Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Optimal continuous variable quantum teleportation with limited resources

Definition
QAlgorithms_Papers_Independent_Intakethird100_P89

by Gohan · Nov 1, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

The objects the formalized statements of Optimal continuous variable quantum teleportation with limited resources (arXiv:1705.03017) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • OscillatorChannel: PDF p.2: an operation on complete single-oscillator coefficient arrays.
  • pairKernel: PDF p.2: basis relabeling of the actual two-mode density, without a cutoff.
  • CenteredResource: PDF p.2: the specified, symmetric, centered Gaussian covariance of an actual normalized trace-class resource. All characteristic series converge. The accepted physicality uses V+iOmega; complex conjugation makes this identical to the source's V-iOmega condition for real V.
  • twoModeMeanPhotonEnergy: PDF p.2: actual per-mode mean photon number, in extended nonnegative reals; infinite energy is retained as infinity, rather than converted to zero.
  • bkTest: PDF pp.2--3: source BK signs, in the argument of exp(i v^T R). D(z) at the output tests D(-g conjugate(z)) on A and D(z) on B. This encodes q_out=q_B+g q_A+g q_in, p_out=p_B-g p_A+g p_in.
  • IsBK: PDF pp.2--3, deterministic BK protocol: exact characteristic-function representation on EVERY input density, including non-Gaussian inputs. The resource, gain and single CPTP map are fixed before the input. This is independent of any asserted Gaussian noise formula.
  • FiniteBKRealization: PDF pp.3--4: a finite-energy Gaussian-resource BK realization with EXACT logarithmic negativity 2r. No photon-energy budget is imposed.
  • FiniteBKSimulates: PDF p.3: finite-energy realizability, distinguished from the closed region.
  • priorDensity: PDF p.4, (11): density of the NORMALIZED Gaussian prior for lambda>0, against planar Lebesgue measure. It is never used as a flat prior at lambda=0.
  • teleportedCoherentOverlap: PDF pp.3--4: squared pure-state overlap. The accepted husimiQ has no 1/pi factor, so it is precisely the overlap, not this paper's Q function. All uses are on trace-class channel outputs; realization is concluded below.
  • gaussianPriorFidelity: PDF p.4, (11): actual prior-averaged overlap, independently of (12).
  • phaseInsensitiveFidelityFormula: PDF p.4, (12): the scalar expression, separate from the actual integral.
  • entanglementAccessibleRegion: PDF pp.3--4, (3),(8): CLOSED scalar region, including limiting channels.
  • scalarValues: PDF p.4: independent scalar objective set over the closed region.
  • scalarOptimum: PDF p.4: the optimum is defined independently of the claimed answer.
  • finiteBKValues: PDF pp.3--4: actually attainable fidelities of finite-energy protocols. A resource and channel precede every input in IsBK.
  • finiteBKOptimum: PDF p.4: operational supremum, independently of the scalar formula.
  • stationaryTau: PDF p.4, immediately before (13): unconstrained stationary transmissivity.
  • optimalTau: PDF p.4: maximizing closed-region channel transmissivity.
  • optimalNoise: PDF p.4: maximizing channel noise.
  • optimumClosedForm: PDF p.4, (13): CLAIMED closed form, not the definition of an optimum.
  • familyDenominator: PDF p.3, (9): denominator, used only where strictly positive.
  • familyMinB: PDF p.3, (9): the lower b coefficient on the nonsingular domain.
  • familyA: PDF p.3, (9): a computed from b, not an independent state hypothesis.
  • familyC: PDF p.3, (9): c computed from b, with source block C=-c sigma_z.
  • familyEnergy: PDF p.3: energy within the displayed family; no global energy minimization.
  • familyMinA: PDF p.3: a on the minimum-energy member of (9).
  • familyMinC: PDF p.3: c on the minimum-energy member of (9).
  • familyMinEnergy: PDF p.3: minimum energy WITHIN family (9).
  • tmssCoefficient: PDF p.3: actual pure TMSS coefficients with the source correlation sign. The alternating sign yields C=-sinh(2r) sigma_z; r is finite.
  • tmssKernel: PDF p.3: full TMSS rank-one coefficient array; normalization is concluded.
  • tmssNoise: PDF p.3, (7): independently defined TMSS noise at gain g and squeezing s.
  • minimumSqueezing: PDF p.3, (10): minimum is defined by the noise equation, NOT its answer. The theorem concludes a least element; no empty real-infimum default is used.
  • squeezingClosedForm: PDF p.3, (10): real arcosh written as log(x+sqrt(x^2-1)). The stated domain ensures its argument is at least one. tau=1 uses (7) directly.
  • heterodynePrepare: PDF p.4: heterodyne POVM d^2beta |beta><beta|/pi followed by coherent preparation |g beta>. The complex overlap retains linearity on trace class.
  • tmssOptimalGain: PDF p.4, (16): optimized positive displacement gain for a fixed TMSS.
  • tmssClosedForm: PDF p.4, (16): claimed fixed-TMSS optimum, distinct from its supremum.
  • tmssValues: PDF p.4: actual fidelities with this fixed TMSS, optimized only over g>0.
  • tmssOptimum: PDF p.4: fixed-TMSS optimum, defined independently of (16).
Definition code
import Mathlib
import Definitions.Def_QAlgorithms_Papers_Independent_Claimed50_P19
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P26
import Definitions.Def_QAlgorithms_Defs_BipartiteSeparability

/-!
Liuzzo-Scorpo et al., arXiv:1705.03017v5, PDF pp.1--4.
Full completed oscillator spaces, hbar=1, interleaved coordinates, doubled
covariance with vacuum I, squared overlap fidelity, and natural logarithms.
Accepted paper definitions are imported under the independent-paper override.
No finite photon cutoff and no infinite-energy density are introduced.

The BK map is specified by its exact all-input Weyl characteristic, the
phase-space representation of the balanced homodyne/feed-forward protocol.
It is defined from the actual resource characteristic, independently of the
Gaussian-channel simulation assertions. It tests every normalized input.
The local centered-Gaussian adapter explicitly requires covariance symmetry;
it never uses Claimed50.P47.HasBipartiteMoments.

Corrections needing independent moderation: (14)'s tau is gain squared;
(10) is extended at tau=1 and its strict comparison has equality exceptions.
The abstract's all-channel assertion is deferred, not silently restricted.
Compilation and independent moderation are pending the batch handoff.
-/

noncomputable section

open MeasureTheory Filter
open scoped BigOperators ENNReal Topology ComplexOrder

attribute [local instance] Classical.propDecidable Classical.decEq

namespace QAlgorithms.Papers.Independent.Intakethird100.P89

open QAlgorithms.Papers.Independent.Dependency50.P33
  (State IsDensity IsTraceClassChannel tensorApply traceSize)
open QAlgorithms.Papers.Independent.Dependency50.P39
  (coherentKet coherentCoefficient operatorOfMatrix husimiQ HasFockMatrix CoherentRealized)
open QAlgorithms.Papers.Independent.Frontier30.P14 (weylCharacteristic)
open QAlgorithms.Papers.Independent.Claimed50.P19
  (Cov FockIndex DensityKernel PhysicalCovariance CharacteristicSummable
   multimodeWeylTrace covarianceCharacteristicValue standardFormCovariance)
open QAlgorithms.Papers.Independent.Frontier30.P08 (QuadratureIndex)
open QAlgorithms.Papers.Independent.Gap48.P26 (PhaseInsensitive coherentKernel)
open QAlgorithms.Entanglement (partialTransposeFirst)

/-!
Repair reuse audit: these paper-specific objects have descriptive names because
accepted short-name matches are different concepts, not reusable equivalents.
Support50.P14.photonEnergy sums all modes on occupation-function indices;
twoModeMeanPhotonEnergy averages two modes on pair indices. Claimed50.P36's
coherentOverlap is a complex overlap of two coherent kets; the present output
fidelity uses the already imported Dependency50.P39.husimiQ. QuditFidelity's
averageFidelity is a finite-dimensional Haar average. Support50.P11's
channelFidelity is finite-dimensional entanglement fidelity. Frontier30.P06's
FeedbackExecution.accessible is accessible information of an executed protocol.
None replaces a Gaussian-prior oscillator integral or the closed (tau,y) region.
The five local bodies and all theorem propositions are otherwise unchanged.
-/

/-- PDF p.2: an operation on complete single-oscillator coefficient arrays. -/
abbrev OscillatorChannel := Matrix ℕ ℕ ℂ → Matrix ℕ ℕ ℂ

/-- PDF p.2: basis relabeling of the actual two-mode density, without a cutoff. -/
def pairKernel (X : Matrix (ℕ × ℕ) (ℕ × ℕ) ℂ) : DensityKernel 2 :=
  fun u v => X (u 0, u 1) (v 0, v 1)

/-- PDF p.2: the specified, symmetric, centered Gaussian covariance of an
actual normalized trace-class resource. All characteristic series converge.
The accepted physicality uses V+iOmega; complex conjugation makes this
identical to the source's V-iOmega condition for real V. -/
def CenteredResource (rho : State (ℕ × ℕ)) (V : Cov 2) : Prop :=
  V.transpose = V ∧ PhysicalCovariance V ∧
    ∀ v : QuadratureIndex 2 → ℝ,
      CharacteristicSummable (pairKernel rho.coefficients) v ∧
      multimodeWeylTrace (pairKernel rho.coefficients) v =
        covarianceCharacteristicValue V 0 v

/-- PDF p.2: actual per-mode mean photon number, in extended nonnegative
reals; infinite energy is retained as infinity, rather than converted to zero. -/
def twoModeMeanPhotonEnergy (rho : State (ℕ × ℕ)) : ℝ≥0∞ :=
  ∑' u : ℕ × ℕ,
    ENNReal.ofReal ((((u.1 : ℝ) + (u.2 : ℝ)) / 2) *
      (rho.coefficients u u).re)

/-- PDF pp.2--3: source BK signs, in the argument of exp(i v^T R).
D(z) at the output tests D(-g conjugate(z)) on A and D(z) on B.
This encodes q_out=q_B+g q_A+g q_in, p_out=p_B-g p_A+g p_in. -/
def bkTest (g : ℝ) (z : ℂ) : QuadratureIndex 2 → ℝ :=
  fun i => if i.1 = 0 then
    (if i.2 = 0 then Real.sqrt 2 * g * z.im else Real.sqrt 2 * g * z.re)
  else (if i.2 = 0 then Real.sqrt 2 * z.im else -Real.sqrt 2 * z.re)

/-- PDF pp.2--3, deterministic BK protocol: exact characteristic-function
representation on EVERY input density, including non-Gaussian inputs.
The resource, gain and single CPTP map are fixed before the input.
This is independent of any asserted Gaussian noise formula. -/
def IsBK (rho : State (ℕ × ℕ)) (g : ℝ) (Phi : OscillatorChannel) : Prop :=
  0 < g ∧ IsTraceClassChannel Phi ∧
    ∀ (sigma : State ℕ) (z : ℂ),
      weylCharacteristic (Phi sigma.coefficients) z =
        weylCharacteristic sigma.coefficients ((g : ℂ) * z) *
          multimodeWeylTrace (pairKernel rho.coefficients) (bkTest g z)

