): biased even/odd private-bit mixture.
  • physicalFin: PDF pp.1--3: same physical AA':BB' matrix, enumerating each party's key/shield pair with the accepted fixed local enumeration.
  • measuredSource: PDF pp.1--2: the actual measured ccq source, reparenthesized A:(B:E). The shield remains unavailable to Eve; it is traced out by associatedCCQ.
  • measuredBE: PDF p.1, Eq.(2): unnormalized conditional BE block for Alice's measured bit.
  • measuredAliceLaw: PDF pp.1--2: Alice's actual measured-bit law.
  • HasMeasuredOneWayRate: PDF pp.1--2: forward public discussion after the specified key measurement. One sequence is chosen before EVERY finite purifying Eve. The accepted protocol includes its public transcript in Eve's output; errors vanish on i.i.d. copies. Zero-mass conditional states use the accepted normalized fallback.
  • weightedSwap: PDF p.2, Eq.(10): the computed weighted swap W_U, with row (i,j), column (j,i).
  • entrySum: PDF p.2: the entrywise l1 sum equals the trace norm of W_U.
  • weightOne: PDF p.2, Eq.(11): the first mixing weight, computed from U.
  • weightTwo: PDF p.2, Eq.(11): the second mixing weight, computed from U.
  • rhoUnitary: PDF p.2, Eqs.(9)--(11): exactly the two X-form blocks of the construction. The two square roots in xForm are kept distinct, as required for complex U. The partial transpose in X2 acts on B' alone. Because flipAlice exchanges 00 with 10 and 11 with 01, its private-bit argument must be xForm (X2 adjoint) to put X2 in the (01,10) block required by Eq.(9).
  • nonzeroEntries: PDF p.2 after Eq.(12): number of nonzero matrix entries in the named basis.
  • IsFlatUnitary: PDF pp.2--3, Observation 1: source's normalized meaning of unimodular.
  • rhoHadamard: PDF p.3, Eqs.(13)--(16): the particular computed 4 by 4 bipartite resource.
  • hadamardRate: PDF p.3, Eq.(16): exact yield in bits per independent input copy. The printed decimal is rounded, not an equality to a terminating decimal.
  • pairKet: PDF p.2, Eqs.(4)--(5): normalized Bell-type vector. For the odd pair the second ket is |10>, correcting the duplicated |01> printed in Eq.(5).
  • chiKet: PDF p.3, Eq.(15): the two normalized chi vectors; true selects chi minus.
  • hadamardFlag: PDF p.3, Eq.(14): shield flags in the four Bell branches.
  • hadamardWeight: PDF p.3, below Eq.(15): the four fixed preparation probabilities.
  • pptSet: PDF p.3, Observation 2: all normalized PPT states, not a mixing line.
  • IsPPTWithKey: PDF p.3, Observation 3: physical PPT states with positive distillable key.
  • squaredDistance: PDF p.3, volume argument: squared Hilbert--Schmidt distance, giving full balls in the real affine space of Hermitian trace-one matrices.
  • embeddedHadamard: PDF p.3, higher-dimensional extension: local isometric embedding of the 4 by 4 example, with zero outside the first four basis labels at each party.
  • noisyEmbeddedHadamard: PDF p.3, Observation 3 proof: add full-dimensional white noise with signal weight q close to one. A ball is centered at this noisy state.
  • IsCCQ: PDF pp.1,3, Eqs.(2),(17): A and B are classical in the named basis.
  • dephaseAB: PDF pp.3--4: computational-basis measurement of both honest key registers.
  • bellPinching: PDF p.4, Eq.(18): actual pinching in the four standard Bell projectors.
  • twirledExpression: PDF p.4, Eq.(18): the experimentally estimated lower-bound expression.
  • realCoherence: PDF p.4, Eq.(19): real-coherence Bell-projector difference.
  • imagCoherence: PDF p.4, Eq.(19): imaginary-coherence difference. We fix the first tilde Bell ket to phase -i and the second to +i; its mean is +2 Im sigma_(first,second).
  • keyObservable: PDF p.4, Eq.(19): all five specified key observables in one finite family.
  • physicalObservable: PDF p.4, Eq.(19): conjugated physical observable on ABA'B'.
Definition code
import Mathlib
import Definitions.Def_QAlgorithms_Defs_BipartiteSeparability
import Definitions.Def_QAlgorithms_Defs_Ensemble
import Definitions.Def_QAlgorithms_Defs_Gates
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P24
import Definitions.Def_QAlgorithms_Papers_Independent_Frontier30_P12
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P42

/-!
Horodecki--Pankowski--Horodecki--Horodecki, quant-ph/0506203v1, PDF pp.1--4.
Coordinates are ABA'B'; physical entanglement always uses AA':BB'.
All operational interfaces are from the pinned accepted library. New definitions
are confined to this module. Twisting is an analytical operation, never a claimed
LOCC implementation. Printed corrections and deferred nodes are in the unit's
MODERATION_NOTES.md and HARD.md. Compilation is pending batch handoff.
-/

noncomputable section

open scoped Matrix BigOperators ComplexConjugate ComplexOrder ENNReal

namespace QAlgorithms.Papers.Independent.Intakenext100.P37

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms.Papers.Independent.Information100.P21 (Op)
open QAlgorithms.Papers.Independent.Dependency50.P24
  (Key IsPrivate xForm twist twistingMatrix associatedCCQ transposeSecond transposeBob partyCut)

/-- PDF p.1, Eq.(1): conjugation by the Pauli X on Alice's KEY qubit only. -/
def flipAlice {a b : ℕ} (rho : Op (Key 2 × (Fin a × Fin b))) : Op (Key 2 × (Fin a × Fin b)) :=
  fun x y => rho
    (((if x.1.1 = 0 then 1 else 0), x.1.2), x.2)
    (((if y.1.1 = 0 then 1 else 0), y.1.2), y.2)

/-- PDF p.1, Eq.(1): biased even/odd private-bit mixture. -/
def privateMixture {a b : ℕ} (p1 p2 : ℝ) (gamma1 gamma2 : Op (Key 2 × (Fin a × Fin b))) :
    Op (Key 2 × (Fin a × Fin b)) := (p1 : ℂ) • gamma1 + (p2 : ℂ) • flipAlice gamma2

/-- PDF pp.1--3: same physical AA':BB' matrix, enumerating each party's
key/shield pair with the accepted fixed local enumeration. -/
def physicalFin {a b : ℕ} (rho : Op (Key 2 × (Fin a × Fin b))) : Op (Fin (2*a) × Fin (2*b)) :=
  fun x y => partyCut rho
    ((Gap48.P42.localIndex 2 a).symm x.1, (Gap48.P42.localIndex 2 b).symm x.2)
    ((Gap48.P42.localIndex 2 a).symm y.1, (Gap48.P42.localIndex 2 b).symm y.2)

/-- PDF pp.1--2: the actual measured ccq source, reparenthesized A:(B:E).
The shield remains unavailable to Eve; it is traced out by associatedCCQ. -/
def measuredSource {a b e : ℕ} (psi : EuclideanSpace ℂ ((Key 2 × (Fin a × Fin b)) × Fin e)) :
    Op (Fin 2 × (Fin 2 × Fin e)) :=
  fun x y => associatedCCQ psi ((x.1,x.2.1),x.2.2) ((y.1,y.2.1),y.2.2)

/-- PDF p.1, Eq.(2): unnormalized conditional BE block for Alice's measured bit. -/
def measuredBE {a b e : ℕ} (psi : EuclideanSpace ℂ ((Key 2 × (Fin a × Fin b)) × Fin e))
    (i : Fin 2) : Op (Fin 2 × Fin e) :=
  fun x y => measuredSource psi (i,x) (i,y)

/-- PDF pp.1--2: Alice's actual measured-bit law. -/
def measuredAliceLaw {a b e : ℕ} (psi : EuclideanSpace ℂ ((Key 2 × (Fin a × Fin b)) × Fin e))
    (i : Fin 2) : ℝ := (measuredBE psi i).trace.re

/-- PDF pp.1--2: forward public discussion after the specified key measurement.
One sequence is chosen before EVERY finite purifying Eve. The accepted protocol
includes its public transcript in Eve's output; errors vanish on i.i.d. copies.
Zero-mass conditional states use the accepted normalized fallback. -/
def HasMeasuredOneWayRate {a b : ℕ} (rho : Op (Key 2 × (Fin a × Fin b))) (r : ℝ) : Prop :=
  ∃ P : (n : ℕ) → Frontier30.P12.CqqProtocol (Fin 2) (Fin 2) n,
    ∀ e : ℕ, 0 < e → ∀ psi : EuclideanSpace ℂ ((Key 2 × (Fin a × Fin b)) × Fin e),
      ‖psi‖ = 1 → traceRight (pureDensity psi) = rho →
      Frontier30.P12.CqqAchieves (measuredAliceLaw psi)
        (fun i => Frontier30.P12.normalized (measuredBE psi i)) P r