/-- PDF pp.3--4: a finite-energy Gaussian-resource BK realization with EXACT
logarithmic negativity 2r. No photon-energy budget is imposed. -/
def FiniteBKRealization (r tau y : ℝ) (Phi : OscillatorChannel) : Prop :=
  ∃ (rho : State (ℕ × ℕ)) (V : Cov 2) (g : ℝ),
    CenteredResource rho V ∧ twoModeMeanPhotonEnergy rho < ⊤ ∧
    traceSize (partialTransposeFirst rho.coefficients) = ENNReal.ofReal (Real.exp (2 * r)) ∧
    IsBK rho g Phi ∧ g ^ 2 = tau ∧ 0 ≤ tau ∧ PhaseInsensitive Phi tau y

/-- PDF p.3: finite-energy realizability, distinguished from the closed region. -/
def FiniteBKSimulates (r tau y : ℝ) : Prop :=
  ∃ Phi : OscillatorChannel, FiniteBKRealization r tau y Phi

/-- PDF p.4, (11): density of the NORMALIZED Gaussian prior for lambda>0,
against planar Lebesgue measure. It is never used as a flat prior at lambda=0. -/
def priorDensity (lambda : ℝ) (alpha : ℂ) : ℝ :=
  (lambda / Real.pi) * Real.exp (-lambda * Complex.normSq alpha)

/-- PDF pp.3--4: squared pure-state overlap. The accepted husimiQ has no
1/pi factor, so it is precisely the overlap, not this paper's Q function.
All uses are on trace-class channel outputs; realization is concluded below. -/
def teleportedCoherentOverlap (Phi : OscillatorChannel) (alpha : ℂ) : ℝ :=
  husimiQ (operatorOfMatrix (Phi (coherentKernel alpha))) alpha

/-- PDF p.4, (11): actual prior-averaged overlap, independently of (12). -/
def gaussianPriorFidelity (lambda : ℝ) (Phi : OscillatorChannel) : ℝ :=
  ∫ alpha : ℂ, priorDensity lambda alpha * teleportedCoherentOverlap Phi alpha

/-- PDF p.4, (12): the scalar expression, separate from the actual integral. -/
def phaseInsensitiveFidelityFormula (lambda tau y : ℝ) : ℝ :=
  2 * lambda / (2 * (1 - Real.sqrt tau) ^ 2 + lambda * (1 + y + tau))

/-- PDF pp.3--4, (3),(8): CLOSED scalar region, including limiting channels. -/
def entanglementAccessibleRegion (r : ℝ) : Set (ℝ × ℝ) :=
  {p | 0 ≤ p.1 ∧ |1 - p.1| ≤ p.2 ∧ Real.exp (-2 * r) * (1 + p.1) ≤ p.2}

/-- PDF p.4: independent scalar objective set over the closed region. -/
def scalarValues (lambda r : ℝ) : Set ℝ :=
  (fun p : ℝ × ℝ => phaseInsensitiveFidelityFormula lambda p.1 p.2) '' entanglementAccessibleRegion r

/-- PDF p.4: the optimum is defined independently of the claimed answer. -/
def scalarOptimum (lambda r : ℝ) : ℝ := sSup (scalarValues lambda r)

/-- PDF pp.3--4: actually attainable fidelities of finite-energy protocols.
A resource and channel precede every input in IsBK. -/
def finiteBKValues (lambda r : ℝ) : Set ℝ :=
  {f | ∃ (tau y : ℝ) (Phi : OscillatorChannel),
    FiniteBKRealization r tau y Phi ∧ f = gaussianPriorFidelity lambda Phi}

/-- PDF p.4: operational supremum, independently of the scalar formula. -/
def finiteBKOptimum (lambda r : ℝ) : ℝ := sSup (finiteBKValues lambda r)

/-- PDF p.4, immediately before (13): unconstrained stationary transmissivity. -/
def stationaryTau (lambda r : ℝ) : ℝ :=
  Real.exp (2 * r) / (Real.exp r + lambda * Real.cosh r) ^ 2

/-- PDF p.4: maximizing closed-region channel transmissivity. -/
def optimalTau (lambda r : ℝ) : ℝ := max (Real.tanh r) (stationaryTau lambda r)

/-- PDF p.4: maximizing channel noise. -/
def optimalNoise (lambda r : ℝ) : ℝ := Real.exp (-2 * r) * (1 + optimalTau lambda r)

/-- PDF p.4, (13): CLAIMED closed form, not the definition of an optimum. -/
def optimumClosedForm (lambda r : ℝ) : ℝ :=
  if stationaryTau lambda r ≤ Real.tanh r then
    lambda / (lambda + (1 - Real.sqrt (Real.tanh r)) ^ 2)
  else Real.exp r * (1 + lambda + Real.tanh r) /
    (2 * Real.exp r + lambda * Real.cosh r)

/-- PDF p.3, (9): denominator, used only where strictly positive. -/
def familyDenominator (r tau : ℝ) : ℝ := tau + 1 - Real.exp (2 * r) * |tau - 1|

/-- PDF p.3, (9): the lower b coefficient on the nonsingular domain. -/
def familyMinB (r tau : ℝ) : ℝ :=
  (Real.exp (2 * r) * tau + Real.exp (-2 * r) - |tau - 1|) /
    familyDenominator r tau

/-- PDF p.3, (9): a computed from b, not an independent state hypothesis. -/
def familyA (r tau b : ℝ) : ℝ := (b + Real.exp (-2 * r) * (tau - 1)) / tau

/-- PDF p.3, (9): c computed from b, with source block C=-c sigma_z. -/
def familyC (r tau b : ℝ) : ℝ := (b - Real.exp (-2 * r)) / Real.sqrt tau

/-- PDF p.3: energy within the displayed family; no global energy minimization. -/
def familyEnergy (r tau b : ℝ) : ℝ := (familyA r tau b + b - 2) / 4

/-- PDF p.3: a on the minimum-energy member of (9). -/
def familyMinA (r tau : ℝ) : ℝ := familyA r tau (familyMinB r tau)

/-- PDF p.3: c on the minimum-energy member of (9). -/
def familyMinC (r tau : ℝ) : ℝ := familyC r tau (familyMinB r tau)

/-- PDF p.3: minimum energy WITHIN family (9). -/
def familyMinEnergy (r tau : ℝ) : ℝ := familyEnergy r tau (familyMinB r tau)

/-- PDF p.3: actual pure TMSS coefficients with the source correlation sign.
The alternating sign yields C=-sinh(2r) sigma_z; r is finite. -/
def tmssCoefficient (r : ℝ) (u : ℕ × ℕ) : ℂ :=
  if u.1 = u.2 then (((-Real.tanh r) ^ u.1 / Real.cosh r : ℝ) : ℂ) else 0

/-- PDF p.3: full TMSS rank-one coefficient array; normalization is concluded. -/
def tmssKernel (r : ℝ) : Matrix (ℕ × ℕ) (ℕ × ℕ) ℂ :=
  fun u v => tmssCoefficient r u * star (tmssCoefficient r v)

/-- PDF p.3, (7): independently defined TMSS noise at gain g and squeezing s. -/
def tmssNoise (g s : ℝ) : ℝ :=
  (1 + g ^ 2) * Real.cosh (2 * s) - 2 * g * Real.sinh (2 * s)

/-- PDF p.3, (10): minimum is defined by the noise equation, NOT its answer.
The theorem concludes a least element; no empty real-infimum default is used. -/
def minimumSqueezing (tau y : ℝ) : ℝ :=
  sInf {s : ℝ | 0 ≤ s ∧ tmssNoise (Real.sqrt tau) s = y}

/-- PDF p.3, (10): real arcosh written as log(x+sqrt(x^2-1)). The stated
domain ensures its argument is at least one. tau=1 uses (7) directly. -/
def squeezingClosedForm (tau y : ℝ) : ℝ :=
  if tau = 1 then Real.log (2 / y) / 2 else
    let x := (y * (1 + tau) - 2 * Real.sqrt (tau * (y ^ 2 - (1 - tau) ^ 2))) /
      (1 - tau) ^ 2
    Real.log (x + Real.sqrt (x ^ 2 - 1)) / 2

/-- PDF p.4: heterodyne POVM d^2beta |beta><beta|/pi followed by coherent
preparation |g beta>. The complex overlap retains linearity on trace class. -/
def heterodynePrepare (g : ℝ) : OscillatorChannel :=
  fun X m n => ∫ beta : ℂ,
    (inner (𝕜 := ℂ) (coherentKet beta) ((operatorOfMatrix X) (coherentKet beta)) /
      (Real.pi : ℂ)) * coherentKernel ((g : ℂ) * beta) m n

/-- PDF p.4, (16): optimized positive displacement gain for a fixed TMSS. -/
def tmssOptimalGain (lambda r : ℝ) : ℝ :=
  (2 + lambda * Real.sinh (2 * r)) / (2 + lambda + lambda * Real.cosh (2 * r))

/-- PDF p.4, (16): claimed fixed-TMSS optimum, distinct from its supremum. -/
def tmssClosedForm (lambda r : ℝ) : ℝ :=
  ((1 / Real.cosh r) ^ 2 + lambda) / (2 + lambda - 2 * Real.tanh r)

/-- PDF p.4: actual fidelities with this fixed TMSS, optimized only over g>0. -/
def tmssValues (lambda r : ℝ) : Set ℝ :=
  {f | ∃ (rho : State (ℕ × ℕ)) (g : ℝ) (Phi : OscillatorChannel),
    rho.coefficients = tmssKernel r ∧ IsBK rho g Phi ∧ f = gaussianPriorFidelity lambda Phi}

/-- PDF p.4: fixed-TMSS optimum, defined independently of (16). -/
def tmssOptimum (lambda r : ℝ) : ℝ := sSup (tmssValues lambda r)

/-! Sanity checks pin coordinate order, signs, normalization conventions,
independent optimization domains and the nonsingular vacuum case. -/

example (X : Matrix (ℕ × ℕ) (ℕ × ℕ) ℂ) :
    pairKernel X ![2, 3] ![4, 5] = X (2, 3) (4, 5) := rfl

example (g : ℝ) (z : ℂ) : bkTest g z (0, 1) = Real.sqrt 2 * g * z.re := by
  simp [bkTest]

example (g : ℝ) (z : ℂ) : bkTest g z (1, 1) = -Real.sqrt 2 * z.re := by
  simp [bkTest]

example (rho : State (ℕ × ℕ)) (g : ℝ) (Phi : OscillatorChannel)
    (h : IsBK rho g Phi) : 0 < g := h.1

example (rho : State (ℕ × ℕ)) (V : Cov 2) (h : CenteredResource rho V) :
    V.transpose = V := h.1

example (lambda : ℝ) : priorDensity lambda 0 = lambda / Real.pi := by
  simp [priorDensity]

example : phaseInsensitiveFidelityFormula 1 1 0 = 1 := by norm_num [phaseInsensitiveFidelityFormula]

example (lambda r f : ℝ) : f ∈ scalarValues lambda r ↔
    ∃ p ∈ entanglementAccessibleRegion r, phaseInsensitiveFidelityFormula lambda p.1 p.2 = f := Iff.rfl

example (lambda r : ℝ) : scalarOptimum lambda r = sSup (scalarValues lambda r) := rfl

example : familyDenominator 0 1 = 2 := by norm_num [familyDenominator]

example : familyMinB 0 1 = 1 := by norm_num [familyMinB, familyDenominator]

example : familyMinA 0 1 = 1 := by norm_num [familyMinA, familyA, familyMinB, familyDenominator]

example : familyMinC 0 1 = 0 := by norm_num [familyMinC, familyC, familyMinB, familyDenominator]

example : familyMinEnergy 0 1 = 0 := by
  norm_num [familyMinEnergy, familyEnergy, familyA, familyMinB, familyDenominator]

example : tmssKernel 0 (0, 0) (0, 0) = 1 := by
  simp [tmssKernel, tmssCoefficient]

example (g : ℝ) : tmssNoise g 0 = 1 + g ^ 2 := by simp [tmssNoise]

example (y : ℝ) : squeezingClosedForm 1 y = Real.log (2 / y) / 2 := by
  simp [squeezingClosedForm]

end QAlgorithms.Papers.Independent.Intakethird100.P89
Source
Optimal continuous variable quantum teleportation with limited resources, arXiv:1705.03017v5, https://arxiv.org/abs/1705.03017v5
Read-back

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

QAlgorithms.Papers.Independent.Intakethird100.P89.CenteredResource

Let ρ\rhoρ be a state indexed by N2\mathbb N^2N2, with coefficient matrix RRR: explicitly, RRR admits sequences aj,bj∈ℓ2(N2;C)a_j,b_j\in\ell^2(\mathbb N^2;\mathbb C)aj​,bj​∈ℓ2(N2;C) such that ∑j∥aj∥∥bj∥<∞\sum_j\|a_j\|\|b_j\|<\infty∑j​∥aj​∥∥bj​∥<∞ and Ru,v=∑jaj(u)bj(v)‾R_{u,v}=\sum_j a_j(u)\overline{b_j(v)}Ru,v​=∑j​aj​(u)bj​(v)​; every finite quadratic form ∑u,v∈Fcu‾Ru,vcv\sum_{u,v\in F}\overline{c_u}R_{u,v}c_v∑u,v∈F​cu​​Ru,v​cv​ is real and nonnegative; and the real diagonal entries are summable with sum 111. Let VVV be a real matrix indexed by Q={0,1}×{0,1}Q=\{0,1\}\times\{0,1\}Q={0,1}×{0,1}. The predicate says that VT=VV^{\mathsf T}=VVT=V, VVV is positive definite, and V+iΩV+i\OmegaV+iΩ is positive semidefinite, where Ω(j,0),(j,1)=1\Omega_{(j,0),(j,1)}=1Ω(j,0),(j,1)​=1, Ω(j,1),(j,0)=−1\Omega_{(j,1),(j,0)}=-1Ω(j,1),(j,0)​=−1, and all other entries vanish. It further requires the following for every v∈RQv\in\mathbb R^Qv∈RQ. Put ζj=(−v(j,1)+iv(j,0))/2\zeta_j=(-v_{(j,1)}+iv_{(j,0)})/\sqrt2ζj​=(−v(j,1)​+iv(j,0)​)/2​, and define

Dz(m,n)=e−∣z∣2/2m! n!∑h=0min⁡(m,n)zm−h(−z‾)n−hh!(m−h)!(n−h)!.D_z(m,n)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{h=0}^{\min(m,n)} \frac{z^{m-h}(-\overline z)^{n-h}}{h!(m-h)!(n-h)!}.Dz​(m,n)=e−∣z∣2/2m!n!​h=0∑min(m,n)​h!(m−h)!(n−h)!zm−h(−z)n−h​.

For every a∈N2a\in\mathbb N^2a∈N2, the sum over b∈N2b\in\mathbb N^2b∈N2 below must be summable; the resulting function of aaa must also be summable; and

∑a∈N2∑b∈N2Ra,bDζ0(b0,a0)Dζ1(b1,a1)=exp⁡ ⁣(−14∑q,q′∈QvqVq,q′vq′).\sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} R_{a,b}D_{\zeta_0}(b_0,a_0)D_{\zeta_1}(b_1,a_1) = \exp\!\left(-\frac14\sum_{q,q'\in Q}v_qV_{q,q'}v_{q'}\right).a∈N2∑​b∈N2∑​Ra,b​Dζ0​​(b0​,a0​)Dζ1​​(b1​,a1​)=exp​−41​q,q′∈Q∑​vq​Vq,q′​vq′​​.

No finite photon-energy condition is included in this predicate.

QAlgorithms.Papers.Independent.Intakethird100.P89.FiniteBKRealization

For arbitrary real r,τ,yr,\tau,yr,τ,y and an arbitrary function Φ\PhiΦ on complex matrices indexed by N\mathbb NN, this predicate asserts the existence of a state ρ\rhoρ indexed by N2\mathbb N^2N2, a real matrix VVV indexed by Q={0,1}×{0,1}Q=\{0,1\}\times\{0,1\}Q={0,1}×{0,1}, and a real ggg, with the following conditions. To specify them explicitly, a matrix XXX on a countable index set AAA is trace class if it has an expansion Xa,b=∑jxj(a)yj(b)‾X_{a,b}=\sum_jx_j(a)\overline{y_j(b)}Xa,b​=∑j​xj​(a)yj​(b)​, where xj,yj∈ℓ2(A;C)x_j,y_j\in\ell^2(A;\mathbb C)xj​,yj​∈ℓ2(A;C) and ∑j∥xj∥∥yj∥<∞\sum_j\|x_j\|\|y_j\|<\infty∑j​∥xj​∥∥yj​∥<∞; it is a density matrix if additionally every finite quadratic form is real and nonnegative and its real diagonal is summable with sum 111. A state consists of such a density matrix. Set

tA(X)=inf⁡xj,yj∈ℓ2(A;C)∑j∥xj∥∥yj∥<∞Xa,b=∑jxj(a)yj(b)‾∑j∥xj∥∥yj∥,t_A(X)=\inf_{\substack{x_j,y_j\in\ell^2(A;\mathbb C)\\ \sum_j\|x_j\|\|y_j\|<\infty\\ X_{a,b}=\sum_jx_j(a)\overline{y_j(b)}}} \sum_j\|x_j\|\|y_j\|,tA​(X)=xj​,yj​∈ℓ2(A;C)∑j​∥xj​∥∥yj​∥<∞Xa,b​=∑j​xj​(a)yj​(b)​​inf​j∑​∥xj​∥∥yj​∥,

as an extended nonnegative real number, with empty infimum +∞+\infty+∞, and define

Dz(m,n)=e−∣z∣2/2m! n!∑h=0min⁡(m,n)zm−h(−z‾)n−hh!(m−h)!(n−h)!,χX(z)=∑m∑nXm,nDz(n,m),D_z(m,n)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{h=0}^{\min(m,n)} \frac{z^{m-h}(-\overline z)^{n-h}}{h!(m-h)!(n-h)!}, \quad \chi_X(z)=\sum_m\sum_nX_{m,n}D_z(n,m),Dz​(m,n)=e−∣z∣2/2m!n!​h=0∑min(m,n)​h!(m−h)!(n−h)!zm−h(−z)n−h​,χX​(z)=m∑​n∑​Xm,n​Dz​(n,m), χρ(2)(w0,w1)=∑a∈N2∑b∈N2ρa,bDw0(b0,a0)Dw1(b1,a1).\chi_\rho^{(2)}(w_0,w_1) =\sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} \rho_{a,b}D_{w_0}(b_0,a_0)D_{w_1}(b_1,a_1).χρ(2)​(w0​,w1​)=a∈N2∑​b∈N2∑​ρa,b​Dw0​​(b0​,a0​)Dw1​​(b1​,a1​).

The matrix VVV must be symmetric and positive definite, and V+iΩV+i\OmegaV+iΩ must be positive semidefinite, where Ω\OmegaΩ has the two diagonal blocks (01−10)\left(\begin{smallmatrix}0&1\\-1&0\end{smallmatrix}\right)(0−1​10​). For every v∈RQv\in\mathbb R^Qv∈RQ, putting ζj=(−v(j,1)+iv(j,0))/2\zeta_j=(-v_{(j,1)}+iv_{(j,0)})/\sqrt2ζj​=(−v(j,1)​+iv(j,0)​)/2​, each inner sum and the outer sum defining χρ(2)(ζ0,ζ1)\chi_\rho^{(2)}(\zeta_0,\zeta_1)χρ(2)​(ζ0​,ζ1​) must be summable, and

χρ(2)(ζ0,ζ1)=e−vTVv/4.\chi_\rho^{(2)}(\zeta_0,\zeta_1)=e^{-v^{\mathsf T}Vv/4}.χρ(2)​(ζ0​,ζ1​)=e−vTVv/4.

The resource must satisfy

∑a,b∈Nmax⁡ ⁣{0,a+b2Re⁡ρ(a,b),(a,b)}<∞,tN2 ⁣((ρ(c,b),(a,d))(a,b),(c,d))=e2r.\sum_{a,b\in\mathbb N} \max\!\left\{0,\frac{a+b}{2}\operatorname{Re}\rho_{(a,b),(a,b)}\right\}<\infty, \qquad t_{\mathbb N^2}\!\left( (\rho_{(c,b),(a,d)})_{(a,b),(c,d)} \right)=e^{2r}.a,b∈N∑​max{0,2a+b​Reρ(a,b),(a,b)​}<∞,tN2​((ρ(c,b),(a,d)​)(a,b),(c,d)​)=e2r.

Furthermore, g>0g>0g>0, g2=τg^2=\taug2=τ, and 0≤τ0\le\tau0≤τ. The function Φ\PhiΦ must preserve trace class and be complex linear on trace-class inputs, and for every d∈Nd\in\mathbb Nd∈N its blockwise extension

S⟼(Φ((S(i,a),(j,b))i,j∈N)k,l)(k,a),(l,b)S\longmapsto \left( \Phi\bigl((S_{(i,a),(j,b)})_{i,j\in\mathbb N}\bigr)_{k,l} \right)_{(k,a),(l,b)}S⟼(Φ((S(i,a),(j,b)​)i,j∈N​)k,l​)(k,a),(l,b)​

must send every density matrix on N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1} to a density matrix. Finally, ∣τ−1∣≤y|\tau-1|\le y∣τ−1∣≤y, and for every one-mode state with coefficient matrix SSS and every z∈Cz\in\mathbb Cz∈C, both identities must hold:

χΦ(S)(z)=χS(gz)χρ(2)(−gz‾,z),χΦ(S)(z)=χS(τ z)e−y∣z∣2/2.\chi_{\Phi(S)}(z)=\chi_S(gz)\chi_\rho^{(2)}(-g\overline z,z), \qquad \chi_{\Phi(S)}(z)=\chi_S(\sqrt\tau\,z)e^{-y|z|^2/2}.χΦ(S)​(z)=χS​(gz)χρ(2)​(−gz,z),χΦ(S)​(z)=χS​(τ​z)e−y∣z∣2/2.

The second identity uses the nonnegative-τ\tauτ branch of the defining phase-insensitive condition; the other conditions in fact force τ>0\tau>0τ>0. No sign assumption on rrr or yyy is supplied separately. The ancillary clause for d=0d=0d=0 is vacuous. Apart from the explicitly required summability for the resource characteristic, complex infinite sums are totalized and are zero when not summable.