/-- PDF p.2, Eq.(10): the computed weighted swap W_U, with row (i,j), column (j,i). -/
def weightedSwap {d : ℕ} (U : Op (Fin d)) : Op (Fin d × Fin d) :=
  fun x y => if x.1 = y.2 ∧ x.2 = y.1 then U x.1 x.2 else 0

/-- PDF p.2: the entrywise l1 sum equals the trace norm of W_U. -/
def entrySum {d : ℕ} (U : Op (Fin d)) : ℝ := ∑ i, ∑ j, ‖U i j‖

/-- PDF p.2, Eq.(11): the first mixing weight, computed from U. -/
def weightOne {d : ℕ} (U : Op (Fin d)) : ℝ := entrySum U / (entrySum U + d)

/-- PDF p.2, Eq.(11): the second mixing weight, computed from U. -/
def weightTwo {d : ℕ} (U : Op (Fin d)) : ℝ := (d : ℝ) / (entrySum U + d)

/-- PDF p.2, Eqs.(9)--(11): exactly the two X-form blocks of the construction.
The two square roots in xForm are kept distinct, as required for complex U.
The partial transpose in X2 acts on B' alone. Because flipAlice exchanges
00 with 10 and 11 with 01, its private-bit argument must be xForm (X2 adjoint)
to put X2 in the (01,10) block required by Eq.(9). -/
def rhoUnitary {d : ℕ} (U : Op (Fin d)) : Op (Key 2 × (Fin d × Fin d)) :=
  privateMixture (weightOne U) (weightTwo U)
    (xForm (((entrySum U)⁻¹ : ℂ) • weightedSwap U))
    (xForm ((((d : ℝ)⁻¹ : ℂ) • transposeSecond (weightedSwap U))ᴴ))

/-- PDF p.2 after Eq.(12): number of nonzero matrix entries in the named basis. -/
def nonzeroEntries {d : ℕ} (U : Op (Fin d)) : ℕ :=
  (Finset.univ.filter fun ij : Fin d × Fin d => U ij.1 ij.2 ≠ 0).card

/-- PDF pp.2--3, Observation 1: source's normalized meaning of unimodular. -/
def IsFlatUnitary {d : ℕ} (U : Op (Fin d)) : Prop :=
  ∀ i j, ‖U i j‖ = 1 / Real.sqrt d

/-- PDF p.3, Eqs.(13)--(16): the particular computed 4 by 4 bipartite resource. -/
def rhoHadamard : Op (Key 2 × (Fin 2 × Fin 2)) :=
  rhoUnitary (fun i j : Fin 2 => QAlgorithms.hadamard (decide (i = 1)) (decide (j = 1)))

/-- PDF p.3, Eq.(16): exact yield in bits per independent input copy.
The printed decimal is rounded, not an equality to a terminating decimal. -/
def hadamardRate : ℝ := 1 - binEntropy2 (Real.sqrt 2 / (1 + Real.sqrt 2))

/-- PDF p.2, Eqs.(4)--(5): normalized Bell-type vector. For the odd pair the
second ket is |10>, correcting the duplicated |01> printed in Eq.(5). -/
def pairKet (odd : Bool) (phase : ℂ) : EuclideanSpace ℂ (Key 2) :=
  WithLp.toLp 2 fun x => ((Real.sqrt 2)⁻¹ : ℂ) *
    ((if x = (0, if odd then 1 else 0) then 1 else 0) +
      (if x = (1, if odd then 0 else 1) then phase else 0))

/-- PDF p.3, Eq.(15): the two normalized chi vectors; true selects chi minus. -/
def chiKet (minus : Bool) : EuclideanSpace ℂ (Key 2) :=
  WithLp.toLp 2 fun x =>
    if x = (0,0) then
      ((Real.sqrt (2 + (if minus then -Real.sqrt 2 else Real.sqrt 2)) / 2 : ℝ) : ℂ)
    else if x = (1,1) then
      (((if minus then -1 else 1) *
        Real.sqrt (2 - (if minus then -Real.sqrt 2 else Real.sqrt 2)) / 2 : ℝ) : ℂ)
    else 0

/-- PDF p.3, Eq.(14): shield flags in the four Bell branches. -/
def hadamardFlag (i : Fin 4) : Op (Key 2) :=
  ![(1/2 : ℂ) • (pureDensity (ket (0,0)) + pureDensity ((pairKet true 1))),
    (1/2 : ℂ) • (pureDensity (ket (1,1)) + pureDensity ((pairKet true (-1)))),
    pureDensity (chiKet false), pureDensity (chiKet true)] i

/-- PDF p.3, below Eq.(15): the four fixed preparation probabilities. -/
def hadamardWeight (i : Fin 4) : ℝ :=
  ![weightOne (fun i j : Fin 2 => QAlgorithms.hadamard (decide (i = 1)) (decide (j = 1))) / 2, weightOne (fun i j : Fin 2 => QAlgorithms.hadamard (decide (i = 1)) (decide (j = 1))) / 2,
    weightTwo (fun i j : Fin 2 => QAlgorithms.hadamard (decide (i = 1)) (decide (j = 1))) / 2, weightTwo (fun i j : Fin 2 => QAlgorithms.hadamard (decide (i = 1)) (decide (j = 1))) / 2] i

/-- PDF p.3, Observation 2: all normalized PPT states, not a mixing line. -/
def pptSet {a b : ℕ} : Set (Op (Key 2 × (Fin a × Fin b))) :=
  {rho | IsDensityMatrix rho ∧ (transposeBob rho).PosSemidef}

/-- PDF p.3, Observation 3: physical PPT states with positive distillable key. -/
def IsPPTWithKey {m n : ℕ} (rho : Op (Fin m × Fin n)) : Prop :=
  IsDensityMatrix rho ∧ (Information100.P21.pt rho).PosSemidef ∧
    0 < Gap48.P42.distillableKey rho

/-- PDF p.3, volume argument: squared Hilbert--Schmidt distance, giving full
balls in the real affine space of Hermitian trace-one matrices. -/
def squaredDistance {I : Type} [Fintype I] (rho sigma : Op I) : ℝ :=
  ∑ i, ∑ j, ‖rho i j - sigma i j‖ ^ 2

/-- PDF p.3, higher-dimensional extension: local isometric embedding of the
4 by 4 example, with zero outside the first four basis labels at each party. -/
def embeddedHadamard (m n : ℕ) : Op (Fin m × Fin n) :=
  fun x y => if h : x.1.val < 4 ∧ x.2.val < 4 ∧ y.1.val < 4 ∧ y.2.val < 4 then
    physicalFin rhoHadamard
      (⟨x.1.val,h.1⟩,⟨x.2.val,h.2.1⟩) (⟨y.1.val,h.2.2.1⟩,⟨y.2.val,h.2.2.2⟩)
    else 0

/-- PDF p.3, Observation 3 proof: add full-dimensional white noise with signal
weight q close to one. A ball is centered at this noisy state. -/
def noisyEmbeddedHadamard (m n : ℕ) (q : ℝ) : Op (Fin m × Fin n) :=
  (q : ℂ) • embeddedHadamard m n + (((1-q) / (m*n) : ℝ) : ℂ) • 1

/-- PDF pp.1,3, Eqs.(2),(17): A and B are classical in the named basis. -/
def IsCCQ {a b e : ℕ} (rho : Op (Fin a × (Fin b × Fin e))) : Prop :=
  ∀ x y, x.1 ≠ y.1 ∨ x.2.1 ≠ y.2.1 → rho x y = 0

/-- PDF pp.3--4: computational-basis measurement of both honest key registers. -/
def dephaseAB {a b e : ℕ} (rho : Op (Fin a × (Fin b × Fin e))) :
    Op (Fin a × (Fin b × Fin e)) :=
  fun x y => if x.1 = y.1 ∧ x.2.1 = y.2.1 then rho x y else 0

/-- PDF p.4, Eq.(18): actual pinching in the four standard Bell projectors. -/
def bellPinching (sigma : Op (Key 2)) : Op (Key 2) :=
  ∑ i : Fin 4, pureDensity (((![pairKet false 1, pairKet false (-1), pairKet true 1, pairKet true (-1)]) i)) * sigma * pureDensity (((![pairKet false 1, pairKet false (-1), pairKet true 1, pairKet true (-1)]) i))

/-- PDF p.4, Eq.(18): the experimentally estimated lower-bound expression. -/
def twirledExpression (sigma : Op (Key 2)) : ℝ :=
  mutualInfoBits (fun k => (sigma k k).re) - vnEntropy (bellPinching sigma)

/-- PDF p.4, Eq.(19): real-coherence Bell-projector difference. -/
def realCoherence (odd : Bool) : Op (Key 2) :=
  pureDensity (pairKet odd 1) - pureDensity (pairKet odd (-1))

/-- PDF p.4, Eq.(19): imaginary-coherence difference. We fix the first tilde
Bell ket to phase -i and the second to +i; its mean is +2 Im sigma_(first,second). -/
def imagCoherence (odd : Bool) : Op (Key 2) :=
  pureDensity (pairKet odd (-Complex.I)) - pureDensity (pairKet odd Complex.I)