QAlgorithms.Papers.Independent.Intakethird100.P89.FiniteBKSimulates

For arbitrary r,τ,y∈Rr,\tau,y\in\mathbb Rr,τ,y∈R, this predicate asserts that there exist a function Φ:CN×N→CN×N\Phi:\mathbb C^{\mathbb N\times\mathbb N}\to\mathbb C^{\mathbb N\times\mathbb N}Φ:CN×N→CN×N, a two-mode state ρ\rhoρ, a real symmetric matrix VVV indexed by Q={0,1}×{0,1}Q=\{0,1\}\times\{0,1\}Q={0,1}×{0,1}, and g∈Rg\in\mathbb Rg∈R satisfying the conditions below. Here a matrix XXX on a countable index set AAA is trace class when Xa,b=∑jxj(a)yj(b)‾X_{a,b}=\sum_jx_j(a)\overline{y_j(b)}Xa,b​=∑j​xj​(a)yj​(b)​ for sequences xj,yj∈ℓ2(A;C)x_j,y_j\in\ell^2(A;\mathbb C)xj​,yj​∈ℓ2(A;C) with ∑j∥xj∥∥yj∥<∞\sum_j\|x_j\|\|y_j\|<\infty∑j​∥xj​∥∥yj​∥<∞; a state has a trace-class coefficient matrix whose finite quadratic forms are real and nonnegative and whose real diagonal is summable with sum 111. Define tA(X)t_A(X)tA​(X) to be the infimum of ∑j∥xj∥∥yj∥\sum_j\|x_j\|\|y_j\|∑j​∥xj​∥∥yj​∥ over all such expansions of XXX, with empty infimum +∞+\infty+∞. Define

Dz(m,n)=e−∣z∣2/2m! n!∑h=0min⁡(m,n)zm−h(−z‾)n−hh!(m−h)!(n−h)!,χX(z)=∑m∑nXm,nDz(n,m),D_z(m,n)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{h=0}^{\min(m,n)} \frac{z^{m-h}(-\overline z)^{n-h}}{h!(m-h)!(n-h)!}, \quad \chi_X(z)=\sum_m\sum_nX_{m,n}D_z(n,m),Dz​(m,n)=e−∣z∣2/2m!n!​h=0∑min(m,n)​h!(m−h)!(n−h)!zm−h(−z)n−h​,χX​(z)=m∑​n∑​Xm,n​Dz​(n,m), χρ(2)(w0,w1)=∑a∈N2∑b∈N2ρa,bDw0(b0,a0)Dw1(b1,a1).\chi_\rho^{(2)}(w_0,w_1) =\sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} \rho_{a,b}D_{w_0}(b_0,a_0)D_{w_1}(b_1,a_1).χρ(2)​(w0​,w1​)=a∈N2∑​b∈N2∑​ρa,b​Dw0​​(b0​,a0​)Dw1​​(b1​,a1​).

The matrix VVV must be positive definite and V+iΩV+i\OmegaV+iΩ positive semidefinite, where Ω\OmegaΩ has diagonal blocks (01−10)\left(\begin{smallmatrix}0&1\\-1&0\end{smallmatrix}\right)(0−1​10​). For every v∈RQv\in\mathbb R^Qv∈RQ, with ζj=(−v(j,1)+iv(j,0))/2\zeta_j=(-v_{(j,1)}+iv_{(j,0)})/\sqrt2ζj​=(−v(j,1)​+iv(j,0)​)/2​, each inner sum and the outer sum defining χρ(2)(ζ0,ζ1)\chi_\rho^{(2)}(\zeta_0,\zeta_1)χρ(2)​(ζ0​,ζ1​) must be summable and their value must be e−vTVv/4e^{-v^{\mathsf T}Vv/4}e−vTVv/4. The further resource requirements are

∑a,b∈Nmax⁡ ⁣{0,a+b2Re⁡ρ(a,b),(a,b)}<∞,tN2 ⁣((ρ(c,b),(a,d))(a,b),(c,d))=e2r.\sum_{a,b\in\mathbb N} \max\!\left\{0,\frac{a+b}{2}\operatorname{Re}\rho_{(a,b),(a,b)}\right\}<\infty, \qquad t_{\mathbb N^2}\!\left( (\rho_{(c,b),(a,d)})_{(a,b),(c,d)} \right)=e^{2r}.a,b∈N∑​max{0,2a+b​Reρ(a,b),(a,b)​}<∞,tN2​((ρ(c,b),(a,d)​)(a,b),(c,d)​)=e2r.

One must have g>0g>0g>0, g2=τg^2=\taug2=τ, 0≤τ0\le\tau0≤τ, and ∣τ−1∣≤y|\tau-1|\le y∣τ−1∣≤y. The function Φ\PhiΦ must preserve trace class and be complex linear on trace-class matrices; for every d∈Nd\in\mathbb Nd∈N, applying Φ\PhiΦ to each matrix block (S(i,a),(j,b))i,j∈N(S_{(i,a),(j,b)})_{i,j\in\mathbb N}(S(i,a),(j,b)​)i,j∈N​ of any density matrix SSS on N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1} must produce a density matrix on that same index set. For every one-mode state with coefficient matrix SSS and every z∈Cz\in\mathbb Cz∈C, it must satisfy

χΦ(S)(z)=χS(gz)χρ(2)(−gz‾,z),χΦ(S)(z)=χS(τ z)e−y∣z∣2/2.\chi_{\Phi(S)}(z)=\chi_S(gz)\chi_\rho^{(2)}(-g\overline z,z), \qquad \chi_{\Phi(S)}(z)=\chi_S(\sqrt\tau\,z)e^{-y|z|^2/2}.χΦ(S)​(z)=χS​(gz)χρ(2)​(−gz,z),χΦ(S)​(z)=χS​(τ​z)e−y∣z∣2/2.

In particular, these requirements force τ>0\tau>0τ>0; there is no separate sign assumption on rrr. The ancillary requirement at d=0d=0d=0 is vacuous. Complex infinite sums without an explicit summability requirement are totalized, with value zero when not summable.

QAlgorithms.Papers.Independent.Intakethird100.P89.IsBK

Let ρ\rhoρ be a state indexed by N2\mathbb N^2N2, let g∈Rg\in\mathbb Rg∈R, and let Φ\PhiΦ be an arbitrary function on complex matrices indexed by N\mathbb NN. For clarity, call a matrix XXX on a countable index set AAA trace class when there are xj,yj∈ℓ2(A;C)x_j,y_j\in\ell^2(A;\mathbb C)xj​,yj​∈ℓ2(A;C) with ∑j∥xj∥∥yj∥<∞\sum_j\|x_j\|\|y_j\|<\infty∑j​∥xj​∥∥yj​∥<∞ and Xa,b=∑jxj(a)yj(b)‾X_{a,b}=\sum_jx_j(a)\overline{y_j(b)}Xa,b​=∑j​xj​(a)yj​(b)​; call it a density matrix when it is trace class, every finite quadratic form is real and nonnegative, and its real diagonal is summable with sum 111. A state consists of such a density matrix. Define

Dz(m,n)=e−∣z∣2/2m! n!∑h=0min⁡(m,n)zm−h(−z‾)n−hh!(m−h)!(n−h)!,χX(z)=∑m∑nXm,nDz(n,m).D_z(m,n)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{h=0}^{\min(m,n)} \frac{z^{m-h}(-\overline z)^{n-h}}{h!(m-h)!(n-h)!}, \quad \chi_X(z)=\sum_m\sum_n X_{m,n}D_z(n,m).Dz​(m,n)=e−∣z∣2/2m!n!​h=0∑min(m,n)​h!(m−h)!(n−h)!zm−h(−z)n−h​,χX​(z)=m∑​n∑​Xm,n​Dz​(n,m).

The predicate requires g>0g>0g>0; requires that Φ\PhiΦ preserve trace-class matrices, addition of trace-class matrices, and multiplication of trace-class matrices by every complex scalar; and requires that, for every natural number ddd and every density matrix SSS on N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1}, the matrix

Φ~(S)(j,a),(k,b)=Φ((S(i,a),(l,b))i,l∈N)j,k\widetilde\Phi(S)_{(j,a),(k,b)} =\Phi\bigl((S_{(i,a),(l,b)})_{i,l\in\mathbb N}\bigr)_{j,k}Φ(S)(j,a),(k,b)​=Φ((S(i,a),(l,b)​)i,l∈N​)j,k​

be a density matrix. Finally, for every state σ\sigmaσ indexed by N\mathbb NN, with coefficient matrix SSS, and every z∈Cz\in\mathbb Cz∈C, it requires

χΦ(S)(z)=χS(gz)∑a∈N2∑b∈N2ρa,bD−gz‾(b0,a0)Dz(b1,a1),\chi_{\Phi(S)}(z)=\chi_S(gz) \sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} \rho_{a,b}D_{-g\overline z}(b_0,a_0)D_z(b_1,a_1),χΦ(S)​(z)=χS​(gz)a∈N2∑​b∈N2∑​ρa,b​D−gz​(b0​,a0​)Dz​(b1​,a1​),

where ρa,b\rho_{a,b}ρa,b​ denotes the coefficients of ρ\rhoρ. No summability hypothesis is separately imposed on these characteristic sums; each complex infinite sum is the totalized sum, equal to zero if its summand is not summable. The ancillary quantifier includes d=0d=0d=0, when there are no density matrices on the empty index set and that clause is vacuous.

QAlgorithms.Papers.Independent.Intakethird100.P89.OscillatorChannel

An oscillator channel, in this abbreviation, is an arbitrary function from complex matrices indexed by N×N\mathbb N\times\mathbb NN×N to complex matrices indexed by N×N\mathbb N\times\mathbb NN×N. No linearity, positivity, continuity, normalization, or summability condition is imposed.

QAlgorithms.Papers.Independent.Intakethird100.P89.bkTest

For every real ggg and complex zzz, this defines the real vector indexed by {0,1}×{0,1}\{0,1\}\times\{0,1\}{0,1}×{0,1} with entries

v(0,0)=2 g Im⁡z,v(0,1)=2 g Re⁡z,v(1,0)=2 Im⁡z,v(1,1)=−2 Re⁡z.v_{(0,0)}=\sqrt2\,g\,\operatorname{Im}z,\qquad v_{(0,1)}=\sqrt2\,g\,\operatorname{Re}z,\qquad v_{(1,0)}=\sqrt2\,\operatorname{Im}z,\qquad v_{(1,1)}=-\sqrt2\,\operatorname{Re}z.v(0,0)​=2​gImz,v(0,1)​=2​gRez,v(1,0)​=2​Imz,v(1,1)​=−2​Rez.

There is no restriction on ggg, including its sign or whether it is zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.entanglementAccessibleRegion

For every real number rrr, this defines the set of real pairs (τ,y)(\tau,y)(τ,y) satisfying

τ≥0,y≥∣1−τ∣,y≥e−2r(1+τ).\tau\ge 0,\qquad y\ge |1-\tau|,\qquad y\ge e^{-2r}(1+\tau).τ≥0,y≥∣1−τ∣,y≥e−2r(1+τ).

There is no restriction on rrr, no upper bound on either coordinate, and τ=0\tau=0τ=0 is included when the two bounds on yyy hold.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyA

For arbitrary real numbers r,τ,br,\tau,br,τ,b, this defines

b+e−2r(τ−1)τ.\frac{b+e^{-2r}(\tau-1)}{\tau}.τb+e−2r(τ−1)​.

In particular, τ=0\tau=0τ=0 is allowed and gives the value 000, regardless of rrr and bbb.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyC

For arbitrary real numbers r,τ,br,\tau,br,τ,b, this defines the real number

b−e−2rτ.\frac{b-e^{-2r}}{\sqrt{\tau}}.τ​b−e−2r​.