/-- PDF p.4, Eq.(19): all five specified key observables in one finite family. -/
def keyObservable (i : Fin 5) : Op (Key 2) :=
  ![(Matrix.diagonal (fun k : Key 2 => if k.1 = k.2 then (1 : ℂ) else -1)), realCoherence false, imagCoherence false,
    realCoherence true, imagCoherence true] i

/-- PDF p.4, Eq.(19): conjugated physical observable on ABA'B'. -/
def physicalObservable {a b : ℕ}
    (U : Key 2 → Matrix.unitaryGroup (Fin a × Fin b) ℂ) (O : Op (Key 2)) :
    Op (Key 2 × (Fin a × Fin b)) :=
  (twistingMatrix U)ᴴ * Matrix.kronecker O (1 : Op (Fin a × Fin b)) * twistingMatrix U

/-- Sanity: W_I has unit (00,00) entry, pinning the row/column convention. -/
example : weightedSwap (1 : Op (Fin 2)) (0,0) (0,0) = 1 := by
  simp [weightedSwap]

/-- Sanity: the flip acts on Alice's key and leaves Bob and both shields fixed. -/
example {a b : ℕ} (rho : Op (Key 2 × (Fin a × Fin b))) (s t : Fin a × Fin b) :
    flipAlice rho ((0,0),s) ((1,1),t) = rho ((1,0),s) ((0,1),t) := by rfl

/-- Sanity: chi has no odd-parity shield amplitude. -/
example (minus : Bool) : chiKet minus (0,1) = 0 := by
  simp [chiKet]

/-- Sanity: an identical matrix has zero squared Hilbert--Schmidt distance. -/
example {I : Type} [Fintype I] (rho : Op I) : squaredDistance rho rho = 0 := by
  simp [squaredDistance]

end QAlgorithms.Papers.Independent.Intakenext100.P37
Source
Low dimensional bound entanglement with one-way distillable cryptographic key, arXiv:quant-ph/0506203v1, https://arxiv.org/abs/quant-ph/0506203v1
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

QAlgorithms.Papers.Independent.Intakenext100.P37.HasMeasuredOneWayRate

For arbitrary natural numbers a,ba,ba,b, an arbitrary complex matrix ρ\rhoρ on ({0,1}2)×(Fin a×Fin b)(\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b)({0,1}2)×(Fina×Finb), and a real number rrr, this property means that there exists one sequence of protocols PnP_nPn​, for every natural number nnn, with the following requirement for every positive natural number eee and every unit vector ψ\psiψ on (({0,1}2)×(Fin a×Fin b))×Fin e((\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b))\times\mathrm{Fin}\,e(({0,1}2)×(Fina×Finb))×Fine satisfying ρv,w=∑uψ(v,u)ψ(w,u)‾\rho_{v,w}=\sum_u\psi(v,u)\overline{\psi(w,u)}ρv,w​=∑u​ψ(v,u)ψ(w,u)​. Define matrices sis_isi​ on {0,1}×Fin e\{0,1\}\times\mathrm{Fin}\,e{0,1}×Fine by (si)(j,u),(j′,v)=0(s_i)_{(j,u),(j',v)}=0(si​)(j,u),(j′,v)​=0 when j≠j′j\ne j'j=j′, and otherwise by ∑s,tψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)‾\sum_{s,t}\psi(((i,j),(s,t)),u)\overline{\psi(((i,j),(s,t)),v)}∑s,t​ψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)​; put pi=Re⁡tr⁡sip_i=\operatorname{Re}\operatorname{tr}s_ipi​=Retrsi​, and Wi=si/piW_i=s_i/p_iWi​=si​/pi​ when pi≠0p_i\ne0pi​=0, while Wi=I/(2e)W_i=I/(2e)Wi​=I/(2e) when pi=0p_i=0pi​=0. Each PnP_nPn​ consists of natural numbers ℓn,mn\ell_n,m_nℓn​,mn​, an encoding array qn(x;t,k)q_n(x;t,k)qn​(x;t,k) for binary words xxx of length nnn, t<ℓnt<\ell_nt<ℓn​, k<mnk<m_nk<mn​, and matrices Dn,t,jD_{n,t,j}Dn,t,j​ on binary words of length nnn. There must exist numbers ϵn≥0\epsilon_n\ge0ϵn​≥0 and a real F≥0F\ge0F≥0, possibly depending on ψ\psiψ, such that for every nnn, ℓn,mn>0\ell_n,m_n>0ℓn​,mn​>0, each qn(x;⋅,⋅)q_n(x;\cdot,\cdot)qn​(x;⋅,⋅) is nonnegative and sums to one, each family (Dn,t,j)j<mn(D_{n,t,j})_{j<m_n}(Dn,t,j​)j<mn​​ is positive semidefinite and sums to the identity, and ℓn≤2nF\ell_n\le2^{nF}ℓn​≤2nF. Writing Bi=tr⁡Fin eWiB_i=\operatorname{tr}_{\mathrm{Fin}\,e}W_iBi​=trFine​Wi​, Ei=tr⁡{0,1}WiE_i=\operatorname{tr}_{\{0,1\}}W_iEi​=tr{0,1}​Wi​, p(x)=∏h<npxhp(x)=\prod_{h<n}p_{x_h}p(x)=∏h<n​pxh​​, Bx(y,z)=∏h<n(Bxh)yh,zhB_x(y,z)=\prod_{h<n}(B_{x_h})_{y_h,z_h}Bx​(y,z)=∏h<n​(Bxh​​)yh​,zh​​, and Ex(u,v)=∏h<n(Exh)uh,vhE_x(u,v)=\prod_{h<n}(E_{x_h})_{u_h,v_h}Ex​(u,v)=∏h<n​(Exh​​)uh​,vh​​, define

Jn(k,j)=∑x∑t<ℓnp(x)qn(x;t,k) Re⁡tr⁡(Dn,t,jBx),Sn,k((t,u),(t′,v))=1t=t′∑xp(x)qn(x;t,k)Ex(u,v).J_n(k,j)=\sum_x\sum_{t<\ell_n}p(x)q_n(x;t,k)\, \operatorname{Re}\operatorname{tr}(D_{n,t,j}B_x),\qquad S_{n,k}((t,u),(t',v)) =\mathbf1_{t=t'}\sum_xp(x)q_n(x;t,k)E_x(u,v).Jn​(k,j)=x∑​t<ℓn​∑​p(x)qn​(x;t,k)Retr(Dn,t,j​Bx​),Sn,k​((t,u),(t′,v))=1t=t′​x∑​p(x)qn​(x;t,k)Ex​(u,v).

The requirements are ∑k≠jJn(k,j)≤ϵn\sum_{k\ne j}J_n(k,j)\le\epsilon_n∑k=j​Jn​(k,j)≤ϵn​, ∑k∣Re⁡tr⁡Sn,k−1/mn∣≤ϵn\sum_k|\operatorname{Re}\operatorname{tr}S_{n,k}-1/m_n|\le\epsilon_n∑k​∣RetrSn,k​−1/mn​∣≤ϵn​, and the existence, for each nnn, of density matrices ηn,k\eta_{n,k}ηn,k​ and one common density matrix σn\sigma_nσn​ on Fin ℓn×(Fin n→Fin e)\mathrm{Fin}\,\ell_n\times(\mathrm{Fin}\,n\to\mathrm{Fin}\,e)Finℓn​×(Finn→Fine) such that (Re⁡tr⁡Sn,k)ηn,k=Sn,k(\operatorname{Re}\operatorname{tr}S_{n,k})\eta_{n,k}=S_{n,k}(RetrSn,k​)ηn,k​=Sn,k​ and ∥ηn,k−σn∥1≤ϵn\|\eta_{n,k}-\sigma_n\|_1\le\epsilon_n∥ηn,k​−σn​∥1​≤ϵn​ for every kkk. Here a density matrix is positive semidefinite with trace one, and ∥X∥1=Re⁡tr⁡X†X\|X\|_1=\operatorname{Re}\operatorname{tr}\sqrt{X^\dagger X}∥X∥1​=RetrX†X​. Finally, ϵn→0\epsilon_n\to0ϵn​→0 and log⁡2(mn)/n→r\log_2(m_n)/n\to rlog2​(mn​)/n→r as n→∞n\to\inftyn→∞. These conditions include n=0n=0n=0, using empty products equal to one and the total real division convention 1/0=01/0=01/0=0. There is no explicit assumption that ρ\rhoρ is a density matrix or that r≥0r\ge0r≥0; if no qualifying purification exists, the universal requirement on purifications is vacuous.

QAlgorithms.Papers.Independent.Intakenext100.P37.IsCCQ

For arbitrary natural numbers a,b,ea,b,ea,b,e and an arbitrary complex square matrix ρ\rhoρ indexed by Fin⁡(a)×(Fin⁡(b)×Fin⁡(e))\operatorname{Fin}(a)\times(\operatorname{Fin}(b)\times\operatorname{Fin}(e))Fin(a)×(Fin(b)×Fin(e)), this predicate means that

ρ(i,j,r),(i′,j′,r′)=0whenever i≠i′ or j≠j′.\rho_{(i,j,r),(i',j',r')}=0 \quad\text{whenever }i\ne i'\text{ or }j\ne j'.ρ(i,j,r),(i′,j′,r′)​=0whenever i=i′ or j=j′.

It imposes no restriction on entries within a fixed pair (i,j)(i,j)(i,j), and requires neither positivity nor trace one. If any dimension is zero, there are no matrix entries and the condition holds vacuously.

QAlgorithms.Papers.Independent.Intakenext100.P37.IsFlatUnitary

For every natural number ddd and complex d×dd\times dd×d matrix UUU, this predicate means that

∣Uij∣=1d(0≤i,j<d).|U_{ij}|=\frac1{\sqrt d}\qquad(0\le i,j<d).∣Uij​∣=d​1​(0≤i,j<d).

It imposes no additional condition that UUU be unitary. When d=0d=0d=0, there are no indices, so the predicate holds vacuously.

QAlgorithms.Papers.Independent.Intakenext100.P37.IsPPTWithKey

For every pair of natural numbers m,nm,nm,n and complex matrix ρ\rhoρ on {0,…,m−1}×{0,…,n−1}\{0,\ldots,m-1\}\times\{0,\ldots,n-1\}{0,…,m−1}×{0,…,n−1}, this predicate requires ρ\rhoρ to be positive semidefinite with trace 111, requires the matrix ρ(i,j),(k,l)Γ=ρ(i,l),(k,j)\rho^\Gamma_{(i,j),(k,l)}=\rho_{(i,l),(k,j)}ρ(i,j),(k,l)Γ​=ρ(i,l),(k,j)​ to be positive semidefinite, and requires 0<K(ρ)0<K(\rho)0<K(ρ). Here K(ρ)K(\rho)K(ρ) is the supremum in [0,∞][0,\infty][0,∞] of max⁡(r,0)\max(r,0)max(r,0) over real rrr satisfying the following condition: r≥0r\ge0r≥0, and for every ϵ,δ>0\epsilon,\delta>0ϵ,δ>0, there is N0∈NN_0\in\mathbb NN0​∈N such that for every integer N≥N0N\ge N_0N≥N0​ with N>0N>0N>0, there exist positive integers d,s,td,s,td,s,t, a positive semidefinite trace-one matrix τ\tauτ on Cs⊗Ct\mathbb C^s\otimes\mathbb C^tCs⊗Ct, unitaries UiU_iUi​ on that space for 0≤i<d0\le i<d0≤i<d, and a finite local-operations-and-classical-communication map Λ\LambdaΛ from local dimensions mN,nNm^N,n^NmN,nN to local dimensions ds,dtds,dtds,dt, for which

∥Λ(ρ⊗N)−γ∥1≤ϵ,r−δ≤log⁡2dN.\left\|\Lambda(\rho^{\otimes N})-\gamma\right\|_1\le\epsilon, \qquad r-\delta\le\frac{\log_2 d}{N}.​Λ(ρ⊗N)−γ​1​≤ϵ,r−δ≤Nlog2​d​.

The copies are reindexed by fixed bijections between length-NNN index tuples and indices of sizes mN,nNm^N,n^NmN,nN, and, after identifying the output indices locally with key and shield indices,

γ((i,α),(j,β)), ((k,α′),(l,β′))={d−1(UiτUk†)(α,β),(α′,β′),i=j, k=l,0,otherwise.\gamma_{((i,\alpha),(j,\beta)),\,((k,\alpha'),(l,\beta'))} = \begin{cases} d^{-1}(U_i\tau U_k^\dagger)_{(\alpha,\beta),(\alpha',\beta')},&i=j,\ k=l,\\ 0,&\text{otherwise}. \end{cases}γ((i,α),(j,β)),((k,α′),(l,β′))​={d−1(Ui​τUk†​)(α,β),(α′,β′)​,0,​i=j, k=l,otherwise.​

The permitted maps are built recursively from identity maps by finite complete local Kraus operations, with subsequent maps allowed to depend on the outcome and the branches summed. The norm is ∥X∥1=Re⁡tr⁡X†X\|X\|_1=\operatorname{Re}\operatorname{tr}\sqrt{X^\dagger X}∥X∥1​=RetrX†X​. If m=0m=0m=0 or n=0n=0n=0, the density-matrix requirement cannot hold.

QAlgorithms.Papers.Independent.Intakenext100.P37.bellPinching

For any complex square matrix σ\sigmaσ indexed by Fin⁡(2)×Fin⁡(2)\operatorname{Fin}(2)\times\operatorname{Fin}(2)Fin(2)×Fin(2), this defines

B(σ)=∑v∈{Φ+,Φ−,Ψ+,Ψ−}∣v⟩⟨v∣ σ ∣v⟩⟨v∣,Φ±=∣00⟩±∣11⟩2,Ψ±=∣01⟩±∣10⟩2.\mathcal B(\sigma)= \sum_{v\in\{\Phi_+,\Phi_-,\Psi_+,\Psi_-\}} |v\rangle\langle v|\,\sigma\,|v\rangle\langle v|, \qquad \Phi_\pm=\frac{|00\rangle\pm|11\rangle}{\sqrt2}, \quad \Psi_\pm=\frac{|01\rangle\pm|10\rangle}{\sqrt2}.B(σ)=v∈{Φ+​,Φ−​,Ψ+​,Ψ−​}∑​∣v⟩⟨v∣σ∣v⟩⟨v∣,Φ±​=2​∣00⟩±∣11⟩​,Ψ±​=2​∣01⟩±∣10⟩​.

Here each computational basis vector has amplitude one at its indicated index and zero elsewhere, and (∣v⟩⟨v∣)x,y=vxvy‾(|v\rangle\langle v|)_{x,y}=v_x\overline{v_y}(∣v⟩⟨v∣)x,y​=vx​vy​​. The definition does not require σ\sigmaσ to be Hermitian, positive, or normalized.

QAlgorithms.Papers.Independent.Intakenext100.P37.chiKet

For every Boolean minus\mathrm{minus}minus, this defines a vector in the complex Euclidean space indexed by {0,1}2\{0,1\}^2{0,1}2, namely

χfalse=2+22∣00⟩+2−22∣11⟩,χtrue=2−22∣00⟩−2+22∣11⟩.\chi_{\mathrm{false}} =\frac{\sqrt{2+\sqrt2}}2|00\rangle +\frac{\sqrt{2-\sqrt2}}2|11\rangle, \qquad \chi_{\mathrm{true}} =\frac{\sqrt{2-\sqrt2}}2|00\rangle -\frac{\sqrt{2+\sqrt2}}2|11\rangle.χfalse​=22+2​​​∣00⟩+22−2​​​∣11⟩,χtrue​=22−2​​​∣00⟩−22+2​​​∣11⟩.

Here ∣ij⟩|ij\rangle∣ij⟩ is the coordinate unit vector at (i,j)(i,j)(i,j); both other coordinates are zero.

QAlgorithms.Papers.Independent.Intakenext100.P37.dephaseAB

For arbitrary natural numbers a,b,ea,b,ea,b,e and an arbitrary complex square matrix ρ\rhoρ indexed by Fin⁡(a)×(Fin⁡(b)×Fin⁡(e))\operatorname{Fin}(a)\times(\operatorname{Fin}(b)\times\operatorname{Fin}(e))Fin(a)×(Fin(b)×Fin(e)), this defines the matrix

(DABρ)(i,j,r),(i′,j′,r′)={ρ(i,j,r),(i′,j′,r′),i=i′ and j=j′,0,otherwise.(\mathcal D_{AB}\rho)_{(i,j,r),(i',j',r')}= \begin{cases} \rho_{(i,j,r),(i',j',r')},&i=i'\text{ and }j=j',\\ 0,&\text{otherwise}. \end{cases}(DAB​ρ)(i,j,r),(i′,j′,r′)​={ρ(i,j,r),(i′,j′,r′)​,0,​i=i′ and j=j′,otherwise.​

Thus it retains every entry within each fixed pair of first and second register indices and removes all entries between different such pairs. No positivity or normalization hypothesis is imposed; zero dimensions are allowed.

QAlgorithms.Papers.Independent.Intakenext100.P37.embeddedHadamard

For every pair of natural numbers m,nm,nm,n, this defines a complex matrix EEE on {0,…,m−1}×{0,…,n−1}\{0,\ldots,m-1\}\times\{0,\ldots,n-1\}{0,…,m−1}×{0,…,n−1} by E(r,s),(u,v)=R(r,s),(u,v)E_{(r,s),(u,v)}=R_{(r,s),(u,v)}E(r,s),(u,v)​=R(r,s),(u,v)​ when all four indices are less than 444, and 000 otherwise. The 4×44\times44×4 bipartite matrix RRR is specified as follows: split each local index into a key bit and a shield bit using the fixed bijection {0,1}×{0,1}≃{0,1,2,3}\{0,1\}\times\{0,1\}\simeq\{0,1,2,3\}{0,1}×{0,1}≃{0,1,2,3} chosen from equality of cardinalities, and regroup the two key bits before the two shield bits. Set Hij=(−1)ij/2H_{ij}=(-1)^{ij}/\sqrt2Hij​=(−1)ij/2​, W(s,t),(u,v)=HstW_{(s,t),(u,v)}=H_{st}W(s,t),(u,v)​=Hst​ when s=v,t=us=v,t=us=v,t=u and 000 otherwise, T(s,t),(u,v)=HsuT_{(s,t),(u,v)}=H_{su}T(s,t),(u,v)​=Hsu​ when s=t,u=vs=t,u=vs=t,u=v and 000 otherwise, and let PPP be the diagonal matrix with entry 111 at shield indices 00,1100,1100,11 and 000 at 01,1001,1001,10. In key order 00,01,10,1100,01,10,1100,01,10,11,

R=14(1+2)(I4/200W0PT00TP0W00I4/2).R=\frac1{4(1+\sqrt2)} \begin{pmatrix} I_4/\sqrt2&0&0&W\\ 0&P&T&0\\ 0&T&P&0\\ W&0&0&I_4/\sqrt2 \end{pmatrix}.R=4(1+2​)1​​I4​/2​00W​0PT0​0TP0​W00I4​/2​​​.