The real square root is zero when τ≤0\tau\le 0τ≤0, and division by zero is defined to give zero, so this value is zero for every τ≤0\tau\le 0τ≤0.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyDenominator

For arbitrary real numbers r,τr,\taur,τ, this defines

τ+1−e2r∣τ−1∣.\tau+1-e^{2r}|\tau-1|.τ+1−e2r∣τ−1∣.

No sign or nonvanishing condition is imposed on this expression.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyEnergy

For arbitrary real numbers r,τ,br,\tau,br,τ,b, this defines the real number

14(b+e−2r(τ−1)τ+b−2).\frac14\left(\frac{b+e^{-2r}(\tau-1)}{\tau}+b-2\right).41​(τb+e−2r(τ−1)​+b−2).

No sign restrictions are imposed. Division by zero gives zero; in particular, at τ=0\tau=0τ=0 the value is (b−2)/4(b-2)/4(b−2)/4.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyMinA

For arbitrary real numbers r,τr,\taur,τ, put

B=e2rτ+e−2r−∣τ−1∣τ+1−e2r∣τ−1∣.B=\frac{e^{2r}\tau+e^{-2r}-|\tau-1|} {\tau+1-e^{2r}|\tau-1|}.B=τ+1−e2r∣τ−1∣e2rτ+e−2r−∣τ−1∣​.

This defines the real number

B+e−2r(τ−1)τ.\frac{B+e^{-2r}(\tau-1)}{\tau}.τB+e−2r(τ−1)​.

Every division by zero gives zero, including the division defining BBB if its denominator vanishes; at τ=0\tau=0τ=0, the defined value is zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyMinB

For arbitrary real numbers r,τr,\taur,τ, this defines

e2rτ+e−2r−∣τ−1∣τ+1−e2r∣τ−1∣.\frac{e^{2r}\tau+e^{-2r}-|\tau-1|} {\tau+1-e^{2r}|\tau-1|}.τ+1−e2r∣τ−1∣e2rτ+e−2r−∣τ−1∣​.

There are no parameter restrictions, and the value is 000 when the denominator vanishes. The definition does not assert a minimum property.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyMinC

For arbitrary real numbers r,τr,\taur,τ, this defines the real number

1τ(e2rτ+e−2r−∣τ−1∣τ+1−e2r∣τ−1∣−e−2r).\frac{1}{\sqrt{\tau}} \left( \frac{e^{2r}\tau+e^{-2r}-|\tau-1|} {\tau+1-e^{2r}|\tau-1|} -e^{-2r} \right).τ​1​(τ+1−e2r∣τ−1∣e2rτ+e−2r−∣τ−1∣​−e−2r).

Every division by zero gives zero. Since the real square root is zero for τ≤0\tau\le0τ≤0, the defined value is zero throughout that range.

QAlgorithms.Papers.Independent.Intakethird100.P89.familyMinEnergy

For arbitrary real numbers r,τr,\taur,τ, put

B=e2rτ+e−2r−∣τ−1∣τ+1−e2r∣τ−1∣.B=\frac{e^{2r}\tau+e^{-2r}-|\tau-1|} {\tau+1-e^{2r}|\tau-1|}.B=τ+1−e2r∣τ−1∣e2rτ+e−2r−∣τ−1∣​.

This defines the real number

14(B+e−2r(τ−1)τ+B−2).\frac14\left(\frac{B+e^{-2r}(\tau-1)}{\tau}+B-2\right).41​(τB+e−2r(τ−1)​+B−2).

There are no sign or nonvanishing assumptions, and every division by zero gives zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.finiteBKOptimum

For arbitrary real numbers λ,r\lambda,rλ,r, this defines the real supremum of the following set of values fff. Write HI=ℓ2(I;C)H_I=\ell^2(I;\mathbb C)HI​=ℓ2(I;C). For each index set III used here, call XXX trace class when it has an expansion Xab=∑juj(a)vj(b)‾X_{ab}=\sum_j u_j(a)\overline{v_j(b)}Xab​=∑j​uj​(a)vj​(b)​ with uj,vj∈HIu_j,v_j\in H_Iuj​,vj​∈HI​ and summable costs ∥uj∥∥vj∥\|u_j\|\|v_j\|∥uj​∥∥vj​∥; let t(X)∈[0,∞]t(X)\in[0,\infty]t(X)∈[0,∞] be the infimum of the sums of those costs over all such expansions, with empty infimum +∞+\infty+∞. Call XXX a density matrix when it is trace class, every finite quadratic form ∑a,b∈Sza‾Xabzb\sum_{a,b\in S}\overline{z_a}X_{ab}z_b∑a,b∈S​za​​Xab​zb​, for finite S⊆IS\subseteq IS⊆I and arbitrary z:I→Cz:I\to\mathbb Cz:I→C, is real and nonnegative, and its real diagonal is summable with sum 111. A value fff belongs to the set precisely when there exist τ,y∈R\tau,y\in\mathbb Rτ,y∈R, a map Φ\PhiΦ on all complex matrices indexed by N\mathbb NN, a density matrix ρ\rhoρ indexed by N2\mathbb N^2N2, a real 4×44\times44×4 matrix VVV, and g>0g>0g>0, with these properties: Φ\PhiΦ preserves trace class and is additive and complex homogeneous on trace-class inputs; for every natural ddd and every density matrix XXX indexed by N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1}, the matrix with entries Φ((X(i,a),(j,b))ij)mn\Phi((X_{(i,a),(j,b)})_{ij})_{mn}Φ((X(i,a),(j,b)​)ij​)mn​, indexed by (m,a),(n,b)(m,a),(n,b)(m,a),(n,b), is a density matrix. Define

D(z)mn=e−∣z∣2/2m! n!∑j=0min⁡(m,n)zm−j(−z‾)n−jj!(m−j)!(n−j)!,χX(z)=∑m∑nXmnD(z)nm.D(z)_{mn} =e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{j=0}^{\min(m,n)} \frac{z^{m-j}(-\overline z)^{n-j}}{j!(m-j)!(n-j)!}, \qquad \chi_X(z)=\sum_m\sum_n X_{mn}D(z)_{nm}.D(z)mn​=e−∣z∣2/2m!n!​j=0∑min(m,n)​j!(m−j)!(n−j)!zm−j(−z)n−j​,χX​(z)=m∑​n∑​Xmn​D(z)nm​.

With coordinates ordered as (0,0),(0,1),(1,0),(1,1)(0,0),(0,1),(1,0),(1,1)(0,0),(0,1),(1,0),(1,1), require VT=VV^{\mathsf T}=VVT=V, positive definiteness of VVV, and positive semidefiniteness of V+idiag⁡(J,J)V+i\operatorname{diag}(J,J)V+idiag(J,J), where J=(01−10)J=\begin{pmatrix}0&1\\-1&0\end{pmatrix}J=(0−1​10​). For every real coordinate vector vvv, putting wj=(−vj,1+ivj,0)/2w_j=(-v_{j,1}+i v_{j,0})/\sqrt2wj​=(−vj,1​+ivj,0​)/2​, require every inner series and the outer series in

Fρ(v)=∑a∈N2∑b∈N2ρabD(w0)b0a0D(w1)b1a1F_\rho(v)=\sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} \rho_{ab}D(w_0)_{b_0a_0}D(w_1)_{b_1a_1}Fρ​(v)=a∈N2∑​b∈N2∑​ρab​D(w0​)b0​a0​​D(w1​)b1​a1​​

to be summable and require Fρ(v)=e−vTVv/4F_\rho(v)=e^{-v^{\mathsf T}Vv/4}Fρ​(v)=e−vTVv/4. Also require

∑amax⁡ ⁣{0,a0+a12Re⁡ρaa}<∞,t((ρ(b0,a1),(a0,b1))ab)=e2r,g2=τ,τ≥0,∣τ−1∣≤y.\sum_a\max\!\left\{0,\frac{a_0+a_1}{2}\operatorname{Re}\rho_{aa}\right\}<\infty, \qquad t\bigl((\rho_{(b_0,a_1),(a_0,b_1)})_{ab}\bigr)=e^{2r}, \qquad g^2=\tau,\quad \tau\ge0,\quad |\tau-1|\le y.a∑​max{0,2a0​+a1​​Reρaa​}<∞,t((ρ(b0​,a1​),(a0​,b1​)​)ab​)=e2r,g2=τ,τ≥0,∣τ−1∣≤y.

For every density matrix σ\sigmaσ indexed by N\mathbb NN and every z∈Cz\in\mathbb Cz∈C, require

χΦ(σ)(z)=χσ(gz)Fρ(2g Im⁡z,2g Re⁡z,2 Im⁡z,−2 Re⁡z)=χσ(τ z)e−y∣z∣2/2;\chi_{\Phi(\sigma)}(z) =\chi_\sigma(gz) F_\rho(\sqrt2g\,\operatorname{Im}z,\sqrt2g\,\operatorname{Re}z, \sqrt2\,\operatorname{Im}z,-\sqrt2\,\operatorname{Re}z) =\chi_\sigma(\sqrt{\tau}\,z)e^{-y|z|^2/2};χΦ(σ)​(z)=χσ​(gz)Fρ​(2​gImz,2​gRez,2​Imz,−2​Rez)=χσ​(τ​z)e−y∣z∣2/2;

the definition's alternative argument −∣τ∣ z‾-\sqrt{|\tau|}\,\overline z−∣τ∣​z applies only when τ<0\tau<0τ<0, which these requirements exclude. Let cα(n)=e−∣α∣2/2αn/n!c_\alpha(n)=e^{-|\alpha|^2/2}\alpha^n/\sqrt{n!}cα​(n)=e−∣α∣2/2αn/n!​, Rαmn=cα(m)cα(n)‾R_\alpha{}_{mn}=c_\alpha(m)\overline{c_\alpha(n)}Rα​mn​=cα​(m)cα​(n)​, let hαh_\alphahα​ be the chosen vector in HNH_{\mathbb N}HN​ with these coordinates if one exists and 000 otherwise, and let B(X)B(X)B(X) be a chosen bounded complex-linear operator with coordinate matrix XXX if one exists and 000 otherwise. The value associated with these witnesses is

f=∫Cλπe−λ∣α∣2Re⁡⟨hα,B(Φ(Rα))hα⟩ dαf=\int_{\mathbb C}\frac{\lambda}{\pi}e^{-\lambda|\alpha|^2} \operatorname{Re}\langle h_\alpha,B(\Phi(R_\alpha))h_\alpha\rangle\,d\alphaf=∫C​πλ​e−λ∣α∣2Re⟨hα​,B(Φ(Rα​))hα​⟩dα

with Lebesgue measure. Neither λ\lambdaλ nor rrr is restricted, and this last integral is not assumed integrable. Infinite sums and integrals are total, giving 000 on nonsummable or nonintegrable inputs unless the relevant property is explicitly required. The finite amplification condition includes d=0d=0d=0, where there are no density matrices to test; g>0g>0g>0 excludes τ=0\tau=0τ=0. The real supremum defining the result is 000 if this set is empty or unbounded above, and no attainment is asserted.