There is no assumption that m,n≥4m,n\ge4m,n≥4: smaller dimensions retain only entries whose indices exist, and if either dimension is zero the resulting matrix has an empty index set.

QAlgorithms.Papers.Independent.Intakenext100.P37.entrySum

For every natural number ddd and every complex d×dd\times dd×d matrix UUU, this defines

entrySum⁡(U)=∑i=0d−1∑j=0d−1∣Uij∣.\operatorname{entrySum}(U)=\sum_{i=0}^{d-1}\sum_{j=0}^{d-1}|U_{ij}|.entrySum(U)=i=0∑d−1​j=0∑d−1​∣Uij​∣.

The norm in each summand is the complex modulus. No unitarity is assumed; for d=0d=0d=0, both index sets are empty and the value is zero.

QAlgorithms.Papers.Independent.Intakenext100.P37.flipAlice

For arbitrary natural numbers a,ba,ba,b, including zero, and any complex matrix ρ\rhoρ indexed by ({0,1}×{0,1})×(Fin a×Fin b)(\{0,1\}\times\{0,1\})\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b)({0,1}×{0,1})×(Fina×Finb), this defines a matrix on the same index set by

flipAlice⁡(ρ)((i,j),s),((i′,j′),t)=ρ((1−i,j),s),((1−i′,j′),t).\operatorname{flipAlice}(\rho)_{((i,j),s),((i',j'),t)} =\rho_{((1-i,j),s),((1-i',j'),t)}.flipAlice(ρ)((i,j),s),((i′,j′),t)​=ρ((1−i,j),s),((1−i′,j′),t)​.

Thus the first binary coordinate is flipped in both matrix indices, while every other coordinate is retained. No positivity or normalization is required; if a=0a=0a=0 or b=0b=0b=0, the matrix index set is empty.

QAlgorithms.Papers.Independent.Intakenext100.P37.hadamardFlag

For every i∈{0,1,2,3}i\in\{0,1,2,3\}i∈{0,1,2,3}, this defines a complex matrix on {0,1}2\{0,1\}^2{0,1}2. Write D(v)x,y=vxvy‾D(v)_{x,y}=v_x\overline{v_y}D(v)x,y​=vx​vy​​, let ∣ij⟩|ij\rangle∣ij⟩ denote the coordinate unit vectors, and put

ϕ±=∣01⟩±∣10⟩2,χ+=2+2∣00⟩+2−2∣11⟩2,χ−=2−2∣00⟩−2+2∣11⟩2.\phi_\pm=\frac{|01\rangle\pm|10\rangle}{\sqrt2},\quad \chi_+=\frac{\sqrt{2+\sqrt2}|00\rangle+\sqrt{2-\sqrt2}|11\rangle}{2},\quad \chi_-=\frac{\sqrt{2-\sqrt2}|00\rangle-\sqrt{2+\sqrt2}|11\rangle}{2}.ϕ±​=2​∣01⟩±∣10⟩​,χ+​=22+2​​∣00⟩+2−2​​∣11⟩​,χ−​=22−2​​∣00⟩−2+2​​∣11⟩​.

The four values, in index order, are

(D(∣00⟩)+D(ϕ+)2, D(∣11⟩)+D(ϕ−)2, D(χ+), D(χ−)).\left( \frac{D(|00\rangle)+D(\phi_+)}2,\ \frac{D(|11\rangle)+D(\phi_-)}2,\ D(\chi_+),\ D(\chi_-) \right).(2D(∣00⟩)+D(ϕ+​)​, 2D(∣11⟩)+D(ϕ−​)​, D(χ+​), D(χ−​)).

QAlgorithms.Papers.Independent.Intakenext100.P37.hadamardRate

This defines the real number

1−h2 ⁣(21+2),h2(p)=−plog⁡p−(1−p)log⁡(1−p)log⁡2,1-h_2\!\left(\frac{\sqrt2}{1+\sqrt2}\right), \qquad h_2(p)=\frac{-p\log p-(1-p)\log(1-p)}{\log2},1−h2​(1+2​2​​),h2​(p)=log2−plogp−(1−p)log(1−p)​,

where log⁡\loglog is the natural logarithm.

QAlgorithms.Papers.Independent.Intakenext100.P37.hadamardWeight

For every i∈{0,1,2,3}i\in\{0,1,2,3\}i∈{0,1,2,3}, this defines the corresponding entry of the real vector

(22(1+2), 22(1+2), 12(1+2), 12(1+2)).\left( \frac{\sqrt2}{2(1+\sqrt2)},\ \frac{\sqrt2}{2(1+\sqrt2)},\ \frac1{2(1+\sqrt2)},\ \frac1{2(1+\sqrt2)} \right).(2(1+2​)2​​, 2(1+2​)2​​, 2(1+2​)1​, 2(1+2​)1​).

These values use the sum of the absolute values of all entries of the matrix 12(111−1)\frac1{\sqrt2}\begin{pmatrix}1&1\\1&-1\end{pmatrix}2​1​(11​1−1​), which is 222\sqrt222​, in the ratios defining the weights.

QAlgorithms.Papers.Independent.Intakenext100.P37.imagCoherence

For either Boolean value odd\mathrm{odd}odd, let (a,b)=(00,11)(a,b)=(00,11)(a,b)=(00,11) when odd\mathrm{odd}odd is false and (a,b)=(01,10)(a,b)=(01,10)(a,b)=(01,10) when it is true. This defines the complex matrix

∣∣a⟩−i∣b⟩2⟩⟨∣a⟩−i∣b⟩2∣−∣∣a⟩+i∣b⟩2⟩⟨∣a⟩+i∣b⟩2∣=i∣a⟩⟨b∣−i∣b⟩⟨a∣\left|\frac{|a\rangle-i|b\rangle}{\sqrt2}\right\rangle \left\langle\frac{|a\rangle-i|b\rangle}{\sqrt2}\right| - \left|\frac{|a\rangle+i|b\rangle}{\sqrt2}\right\rangle \left\langle\frac{|a\rangle+i|b\rangle}{\sqrt2}\right| = i|a\rangle\langle b|-i|b\rangle\langle a|​2​∣a⟩−i∣b⟩​⟩⟨2​∣a⟩−i∣b⟩​​−​2​∣a⟩+i∣b⟩​⟩⟨2​∣a⟩+i∣b⟩​​=i∣a⟩⟨b∣−i∣b⟩⟨a∣

on Fin⁡(2)×Fin⁡(2)\operatorname{Fin}(2)\times\operatorname{Fin}(2)Fin(2)×Fin(2), where i2=−1i^2=-1i2=−1 and the indicated computational basis vectors have one nonzero amplitude, equal to one.

QAlgorithms.Papers.Independent.Intakenext100.P37.keyObservable

For an index i∈{0,1,2,3,4}i\in\{0,1,2,3,4\}i∈{0,1,2,3,4}, this selects the corresponding complex matrix on the two key indices from the ordered list

O0=diag⁡(1,−1,−1,1),O1=∣00⟩⟨11∣+∣11⟩⟨00∣,O2=i∣00⟩⟨11∣−i∣11⟩⟨00∣,O3=∣01⟩⟨10∣+∣10⟩⟨01∣,O4=i∣01⟩⟨10∣−i∣10⟩⟨01∣.\begin{aligned} O_0&=\operatorname{diag}(1,-1,-1,1),\\ O_1&=|00\rangle\langle11|+|11\rangle\langle00|,\\ O_2&=i|00\rangle\langle11|-i|11\rangle\langle00|,\\ O_3&=|01\rangle\langle10|+|10\rangle\langle01|,\\ O_4&=i|01\rangle\langle10|-i|10\rangle\langle01|. \end{aligned}O0​O1​O2​O3​O4​​=diag(1,−1,−1,1),=∣00⟩⟨11∣+∣11⟩⟨00∣,=i∣00⟩⟨11∣−i∣11⟩⟨00∣,=∣01⟩⟨10∣+∣10⟩⟨01∣,=i∣01⟩⟨10∣−i∣10⟩⟨01∣.​

The diagonal is written in computational basis order 00,01,10,1100,01,10,1100,01,10,11, and ∣x⟩⟨y∣|x\rangle\langle y|∣x⟩⟨y∣ has entry one in row xxx, column yyy, and zero elsewhere. The index is restricted to these five values.

QAlgorithms.Papers.Independent.Intakenext100.P37.measuredAliceLaw

For arbitrary natural numbers a,b,ea,b,ea,b,e, any complex Euclidean vector ψ\psiψ indexed by (({0,1}2)×(Fin a×Fin b))×Fin e((\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b))\times\mathrm{Fin}\,e(({0,1}2)×(Fina×Finb))×Fine, and any i∈{0,1}i\in\{0,1\}i∈{0,1}, this defines the real number

measuredAliceLaw⁡(ψ,i)=∑j∈{0,1}∑u∈Fin e∑s∈Fin a∑t∈Fin b∣ψ(((i,j),(s,t)),u)∣2.\operatorname{measuredAliceLaw}(\psi,i) =\sum_{j\in\{0,1\}}\sum_{u\in\mathrm{Fin}\,e} \sum_{s\in\mathrm{Fin}\,a}\sum_{t\in\mathrm{Fin}\,b} \left|\psi(((i,j),(s,t)),u)\right|^2.measuredAliceLaw(ψ,i)=j∈{0,1}∑​u∈Fine∑​s∈Fina∑​t∈Finb∑​∣ψ(((i,j),(s,t)),u)∣2.

This is the real part of the trace of the matrix whose ((j,u),(j′,v))((j,u),(j',v))((j,u),(j′,v)) entry is zero for j≠j′j\ne j'j=j′ and otherwise is the shield sum of ψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)‾\psi(((i,j),(s,t)),u)\overline{\psi(((i,j),(s,t)),v)}ψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)​. No unit norm assumption is imposed, so the two values need not sum to one; any empty summation index set makes the value zero.

QAlgorithms.Papers.Independent.Intakenext100.P37.measuredBE

For arbitrary natural numbers a,b,ea,b,ea,b,e, any complex Euclidean vector ψ\psiψ indexed by (({0,1}2)×(Fin a×Fin b))×Fin e((\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b))\times\mathrm{Fin}\,e(({0,1}2)×(Fina×Finb))×Fine, and any i∈{0,1}i\in\{0,1\}i∈{0,1}, this defines a matrix on {0,1}×Fin e\{0,1\}\times\mathrm{Fin}\,e{0,1}×Fine by

measuredBE⁡(ψ,i)(j,u),(j′,v)={∑s∈Fin a∑t∈Fin bψ(((i,j),(s,t)),u) ψ(((i,j),(s,t)),v)‾,j=j′,0,j≠j′.\operatorname{measuredBE}(\psi,i)_{(j,u),(j',v)} = \begin{cases} \displaystyle\sum_{s\in\mathrm{Fin}\,a}\sum_{t\in\mathrm{Fin}\,b} \psi(((i,j),(s,t)),u)\, \overline{\psi(((i,j),(s,t)),v)},&j=j',\\ 0,&j\ne j'. \end{cases}measuredBE(ψ,i)(j,u),(j′,v)​=⎩⎨⎧​s∈Fina∑​t∈Finb∑​ψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)​,0,​j=j′,j=j′.​

No normalization of ψ\psiψ is required. The matrix is zero if either shield index set is empty, and has an empty index set if e=0e=0e=0.

QAlgorithms.Papers.Independent.Intakenext100.P37.measuredSource

For arbitrary natural numbers a,b,ea,b,ea,b,e and any complex Euclidean vector ψ\psiψ indexed by (({0,1}2)×(Fin a×Fin b))×Fin e((\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b))\times\mathrm{Fin}\,e(({0,1}2)×(Fina×Finb))×Fine, this defines a matrix on {0,1}×({0,1}×Fin e)\{0,1\}\times(\{0,1\}\times\mathrm{Fin}\,e){0,1}×({0,1}×Fine) by

measuredSource⁡(ψ)(i,(j,u)),(i′,(j′,v))={∑s∈Fin a∑t∈Fin bψ(((i,j),(s,t)),u) ψ(((i,j),(s,t)),v)‾,i=i′ and j=j′,0,otherwise.\operatorname{measuredSource}(\psi)_{(i,(j,u)),(i',(j',v))} = \begin{cases} \displaystyle\sum_{s\in\mathrm{Fin}\,a}\sum_{t\in\mathrm{Fin}\,b} \psi(((i,j),(s,t)),u)\, \overline{\psi(((i,j),(s,t)),v)},&i=i'\text{ and }j=j',\\ 0,&\text{otherwise}. \end{cases}measuredSource(ψ)(i,(j,u)),(i′,(j′,v))​=⎩⎨⎧​s∈Fina∑​t∈Finb∑​ψ(((i,j),(s,t)),u)ψ(((i,j),(s,t)),v)​,0,​i=i′ and j=j′,otherwise.​

There is no unit norm assumption. Empty shield index sets make the sums zero, and e=0e=0e=0 makes the output index set empty.

QAlgorithms.Papers.Independent.Intakenext100.P37.noisyEmbeddedHadamard

For arbitrary natural numbers m,nm,nm,n, including zero, and an arbitrary real number qqq, this defines the complex matrix

q ρ^H+1−qmnIq\,\widehat\rho_H+\frac{1-q}{mn}Iqρ​H​+mn1−q​I

on Fin⁡(m)×Fin⁡(n)\operatorname{Fin}(m)\times\operatorname{Fin}(n)Fin(m)×Fin(n). Here ρ^H(x,y)\widehat\rho_H(x,y)ρ​H​(x,y) is zero unless both coordinates of both xxx and yyy are less than 444; otherwise it is the corresponding entry of the following fixed 4×44\times44×4 bipartite state, with each local index split into a key index and a shield index using the fixed chosen bijection Fin⁡(2)×Fin⁡(2)≃Fin⁡(4)\operatorname{Fin}(2)\times\operatorname{Fin}(2)\simeq\operatorname{Fin}(4)Fin(2)×Fin(2)≃Fin(4). Put

H=12(111−1),W(i,j),(k,l)={Hij,i=l, j=k,0,otherwise,H=\frac1{\sqrt2}\begin{pmatrix}1&1\\1&-1\end{pmatrix}, \qquad W_{(i,j),(k,l)}= \begin{cases}H_{ij},&i=l,\ j=k,\\0,&\text{otherwise},\end{cases}H=2​1​(11​1−1​),W(i,j),(k,l)​={Hij​,0,​i=l, j=k,otherwise,​

and write Z(i,j),(k,l)T2=Z(i,l),(k,j)Z^{T_2}_{(i,j),(k,l)}=Z_{(i,l),(k,j)}Z(i,j),(k,l)T2​​=Z(i,l),(k,j)​. For any shield matrix ZZZ, define, in key order 00,01,10,1100,01,10,1100,01,10,11,

X(Z)=12(ZZ∗00Z00000000Z∗00Z∗Z).\mathcal X(Z)=\frac12 \begin{pmatrix} \sqrt{ZZ^*}&0&0&Z\\ 0&0&0&0\\ 0&0&0&0\\ Z^*&0&0&\sqrt{Z^*Z} \end{pmatrix}.X(Z)=21​​ZZ∗​00Z∗​0000​0000​Z00Z∗Z​​​.

The fixed state is

ρH=21+2X ⁣(W22)+11+2FX ⁣(WT22)F∗,\rho_H= \frac{\sqrt2}{1+\sqrt2}\mathcal X\!\left(\frac{W}{2\sqrt2}\right) +\frac1{1+\sqrt2}F\mathcal X\!\left(\frac{W^{T_2}}2\right)F^*,ρH​=1+2​2​​X(22​W​)+1+2​1​FX(2WT2​​)F∗,

where FFF flips Alice’s key index and preserves all other indices. Thus m<4m<4m<4 or n<4n<4n<4 retains only the entries whose indices exist, while larger dimensions add zero entries. There is no restriction 0≤q≤10\le q\le10≤q≤1, nor any assertion here that the resulting matrix is a density matrix. Division by zero is defined to be zero; if either dimension is zero, the resulting matrix has an empty index set.

QAlgorithms.Papers.Independent.Intakenext100.P37.nonzeroEntries

For every natural number ddd and complex d×dd\times dd×d matrix UUU, this defines the number of its nonzero entries:

#{(i,j):0≤i,j<d, Uij≠0}.\#\{(i,j):0\le i,j<d,\ U_{ij}\ne0\}.#{(i,j):0≤i,j<d, Uij​=0}.

No unitarity assumption is imposed. When d=0d=0d=0, the count is 000.