QAlgorithms.Papers.Independent.Intakethird100.P89.finiteBKValues

For arbitrary real numbers λ\lambdaλ and rrr, this defines the set of real numbers fff for which there exist real numbers τ,y\tau,yτ,y, a map Φ\PhiΦ from all complex matrices indexed by N\mathbb NN to such matrices, a density matrix ρ\rhoρ indexed by N2\mathbb N^2N2, a real 4×44\times44×4 matrix VVV, and a real g>0g>0g>0, satisfying the conditions below. Here, for any index set III occurring in these conditions, write HI=ℓ2(I;C)H_I=\ell^2(I;\mathbb C)HI​=ℓ2(I;C); a matrix XXX is called trace class precisely when there are sequences uj,vj∈HIu_j,v_j\in H_Iuj​,vj​∈HI​ such that ∑j∥uj∥∥vj∥\sum_j\|u_j\|\|v_j\|∑j​∥uj​∥∥vj​∥ is summable and Xab=∑juj(a)vj(b)‾X_{ab}=\sum_j u_j(a)\overline{v_j(b)}Xab​=∑j​uj​(a)vj​(b)​ for every a,ba,ba,b. Write t(X)t(X)t(X) for the infimum, over all such expansions, of their costs ∑j∥uj∥∥vj∥\sum_j\|u_j\|\|v_j\|∑j​∥uj​∥∥vj​∥, viewed in [0,∞][0,\infty][0,∞]; an empty infimum is +∞+\infty+∞. A density matrix means a trace-class matrix XXX such that, for every finite S⊆IS\subseteq IS⊆I and every z:I→Cz:I\to\mathbb Cz:I→C, ∑a,b∈Sza‾Xabzb\sum_{a,b\in S}\overline{z_a}X_{ab}z_b∑a,b∈S​za​​Xab​zb​ is real and nonnegative, and the real diagonal is summable with ∑aRe⁡Xaa=1\sum_a\operatorname{Re}X_{aa}=1∑a​ReXaa​=1. The map Φ\PhiΦ must preserve trace class, preserve sums of trace-class matrices and complex scalar multiples of trace-class matrices, and, for every natural number ddd and every density matrix XXX on N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1}, the matrix

Φ~d(X)(m,a),(n,b)=Φ((X(i,a),(j,b))i,j∈N)mn\widetilde\Phi_d(X)_{(m,a),(n,b)} =\Phi\bigl((X_{(i,a),(j,b)})_{i,j\in\mathbb N}\bigr)_{mn}Φd​(X)(m,a),(n,b)​=Φ((X(i,a),(j,b)​)i,j∈N​)mn​

must be a density matrix; the d=0d=0d=0 condition has no density matrices to test. To specify the remaining conditions, let

D(z)mn=e−∣z∣2/2m! n!∑j=0min⁡(m,n)zm−j(−z‾)n−jj!(m−j)!(n−j)!,χX(z)=∑m∑nXmnD(z)nm.D(z)_{mn} =e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{j=0}^{\min(m,n)} \frac{z^{m-j}(-\overline z)^{n-j}}{j!(m-j)!(n-j)!}, \qquad \chi_X(z)=\sum_m\sum_n X_{mn}D(z)_{nm}.D(z)mn​=e−∣z∣2/2m!n!​j=0∑min(m,n)​j!(m−j)!(n−j)!zm−j(−z)n−j​,χX​(z)=m∑​n∑​Xmn​D(z)nm​.

Index the coordinates of VVV by (j,k)∈{0,1}2(j,k)\in\{0,1\}^2(j,k)∈{0,1}2, put Ω=diag⁡(J,J)\Omega=\operatorname{diag}(J,J)Ω=diag(J,J) with J=(01−10)J=\begin{pmatrix}0&1\\-1&0\end{pmatrix}J=(0−1​10​), and, for v∈R{0,1}2v\in\mathbb R^{\{0,1\}^2}v∈R{0,1}2, put wj=(−vj,1+ivj,0)/2w_j=(-v_{j,1}+i v_{j,0})/\sqrt2wj​=(−vj,1​+ivj,0​)/2​. The matrix VVV must be symmetric and positive definite, V+iΩV+i\OmegaV+iΩ must be positive semidefinite, and

Fρ(v):=∑a∈N2∑b∈N2ρabD(w0)b0a0D(w1)b1a1=e−vTVv/4F_\rho(v):= \sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} \rho_{ab}D(w_0)_{b_0a_0}D(w_1)_{b_1a_1} =e^{-v^{\mathsf T}Vv/4}Fρ​(v):=a∈N2∑​b∈N2∑​ρab​D(w0​)b0​a0​​D(w1​)b1​a1​​=e−vTVv/4

must hold for every vvv, with every inner series and then the outer series summable. In addition,

∑a∈N2max⁡ ⁣{0,a0+a12Re⁡ρaa}<∞,t(ρΓ)=e2r,ρabΓ=ρ(b0,a1),(a0,b1).\sum_{a\in\mathbb N^2}\max\!\left\{0,\frac{a_0+a_1}{2}\operatorname{Re}\rho_{aa}\right\}<\infty, \qquad t(\rho^\Gamma)=e^{2r}, \qquad \rho^\Gamma_{ab}=\rho_{(b_0,a_1),(a_0,b_1)}.a∈N2∑​max{0,2a0​+a1​​Reρaa​}<∞,t(ρΓ)=e2r,ρabΓ​=ρ(b0​,a1​),(a0​,b1​)​.

One must have g2=τg^2=\taug2=τ, τ≥0\tau\ge0τ≥0, and ∣τ−1∣≤y|\tau-1|\le y∣τ−1∣≤y. For every density matrix σ\sigmaσ indexed by N\mathbb NN and every z∈Cz\in\mathbb Cz∈C, both identities

χΦ(σ)(z)=χσ(gz)Fρ(bg(z)),χΦ(σ)(z)=χσ ⁣({τ z,τ≥0,−∣τ∣ z‾,τ<0)e−y∣z∣2/2\chi_{\Phi(\sigma)}(z)=\chi_\sigma(gz)F_\rho(b_g(z)), \qquad \chi_{\Phi(\sigma)}(z)= \chi_\sigma\!\left( \begin{cases} \sqrt{\tau}\,z,&\tau\ge0,\\ -\sqrt{|\tau|}\,\overline z,&\tau<0 \end{cases}\right)e^{-y|z|^2/2}χΦ(σ)​(z)=χσ​(gz)Fρ​(bg​(z)),χΦ(σ)​(z)=χσ​({τ​z,−∣τ∣​z,​τ≥0,τ<0​)e−y∣z∣2/2

are required, where bg(z)=(2g Im⁡z,2g Re⁡z,2 Im⁡z,−2 Re⁡z)b_g(z)=(\sqrt2g\,\operatorname{Im}z,\sqrt2g\,\operatorname{Re}z,\sqrt2\,\operatorname{Im}z,-\sqrt2\,\operatorname{Re}z)bg​(z)=(2​gImz,2​gRez,2​Imz,−2​Rez) in the stated coordinate order; the negative-τ\tauτ branch is excluded by the preceding requirements. Finally, set cα(n)=e−∣α∣2/2αn/n!c_\alpha(n)=e^{-|\alpha|^2/2}\alpha^n/\sqrt{n!}cα​(n)=e−∣α∣2/2αn/n!​, Rαmn=cα(m)cα(n)‾R_\alpha{}_{mn}=c_\alpha(m)\overline{c_\alpha(n)}Rα​mn​=cα​(m)cα​(n)​, let hα∈HNh_\alpha\in H_{\mathbb N}hα​∈HN​ be the chosen vector with coordinates cα(n)c_\alpha(n)cα​(n) when such a vector exists and 000 otherwise, and let B(X)B(X)B(X) be a chosen bounded complex-linear operator on HNH_{\mathbb N}HN​ whose matrix in the coordinate unit vectors is XXX, or 000 if none exists. The required value of fff is

f=∫Cλπe−λ∣α∣2Re⁡⟨hα,B(Φ(Rα))hα⟩ dα,f=\int_{\mathbb C}\frac{\lambda}{\pi}e^{-\lambda|\alpha|^2} \operatorname{Re}\langle h_\alpha,B(\Phi(R_\alpha))h_\alpha\rangle\,d\alpha,f=∫C​πλ​e−λ∣α∣2Re⟨hα​,B(Φ(Rα​))hα​⟩dα,

using Lebesgue measure. There is no positivity assumption on λ\lambdaλ or rrr, no integrability hypothesis on this last integrand, and no assertion that witnesses exist; in particular, g>0g>0g>0 excludes τ=0\tau=0τ=0. The integrals and infinite sums use their total definitions, giving 000 when the corresponding integrability or summability fails, except where those properties are explicitly required.

QAlgorithms.Papers.Independent.Intakethird100.P89.gaussianPriorFidelity

For any real λ\lambdaλ and any function Φ\PhiΦ on complex matrices indexed by N\mathbb NN, this defines

∫Cλπe−λ∣α∣2Re⁡⟨uα,Aαuα⟩ dα,\int_{\mathbb C} \frac{\lambda}{\pi}e^{-\lambda|\alpha|^2} \operatorname{Re}\langle u_\alpha,A_\alpha u_\alpha\rangle \,d\alpha,∫C​πλ​e−λ∣α∣2Re⟨uα​,Aα​uα​⟩dα,

where dαd\alphadα is Lebesgue measure on C\mathbb CC, cα(n)=e−∣α∣2/2αn/n!c_\alpha(n)=e^{-|\alpha|^2/2}\alpha^n/\sqrt{n!}cα​(n)=e−∣α∣2/2αn/n!​, and uαu_\alphauα​ is a chosen vector in ℓ2(N;C)\ell^2(\mathbb N;\mathbb C)ℓ2(N;C) with these coordinates if one exists, otherwise zero. Here AαA_\alphaAα​ is a chosen bounded complex-linear operator on that space with matrix entries (Aαen)m=Φ(Kα)m,n(A_\alpha e_n)_m=\Phi(K_\alpha)_{m,n}(Aα​en​)m​=Φ(Kα​)m,n​, where Kα(m,n)=cα(m)cα(n)‾K_\alpha(m,n)=c_\alpha(m)\overline{c_\alpha(n)}Kα​(m,n)=cα​(m)cα​(n)​, if such an operator exists, otherwise zero. No positivity assumption on λ\lambdaλ, channel condition on Φ\PhiΦ, or integrability hypothesis is imposed. The integral is the totalized real Bochner integral, equal to zero when the integrand is not integrable; it is also zero at λ=0\lambda=0λ=0.

QAlgorithms.Papers.Independent.Intakethird100.P89.heterodynePrepare

For every real number ggg, this defines a function from arbitrary complex matrices X=(Xmn)m,n∈NX=(X_{mn})_{m,n\in\mathbb N}X=(Xmn​)m,n∈N​ to complex matrices indexed by N\mathbb NN. Let H=ℓ2(N;C)H=\ell^2(\mathbb N;\mathbb C)H=ℓ2(N;C), let ene_nen​ be its coordinate unit vectors, and let AXA_XAX​ be a chosen bounded complex-linear operator on HHH satisfying (AXen)m=Xmn(A_Xe_n)_m=X_{mn}(AX​en​)m​=Xmn​ for all m,nm,nm,n, if such an operator exists, and the zero operator otherwise. For each β∈C\beta\in\mathbb Cβ∈C, let hβh_\betahβ​ be a chosen vector in HHH with coordinates e−∣β∣2/2βj/j!e^{-|\beta|^2/2}\beta^j/\sqrt{j!}e−∣β∣2/2βj/j!​ for every j∈Nj\in\mathbb Nj∈N, if such a vector exists, and the zero vector otherwise. The output entry at any m,n∈Nm,n\in\mathbb Nm,n∈N is the complex Bochner integral

∫C⟨hβ,AXhβ⟩π(e−∣gβ∣2/2(gβ)mm!)(e−∣gβ∣2/2(gβ)nn!)‾ dβ,\int_{\mathbb C} \frac{\langle h_\beta,A_Xh_\beta\rangle}{\pi} \left(\frac{e^{-|g\beta|^2/2}(g\beta)^m}{\sqrt{m!}}\right) \overline{\left(\frac{e^{-|g\beta|^2/2}(g\beta)^n}{\sqrt{n!}}\right)} \,d\beta,∫C​π⟨hβ​,AX​hβ​⟩​(m!​e−∣gβ∣2/2(gβ)m​)(n!​e−∣gβ∣2/2(gβ)n​)​dβ,

with respect to Lebesgue volume on C\mathbb CC. No density, positivity, or trace-class assumption is imposed on XXX, and ggg is unrestricted, including zero; the integral is defined to be zero if its integrand is not integrable.