QAlgorithms.Papers.Independent.Intakenext100.P37.pairKet

For every Boolean odd\mathrm{odd}odd and every complex number zzz, this defines the vector

{(∣00⟩+z∣11⟩)/2,odd=false,(∣01⟩+z∣10⟩)/2,odd=true,\begin{cases} \bigl(|00\rangle+z|11\rangle\bigr)/\sqrt2,&\mathrm{odd}=\mathrm{false},\\ \bigl(|01\rangle+z|10\rangle\bigr)/\sqrt2,&\mathrm{odd}=\mathrm{true}, \end{cases}{(∣00⟩+z∣11⟩)/2​,(∣01⟩+z∣10⟩)/2​,​odd=false,odd=true,​

in the complex Euclidean space indexed by pairs from {0,1}\{0,1\}{0,1}, where ∣ij⟩|ij\rangle∣ij⟩ has coordinate 111 at (i,j)(i,j)(i,j) and 000 elsewhere. The complex parameter is unrestricted; its modulus need not be 111.

QAlgorithms.Papers.Independent.Intakenext100.P37.physicalFin

For arbitrary natural numbers a,ba,ba,b and any complex matrix ρ\rhoρ on ({0,1}2)×(Fin a×Fin b)(\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b)({0,1}2)×(Fina×Finb), let ea:{0,1}×Fin a→Fin(2a)e_a:\{0,1\}\times\mathrm{Fin}\,a\to\mathrm{Fin}(2a)ea​:{0,1}×Fina→Fin(2a) and eb:{0,1}×Fin b→Fin(2b)e_b:\{0,1\}\times\mathrm{Fin}\,b\to\mathrm{Fin}(2b)eb​:{0,1}×Finb→Fin(2b) be the chosen equivalences obtained from equality of finite cardinalities. This defines a matrix on Fin(2a)×Fin(2b)\mathrm{Fin}(2a)\times\mathrm{Fin}(2b)Fin(2a)×Fin(2b) by

physicalFin⁡(ρ)(ea(i,s),eb(j,t)),(ea(i′,s′),eb(j′,t′))=ρ((i,j),(s,t)),((i′,j′),(s′,t′)).\operatorname{physicalFin}(\rho)_{(e_a(i,s),e_b(j,t)),(e_a(i',s'),e_b(j',t'))} =\rho_{((i,j),(s,t)),((i',j'),(s',t'))}.physicalFin(ρ)(ea​(i,s),eb​(j,t)),(ea​(i′,s′),eb​(j′,t′))​=ρ((i,j),(s,t)),((i′,j′),(s′,t′))​.

The equivalences are the particular cardinality based choices in the definition, with no specified arithmetic enumeration. No positivity or normalization is assumed, and a=0a=0a=0 or b=0b=0b=0 gives an empty index set.

QAlgorithms.Papers.Independent.Intakenext100.P37.physicalObservable

For arbitrary natural numbers a,ba,ba,b, a family of unitary matrices UkU_kUk​ on the shield space indexed by Fin⁡(a)×Fin⁡(b)\operatorname{Fin}(a)\times\operatorname{Fin}(b)Fin(a)×Fin(b), one for each key pair k∈Fin⁡(2)×Fin⁡(2)k\in\operatorname{Fin}(2)\times\operatorname{Fin}(2)k∈Fin(2)×Fin(2), and an arbitrary complex matrix OOO on the key pairs, this defines

TU∗(O⊗I)TU,(TU)(k,s),(l,t)={(Uk)s,t,k=l,0,k≠l.T_U^*(O\otimes I)T_U, \qquad (T_U)_{(k,s),(l,t)}= \begin{cases}(U_k)_{s,t},&k=l,\\0,&k\ne l.\end{cases}TU∗​(O⊗I)TU​,(TU​)(k,s),(l,t)​={(Uk​)s,t​,0,​k=l,k=l.​

Equivalently, its entry in row (k,s)(k,s)(k,s), column (l,t)(l,t)(l,t) is Ok,l(Uk∗Ul)s,tO_{k,l}(U_k^*U_l)_{s,t}Ok,l​(Uk∗​Ul​)s,t​. Neither Hermiticity nor positivity of OOO is assumed. Zero shield dimensions are allowed, in which case the output matrix has an empty index set.

QAlgorithms.Papers.Independent.Intakenext100.P37.pptSet

For every pair of natural numbers a,ba,ba,b, this defines the set of complex matrices ρ\rhoρ indexed by {0,1}2×({0,…,a−1}×{0,…,b−1})\{0,1\}^2\times(\{0,\ldots,a-1\}\times\{0,\ldots,b-1\}){0,1}2×({0,…,a−1}×{0,…,b−1}) such that ρ\rhoρ is positive semidefinite, tr⁡ρ=1\operatorname{tr}\rho=1trρ=1, and the matrix ρΓ\rho^\GammaρΓ is positive semidefinite, where

(ρΓ)((i,j),(s,t)), ((k,l),(u,v))=ρ((i,l),(s,v)), ((k,j),(u,t)).(\rho^\Gamma)_{((i,j),(s,t)),\,((k,l),(u,v))} =\rho_{((i,l),(s,v)),\,((k,j),(u,t))}.(ρΓ)((i,j),(s,t)),((k,l),(u,v))​=ρ((i,l),(s,v)),((k,j),(u,t))​.

Thus the transpose exchanges both the second key coordinate and the second shield coordinate between row and column. If a=0a=0a=0 or b=0b=0b=0, the index set is empty and the trace-one requirement makes this set empty.

QAlgorithms.Papers.Independent.Intakenext100.P37.privateMixture

For arbitrary natural numbers a,ba,ba,b, arbitrary real numbers p1,p2p_1,p_2p1​,p2​, and arbitrary complex matrices γ1,γ2\gamma_1,\gamma_2γ1​,γ2​ on ({0,1}2)×(Fin a×Fin b)(\{0,1\}^2)\times(\mathrm{Fin}\,a\times\mathrm{Fin}\,b)({0,1}2)×(Fina×Finb), this defines

privateMixture⁡(p1,p2,γ1,γ2)((i,j),s),((i′,j′),t)=p1(γ1)((i,j),s),((i′,j′),t)+p2(γ2)((1−i,j),s),((1−i′,j′),t).\operatorname{privateMixture}(p_1,p_2,\gamma_1,\gamma_2)_{((i,j),s),((i',j'),t)} =p_1(\gamma_1)_{((i,j),s),((i',j'),t)} +p_2(\gamma_2)_{((1-i,j),s),((1-i',j'),t)}.privateMixture(p1​,p2​,γ1​,γ2​)((i,j),s),((i′,j′),t)​=p1​(γ1​)((i,j),s),((i′,j′),t)​+p2​(γ2​)((1−i,j),s),((1−i′,j′),t)​.

The real coefficients are regarded as complex scalars. They need not be nonnegative or sum to one, and the matrices need not be density matrices or private states. Zero values of aaa or bbb are allowed, giving an empty matrix index set.

QAlgorithms.Papers.Independent.Intakenext100.P37.realCoherence

For either Boolean value odd\mathrm{odd}odd, let (a,b)=(00,11)(a,b)=(00,11)(a,b)=(00,11) when odd\mathrm{odd}odd is false and (a,b)=(01,10)(a,b)=(01,10)(a,b)=(01,10) when it is true. This defines the complex matrix

∣∣a⟩+∣b⟩2⟩⟨∣a⟩+∣b⟩2∣−∣∣a⟩−∣b⟩2⟩⟨∣a⟩−∣b⟩2∣=∣a⟩⟨b∣+∣b⟩⟨a∣\left|\frac{|a\rangle+|b\rangle}{\sqrt2}\right\rangle \left\langle\frac{|a\rangle+|b\rangle}{\sqrt2}\right| - \left|\frac{|a\rangle-|b\rangle}{\sqrt2}\right\rangle \left\langle\frac{|a\rangle-|b\rangle}{\sqrt2}\right| = |a\rangle\langle b|+|b\rangle\langle a|​2​∣a⟩+∣b⟩​⟩⟨2​∣a⟩+∣b⟩​​−​2​∣a⟩−∣b⟩​⟩⟨2​∣a⟩−∣b⟩​​=∣a⟩⟨b∣+∣b⟩⟨a∣

on Fin⁡(2)×Fin⁡(2)\operatorname{Fin}(2)\times\operatorname{Fin}(2)Fin(2)×Fin(2), where the indicated computational basis vectors have one nonzero amplitude, equal to one.

QAlgorithms.Papers.Independent.Intakenext100.P37.rhoHadamard

This defines a complex matrix on ({0,1}×{0,1})×({0,1}×{0,1})(\{0,1\}\times\{0,1\})\times(\{0,1\}\times\{0,1\})({0,1}×{0,1})×({0,1}×{0,1}). Explicitly, put Hij=(−1)ij/2H_{ij}=(-1)^{ij}/\sqrt2Hij​=(−1)ij/2​ for i,j∈{0,1}i,j\in\{0,1\}i,j∈{0,1}, and define matrices on the second pair of indices by

W(s,t),(u,v)={Hsts=v, t=u,0otherwise,T(s,t),(u,v)={Hsus=t, u=v,0otherwise,P(s,t),(u,v)={1(s,t)=(u,v), s=t,0otherwise.W_{(s,t),(u,v)}= \begin{cases}H_{st}&s=v,\ t=u,\\0&\text{otherwise},\end{cases} \quad T_{(s,t),(u,v)}= \begin{cases}H_{su}&s=t,\ u=v,\\0&\text{otherwise},\end{cases} \quad P_{(s,t),(u,v)}= \begin{cases}1&(s,t)=(u,v),\ s=t,\\0&\text{otherwise}.\end{cases}W(s,t),(u,v)​={Hst​0​s=v, t=u,otherwise,​T(s,t),(u,v)​={Hsu​0​s=t, u=v,otherwise,​P(s,t),(u,v)​={10​(s,t)=(u,v), s=t,otherwise.​

In the order 00,01,10,1100,01,10,1100,01,10,11 for the first pair of indices, the defined matrix is

14(1+2)(I4/200W0PT00TP0W00I4/2).\frac1{4(1+\sqrt2)} \begin{pmatrix} I_4/\sqrt2&0&0&W\\ 0&P&T&0\\ 0&T&P&0\\ W&0&0&I_4/\sqrt2 \end{pmatrix}.4(1+2​)1​​I4​/2​00W​0PT0​0TP0​W00I4​/2​​​.