QAlgorithms.Papers.Independent.Intakethird100.P89.minimumSqueezing

For arbitrary real numbers τ,y\tau,yτ,y, this defines the real infimum

inf⁡{s∈R:s≥0,(1+(τ)2)cosh⁡(2s)−2τsinh⁡(2s)=y}.\inf\left\{s\in\mathbb R: s\ge0,\quad (1+(\sqrt{\tau})^2)\cosh(2s) -2\sqrt{\tau}\sinh(2s)=y \right\}.inf{s∈R:s≥0,(1+(τ​)2)cosh(2s)−2τ​sinh(2s)=y}.

Here τ=0\sqrt{\tau}=0τ​=0 when τ≤0\tau\le0τ≤0. The set of admissible sss may be empty, in which case the real infimum used here is defined to be zero; the definition does not assert that an admissible sss exists or attains the infimum.

QAlgorithms.Papers.Independent.Intakethird100.P89.optimalNoise

For arbitrary real numbers λ,r\lambda,rλ,r, this defines

e−2r(1+max⁡{tanh⁡r,e2r(er+λcosh⁡r)2}),e^{-2r}\left(1+ \max\left\{\tanh r,\frac{e^{2r}}{(e^r+\lambda\cosh r)^2}\right\}\right),e−2r(1+max{tanhr,(er+λcoshr)2e2r​}),

with the fraction interpreted as 000 when its denominator is zero. There are no hypotheses restricting λ\lambdaλ or rrr.

QAlgorithms.Papers.Independent.Intakethird100.P89.optimalTau

For arbitrary real numbers λ,r\lambda,rλ,r, this defines

max⁡{tanh⁡r,e2r(er+λcosh⁡r)2},\max\left\{\tanh r,\frac{e^{2r}}{(e^r+\lambda\cosh r)^2}\right\},max{tanhr,(er+λcoshr)2e2r​},

where a zero denominator makes the fraction 000. The definition imposes no parameter restrictions and asserts no optimization property.

QAlgorithms.Papers.Independent.Intakethird100.P89.optimumClosedForm

For arbitrary real numbers λ,r\lambda,rλ,r, this defines the real number

{λλ+(1−tanh⁡r)2,e2r(er+λcosh⁡r)2≤tanh⁡r,er(1+λ+tanh⁡r)2er+λcosh⁡r,e2r(er+λcosh⁡r)2>tanh⁡r.\begin{cases} \displaystyle \frac{\lambda}{\lambda+(1-\sqrt{\tanh r})^2}, & \displaystyle\frac{e^{2r}}{(e^r+\lambda\cosh r)^2}\le\tanh r,\\[8pt] \displaystyle \frac{e^r(1+\lambda+\tanh r)}{2e^r+\lambda\cosh r}, & \displaystyle\frac{e^{2r}}{(e^r+\lambda\cosh r)^2}>\tanh r. \end{cases}⎩⎨⎧​λ+(1−tanhr​)2λ​,2er+λcoshrer(1+λ+tanhr)​,​(er+λcoshr)2e2r​≤tanhr,(er+λcoshr)2e2r​>tanhr.​

Equality in the comparison selects the first branch. All fractions, including the fraction used in that comparison, are 000 when their denominator is zero; the real square root is 000 on negative arguments. No restrictions on λ,r\lambda,rλ,r, or equality with an independently defined optimum, are asserted.

QAlgorithms.Papers.Independent.Intakethird100.P89.pairKernel

For any complex matrix XXX whose row and column indices are pairs of natural numbers, this defines a complex matrix indexed instead by functions from {0,1}\{0,1\}{0,1} to N\mathbb NN, by

Ku,v=X(u0,u1),(v0,v1).K_{u,v}=X_{(u_0,u_1),(v_0,v_1)}.Ku,v​=X(u0​,u1​),(v0​,v1​)​.

Thus it changes the indexing of the entries. The input XXX is arbitrary; neither XXX nor the resulting matrix is required to be a density matrix.

QAlgorithms.Papers.Independent.Intakethird100.P89.phaseInsensitiveFidelityFormula

For every three real numbers λ,τ,y\lambda,\tau,yλ,τ,y, this defines

2λ 2(1−τ)2+λ(1+y+τ) .\frac{2\lambda} {\,2(1-\sqrt\tau)^2+\lambda(1+y+\tau)\,}.2(1−τ​)2+λ(1+y+τ)2λ​.

No parameter restrictions are imposed. The real square root is zero for τ≤0\tau\le0τ≤0, and division by a zero denominator has value zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.priorDensity

For every real λ\lambdaλ and complex α\alphaα, this defines the real number

pλ(α)=λπe−λ∣α∣2.p_\lambda(\alpha)=\frac{\lambda}{\pi}e^{-\lambda|\alpha|^2}.pλ​(α)=πλ​e−λ∣α∣2.

There is no positivity assumption on λ\lambdaλ. At λ=0\lambda=0λ=0 this function is identically zero; for λ<0\lambda<0λ<0 it is negative everywhere.

QAlgorithms.Papers.Independent.Intakethird100.P89.scalarOptimum

For arbitrary real numbers λ\lambdaλ and rrr, this defines the real supremum of

{2λ2(1−τ)2+λ(1+y+τ)  |  τ,y∈R,τ≥0,∣1−τ∣≤y,e−2r(1+τ)≤y},\left\{ \frac{2\lambda} {2(1-\sqrt{\tau})^2+\lambda(1+y+\tau)} \;\middle|\; \tau,y\in\mathbb R,\quad \tau\ge0,\quad |1-\tau|\le y,\quad e^{-2r}(1+\tau)\le y \right\},{2(1−τ​)2+λ(1+y+τ)2λ​​τ,y∈R,τ≥0,∣1−τ∣≤y,e−2r(1+τ)≤y},

where division by zero gives 000. This set is nonempty, since one may take τ=0\tau=0τ=0 and y=max⁡{1,e−2r}y=\max\{1,e^{-2r}\}y=max{1,e−2r}. The real supremum operation returns 000 if the set is unbounded above; this definition does not assert that a largest value is attained.

QAlgorithms.Papers.Independent.Intakethird100.P89.scalarValues

For arbitrary real numbers λ\lambdaλ and rrr, this defines the set

{2λ2(1−τ)2+λ(1+y+τ)  |  τ,y∈R,τ≥0,∣1−τ∣≤y,e−2r(1+τ)≤y}.\left\{ \frac{2\lambda} {2(1-\sqrt{\tau})^2+\lambda(1+y+\tau)} \;\middle|\; \tau,y\in\mathbb R,\quad \tau\ge0,\quad |1-\tau|\le y,\quad e^{-2r}(1+\tau)\le y \right\}.{2(1−τ​)2+λ(1+y+τ)2λ​​τ,y∈R,τ≥0,∣1−τ∣≤y,e−2r(1+τ)≤y}.

Neither λ\lambdaλ nor rrr is assumed positive. Division is total: a value with zero denominator is 000.

QAlgorithms.Papers.Independent.Intakethird100.P89.squeezingClosedForm

For arbitrary real numbers τ,y\tau,yτ,y, this defines

{12log⁡(2/y),τ=1,12log⁡ ⁣(x+x2−1),τ≠1,x=y(1+τ)−2τ(y2−(1−τ)2)(1−τ)2in the second branch.\begin{cases} \dfrac12\log(2/y),&\tau=1,\\[5pt] \dfrac12\log\!\left(x+\sqrt{x^2-1}\right),&\tau\ne1, \end{cases} \qquad x=\frac{y(1+\tau)-2\sqrt{\tau\bigl(y^2-(1-\tau)^2\bigr)}} {(1-\tau)^2} \quad\text{in the second branch}.⎩⎨⎧​21​log(2/y),21​log(x+x2−1​),​τ=1,τ=1,​x=(1−τ)2y(1+τ)−2τ(y2−(1−τ)2)​​in the second branch.

No domain restrictions on τ,y\tau,yτ,y are assumed. Real square roots of negative numbers give zero, division by zero gives zero, and the real logarithm satisfies log⁡0=0\log0=0log0=0 and log⁡x=log⁡∣x∣\log x=\log|x|logx=log∣x∣ for nonzero negative xxx; in particular, the first branch gives zero when y=0y=0y=0.

QAlgorithms.Papers.Independent.Intakethird100.P89.stationaryTau

For arbitrary real numbers λ,r\lambda,rλ,r, this defines

e2r(er+λcosh⁡r)2.\frac{e^{2r}}{(e^r+\lambda\cosh r)^2}.(er+λcoshr)2e2r​.

No positivity assumptions are imposed on either parameter. If er+λcosh⁡r=0e^r+\lambda\cosh r=0er+λcoshr=0, the value is 000 by total division.

QAlgorithms.Papers.Independent.Intakethird100.P89.teleportedCoherentOverlap

For any function Φ\PhiΦ on complex matrices indexed by N\mathbb NN and any α∈C\alpha\in\mathbb Cα∈C, put

cα(n)=e−∣α∣2/2αnn!,Kα(m,n)=cα(m)cα(n)‾.c_\alpha(n)=e^{-|\alpha|^2/2}\frac{\alpha^n}{\sqrt{n!}}, \qquad K_\alpha(m,n)=c_\alpha(m)\overline{c_\alpha(n)}.cα​(n)=e−∣α∣2/2n!​αn​,Kα​(m,n)=cα​(m)cα​(n)​.

Let uα∈ℓ2(N;C)u_\alpha\in\ell^2(\mathbb N;\mathbb C)uα​∈ℓ2(N;C) be a chosen vector with coordinates cα(n)c_\alpha(n)cα​(n), if such a vector exists, and otherwise let uα=0u_\alpha=0uα​=0. Let AAA be a chosen bounded complex-linear operator on ℓ2(N;C)\ell^2(\mathbb N;\mathbb C)ℓ2(N;C) satisfying (Aen)m=Φ(Kα)m,n(Ae_n)_m=\Phi(K_\alpha)_{m,n}(Aen​)m​=Φ(Kα​)m,n​ for all m,nm,nm,n, if such an operator exists, and otherwise let A=0A=0A=0; here ene_nen​ is the vector equal to 111 at nnn and zero elsewhere. The definition returns

Re⁡⟨uα,Auα⟩.\operatorname{Re}\langle u_\alpha,A u_\alpha\rangle.Re⟨uα​,Auα​⟩.

No channel or density condition on Φ\PhiΦ is assumed, and in particular the operator-selection fallback can make the value zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssClosedForm

For arbitrary real numbers λ\lambdaλ and rrr, this defines the real number

cosh⁡(r)−2+λ2+λ−2tanh⁡r.\frac{\cosh(r)^{-2}+\lambda}{2+\lambda-2\tanh r}.2+λ−2tanhrcosh(r)−2+λ​.