QAlgorithms.Papers.Independent.Intakenext100.P37.rhoUnitary

For every natural number ddd, including 000, and every complex d×dd\times dd×d matrix UUU, with no unitarity assumption, this defines a complex matrix on {0,1}2×({0,…,d−1}2)\{0,1\}^2\times(\{0,\ldots,d-1\}^2){0,1}2×({0,…,d−1}2) as follows. Put s=∑i,j=0d−1∣Uij∣s=\sum_{i,j=0}^{d-1}|U_{ij}|s=∑i,j=0d−1​∣Uij​∣, and define the shield matrix WWW by W(i,j),(k,l)=UijW_{(i,j),(k,l)}=U_{ij}W(i,j),(k,l)​=Uij​ when i=li=li=l and j=kj=kj=k, and 000 otherwise. For any shield matrix ZZZ, write Z(i,j),(k,l)Γ=Z(i,l),(k,j)Z^\Gamma_{(i,j),(k,l)}=Z_{(i,l),(k,j)}Z(i,j),(k,l)Γ​=Z(i,l),(k,j)​, and define the following matrix, whose block order on the two key coordinates is 00,01,10,1100,01,10,1100,01,10,11:

X(Z)=12(ZZ†00Z00000000Z†00Z†Z),\mathcal X(Z)=\frac12 \begin{pmatrix} \sqrt{ZZ^\dagger}&0&0&Z\\ 0&0&0&0\\ 0&0&0&0\\ Z^\dagger&0&0&\sqrt{Z^\dagger Z} \end{pmatrix},X(Z)=21​​ZZ†​00Z†​0000​0000​Z00Z†Z​​​,

where the square roots are defined by continuous functional calculus. Let F(Y)\mathcal F(Y)F(Y) denote the matrix obtained by flipping Alice’s key coordinate in both indices, so that F(Y)((i,j),u),((k,l),v)=Y((1−i,j),u),((1−k,l),v)\mathcal F(Y)_{((i,j),u),((k,l),v)}=Y_{((1-i,j),u),((1-k,l),v)}F(Y)((i,j),u),((k,l),v)​=Y((1−i,j),u),((1−k,l),v)​. The defined matrix is

ρU=ss+d X(s−1W)+ds+d F ⁣(X ⁣((d−1WΓ)†)).\rho_U= \frac{s}{s+d}\,\mathcal X(s^{-1}W) +\frac{d}{s+d}\, \mathcal F\!\left(\mathcal X\!\left((d^{-1}W^\Gamma)^\dagger\right)\right).ρU​=s+ds​X(s−1W)+s+dd​F(X((d−1WΓ)†)).

All scalar inverses use the convention 0−1=00^{-1}=00−1=0. In particular, U=0U=0U=0 gives the zero matrix; when d=0d=0d=0, the indexing set is empty, s=0s=0s=0, and both displayed weights are 000.

QAlgorithms.Papers.Independent.Intakenext100.P37.squaredDistance

For every finite type III, equipped with its finite enumeration, and every pair of complex matrices ρ,σ\rho,\sigmaρ,σ indexed by III, this defines

∑i∈I∑j∈I∣ρij−σij∣2.\sum_{i\in I}\sum_{j\in I}|\rho_{ij}-\sigma_{ij}|^2.i∈I∑​j∈I∑​∣ρij​−σij​∣2.

No positivity or trace assumptions are imposed. If III is empty, the value is 000.

QAlgorithms.Papers.Independent.Intakenext100.P37.twirledExpression

For any complex square matrix σ\sigmaσ indexed by Fin⁡(2)×Fin⁡(2)\operatorname{Fin}(2)\times\operatorname{Fin}(2)Fin(2)×Fin(2), put pij=Re⁡σ(i,j),(i,j)p_{ij}=\operatorname{Re}\sigma_{(i,j),(i,j)}pij​=Reσ(i,j),(i,j)​, pi∙=∑jpijp_{i\bullet}=\sum_jp_{ij}pi∙​=∑j​pij​, and p∙j=∑ipijp_{\bullet j}=\sum_ip_{ij}p∙j​=∑i​pij​. This defines

∑i,j=01pij ℓ2 ⁣(pijpi∙p∙j)−S ⁣(∑v∈{Φ+,Φ−,Ψ+,Ψ−}∣v⟩⟨v∣σ∣v⟩⟨v∣),\sum_{i,j=0}^1p_{ij}\, \ell_2\!\left(\frac{p_{ij}}{p_{i\bullet}p_{\bullet j}}\right) -S\!\left(\sum_{v\in\{\Phi_+,\Phi_-,\Psi_+,\Psi_-\}} |v\rangle\langle v|\sigma|v\rangle\langle v|\right),i,j=0∑1​pij​ℓ2​(pi∙​p∙j​pij​​)−S​v∈{Φ+​,Φ−​,Ψ+​,Ψ−​}∑​∣v⟩⟨v∣σ∣v⟩⟨v∣​,

where Φ±=(∣00⟩±∣11⟩)/2\Phi_\pm=(|00\rangle\pm|11\rangle)/\sqrt2Φ±​=(∣00⟩±∣11⟩)/2​, Ψ±=(∣01⟩±∣10⟩)/2\Psi_\pm=(|01\rangle\pm|10\rangle)/\sqrt2Ψ±​=(∣01⟩±∣10⟩)/2​, and S(X)=−Re⁡tr⁡f(X)S(X)=-\operatorname{Re}\operatorname{tr}f(X)S(X)=−Retrf(X) for f(t)=tℓ2(t)f(t)=t\ell_2(t)f(t)=tℓ2​(t), using real spectral functional calculus, whose value is zero on a non-Hermitian matrix. The total logarithm used here is ℓ2(t)=log⁡∣t∣/log⁡2\ell_2(t)=\log|t|/\log2ℓ2​(t)=log∣t∣/log2 for t≠0t\ne0t=0, with ℓ2(0)=0\ell_2(0)=0ℓ2​(0)=0, and division by zero is zero. No hypothesis makes the numbers pijp_{ij}pij​ a probability distribution or makes σ\sigmaσ a density matrix.

QAlgorithms.Papers.Independent.Intakenext100.P37.weightOne

For every natural number ddd and every complex d×dd\times dd×d matrix UUU, this defines

weightOne⁡(U)=∑i,j<d∣Uij∣∑i,j<d∣Uij∣+d.\operatorname{weightOne}(U) =\frac{\displaystyle\sum_{i,j<d}|U_{ij}|} {\displaystyle\sum_{i,j<d}|U_{ij}|+d}.weightOne(U)=i,j<d∑​∣Uij​∣+di,j<d∑​∣Uij​∣​.

No unitarity is required. When d=0d=0d=0, the numerator and denominator are zero and the defined value is zero, using total real division; for d>0d>0d>0, the denominator is positive.

QAlgorithms.Papers.Independent.Intakenext100.P37.weightTwo

For every natural number ddd and every complex d×dd\times dd×d matrix UUU, this defines

weightTwo⁡(U)=d∑i,j<d∣Uij∣+d.\operatorname{weightTwo}(U) =\frac{d}{\displaystyle\sum_{i,j<d}|U_{ij}|+d}.weightTwo(U)=i,j<d∑​∣Uij​∣+dd​.

No unitarity is assumed. When d=0d=0d=0, this is 0/00/00/0, whose defined real value is zero; for d>0d>0d>0, the denominator is positive.

QAlgorithms.Papers.Independent.Intakenext100.P37.weightedSwap

For every natural number ddd, including zero, and every complex d×dd\times dd×d matrix UUU, this defines a matrix on Fin d×Fin d\mathrm{Fin}\,d\times\mathrm{Fin}\,dFind×Find by

weightedSwap⁡(U)(i,j),(k,l)={Uij,i=l and j=k,0,otherwise.\operatorname{weightedSwap}(U)_{(i,j),(k,l)} = \begin{cases} U_{ij},&i=l\text{ and }j=k,\\ 0,&\text{otherwise}. \end{cases}weightedSwap(U)(i,j),(k,l)​={Uij​,0,​i=l and j=k,otherwise.​

No unitarity or other restriction on UUU is assumed. For d=0d=0d=0, the matrix index set is empty.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me