There are no positivity assumptions on either argument. Division is total: if the denominator is zero, the defined value is zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssCoefficient

For every real number rrr and every pair u=(u1,u2)u=(u_1,u_2)u=(u1​,u2​) of natural numbers, this defines the complex number

{(−tanh⁡r)u1cosh⁡r,u1=u2,0,u1≠u2,\begin{cases} \dfrac{(-\tanh r)^{u_1}}{\cosh r},&u_1=u_2,\\[4pt] 0,&u_1\ne u_2, \end{cases}⎩⎨⎧​coshr(−tanhr)u1​​,0,​u1​=u2​,u1​=u2​,​

where the displayed real quotient is embedded into C\mathbb CC. The indices may be zero, with the zeroth power equal to 111, including when its base is zero; cosh⁡r\cosh rcoshr is positive for every real rrr.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssKernel

For every real number rrr, this defines a complex matrix indexed by pairs of natural numbers. For u=(u1,u2)u=(u_1,u_2)u=(u1​,u2​) and v=(v1,v2)v=(v_1,v_2)v=(v1​,v2​), its entry is

{(−tanh⁡r)u1+v1cosh⁡2r,u1=u2 and v1=v2,0,otherwise,\begin{cases} \dfrac{(-\tanh r)^{u_1+v_1}}{\cosh^2 r}, &u_1=u_2\ \text{and}\ v_1=v_2,\\[4pt] 0,&\text{otherwise}, \end{cases}⎩⎨⎧​cosh2r(−tanhr)u1​+v1​​,0,​u1​=u2​ and v1​=v2​,otherwise,​

with the real value embedded into C\mathbb CC. This is the coefficient at uuu multiplied by the complex conjugate of the coefficient at vvv; both coefficients are real. Zero indices are included, and zeroth powers equal 111.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssNoise

For arbitrary real numbers g,sg,sg,s, this defines the real number

(1+g2)cosh⁡(2s)−2gsinh⁡(2s).(1+g^2)\cosh(2s)-2g\sinh(2s).(1+g2)cosh(2s)−2gsinh(2s).

Neither argument is required to be positive or nonnegative.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssOptimalGain

For arbitrary real numbers λ,r\lambda,rλ,r, this defines the real number

2+λsinh⁡(2r)2+λ+λcosh⁡(2r).\frac{2+\lambda\sinh(2r)} {2+\lambda+\lambda\cosh(2r)}.2+λ+λcosh(2r)2+λsinh(2r)​.

No positivity assumptions are imposed on either argument. If the denominator vanishes, the defined value is zero.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssOptimum

For arbitrary real λ,r\lambda,rλ,r, this defines the real supremum of the following set of fidelities. Write HL=ℓ2(L;C)H_L=\ell^2(L;\mathbb C)HL​=ℓ2(L;C); a trace-class kernel means a kernel admitting Xij=∑kuk(i)vk(j)‾X_{ij}=\sum_k u_k(i)\overline{v_k(j)}Xij​=∑k​uk​(i)vk​(j)​ with uk,vk∈HLu_k,v_k\in H_Luk​,vk​∈HL​ and ∑k∥uk∥∥vk∥<∞\sum_k\|u_k\|\|v_k\|<\infty∑k​∥uk​∥∥vk​∥<∞, and a density kernel additionally has real nonnegative finite quadratic forms, a summable real diagonal, and real diagonal sum one. Range over density kernels RRR on N2\mathbb N^2N2 with

Rab=tr(a)tr(b)‾,tr(a)=1a0=a1(−tanh⁡r)a0cosh⁡r,R_{ab}=t_r(a)\overline{t_r(b)},\qquad t_r(a)=\mathbf1_{a_0=a_1}\frac{(-\tanh r)^{a_0}}{\cosh r},Rab​=tr​(a)tr​(b)​,tr​(a)=1a0​=a1​​coshr(−tanhr)a0​​,

positive real gains ggg, and functions Φ\PhiΦ on complex kernels on N\mathbb NN. Require Φ\PhiΦ to preserve trace-class kernels and to be additive and complex homogeneous on them; also require every finite ampliation, including size zero, to preserve density kernels, where the ampliation acts by applying Φ\PhiΦ to the first-coordinate kernel with the ancillary coordinates fixed. The size-zero requirement has no density inputs. With

Dmn(z)=e−∣z∣2/2m!n!∑j=0min⁡(m,n)zm−j(−z‾)n−jj!(m−j)!(n−j)!,χX(z)=∑m∑nXmnDnm(z),D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!n!} \sum_{j=0}^{\min(m,n)} \frac{z^{m-j}(-\overline z)^{n-j}}{j!(m-j)!(n-j)!}, \qquad \chi_X(z)=\sum_m\sum_n X_{mn}D_{nm}(z),Dmn​(z)=e−∣z∣2/2m!n!​j=0∑min(m,n)​j!(m−j)!(n−j)!zm−j(−z)n−j​,χX​(z)=m∑​n∑​Xmn​Dnm​(z),

require for every one-mode density kernel SSS and z∈Cz\in\mathbb Cz∈C that

χΦ(S)(z)=χS(gz)∑a∑bRabDb0a0(−gz‾)Db1a1(z).\chi_{\Phi(S)}(z)=\chi_S(gz) \sum_a\sum_b R_{ab}D_{b_0a_0}(-g\overline z)D_{b_1a_1}(z).χΦ(S)​(z)=χS​(gz)a∑​b∑​Rab​Db0​a0​​(−gz)Db1​a1​​(z).

Set cα(n)=e−∣α∣2/2αn/n!c_\alpha(n)=e^{-|\alpha|^2/2}\alpha^n/\sqrt{n!}cα​(n)=e−∣α∣2/2αn/n!​ and Pα(m,n)=cα(m)cα(n)‾P_\alpha(m,n)=c_\alpha(m)\overline{c_\alpha(n)}Pα​(m,n)=cα​(m)cα​(n)​; let hαh_\alphahα​ be a chosen square-summable vector with those coefficients, or zero if none exists, and let O(X)O(X)O(X) be a chosen bounded operator with basis matrix XXX, or zero if none exists. The values whose supremum is taken are

∫Cλπe−λ∣α∣2Re⁡⟨hα,O(Φ(Pα))hα⟩ dα.\int_{\mathbb C}\frac{\lambda}{\pi}e^{-\lambda|\alpha|^2} \operatorname{Re}\langle h_\alpha,O(\Phi(P_\alpha))h_\alpha\rangle\,d\alpha.∫C​πλ​e−λ∣α∣2Re⟨hα​,O(Φ(Pα​))hα​⟩dα.

All sums and integrals are total operations, so no convergence or integrability assumptions are added by this definition. This is a supremum in R\mathbb RR, not an extended-real supremum: its value is zero if the set is empty or is not bounded above.

QAlgorithms.Papers.Independent.Intakethird100.P89.tmssValues

For arbitrary real numbers λ,r\lambda,rλ,r, this defines the set of real numbers fff obtained as follows. Put HL=ℓ2(L;C)H_L=\ell^2(L;\mathbb C)HL​=ℓ2(L;C). A kernel XXX on an index set LLL is called trace class here when it has an expansion Xij=∑k≥0uk(i)vk(j)‾X_{ij}=\sum_{k\ge0}u_k(i)\overline{v_k(j)}Xij​=∑k≥0​uk​(i)vk​(j)​, where uk,vk∈HLu_k,v_k\in H_Luk​,vk​∈HL​ and ∑k∥uk∥∥vk∥<∞\sum_k\|u_k\|\|v_k\|<\infty∑k​∥uk​∥∥vk​∥<∞; it is a density kernel when, additionally, every finite quadratic form ∑i,j∈Szi‾Xijzj\sum_{i,j\in S}\overline{z_i}X_{ij}z_j∑i,j∈S​zi​​Xij​zj​ is real and nonnegative, its real diagonal is summable, and ∑iRe⁡Xii=1\sum_i\operatorname{Re}X_{ii}=1∑i​ReXii​=1. Choose a density kernel RRR on N2\mathbb N^2N2, a real g>0g>0g>0, and a function Φ\PhiΦ from complex kernels on N\mathbb NN to themselves. Require

Rab=tr(a)tr(b)‾,tr(a)={(−tanh⁡r)a0/cosh⁡r,a0=a1,0,a0≠a1.R_{ab}=t_r(a)\overline{t_r(b)},\qquad t_r(a)= \begin{cases} (-\tanh r)^{a_0}/\cosh r,&a_0=a_1,\\ 0,&a_0\ne a_1. \end{cases}Rab​=tr​(a)tr​(b)​,tr​(a)={(−tanhr)a0​/coshr,0,​a0​=a1​,a0​=a1​.​

Require Φ\PhiΦ to preserve trace-class kernels, to be additive and complex homogeneous on them, and, for every integer d≥0d\ge0d≥0, to send every density kernel XXX on N×{0,…,d−1}\mathbb N\times\{0,\ldots,d-1\}N×{0,…,d−1} to a density kernel under the entrywise ampliation

(Φ⊗id)(X)(i,a),(j,b)=Φ((X(k,a),(ℓ,b))k,ℓ)ij.(\Phi\otimes\mathrm{id})(X)_{(i,a),(j,b)} =\Phi\bigl((X_{(k,a),(\ell,b)})_{k,\ell}\bigr)_{ij}.(Φ⊗id)(X)(i,a),(j,b)​=Φ((X(k,a),(ℓ,b)​)k,ℓ​)ij​.

The d=0d=0d=0 condition has no density inputs. Define

Dmn(z)=e−∣z∣2/2m!n!∑j=0min⁡(m,n)zm−j(−z‾)n−jj!(m−j)!(n−j)!,χX(z)=∑m∑nXmnDnm(z).D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!n!} \sum_{j=0}^{\min(m,n)} \frac{z^{m-j}(-\overline z)^{n-j}}{j!(m-j)!(n-j)!}, \quad \chi_X(z)=\sum_m\sum_n X_{mn}D_{nm}(z).Dmn​(z)=e−∣z∣2/2m!n!​j=0∑min(m,n)​j!(m−j)!(n−j)!zm−j(−z)n−j​,χX​(z)=m∑​n∑​Xmn​Dnm​(z).

Require, for every density kernel SSS on N\mathbb NN and every z∈Cz\in\mathbb Cz∈C,

χΦ(S)(z)=χS(gz)∑a∈N2∑b∈N2RabDb0a0(−gz‾)Db1a1(z).\chi_{\Phi(S)}(z)=\chi_S(gz) \sum_{a\in\mathbb N^2}\sum_{b\in\mathbb N^2} R_{ab}D_{b_0a_0}(-g\overline z)D_{b_1a_1}(z).χΦ(S)​(z)=χS​(gz)a∈N2∑​b∈N2∑​Rab​Db0​a0​​(−gz)Db1​a1​​(z).

For α∈C\alpha\in\mathbb Cα∈C, let cα(n)=e−∣α∣2/2αn/n!c_\alpha(n)=e^{-|\alpha|^2/2}\alpha^n/\sqrt{n!}cα​(n)=e−∣α∣2/2αn/n!​, let Pα(m,n)=cα(m)cα(n)‾P_\alpha(m,n)=c_\alpha(m)\overline{c_\alpha(n)}Pα​(m,n)=cα​(m)cα​(n)​, and let hα∈HNh_\alpha\in H_{\mathbb N}hα​∈HN​ be a chosen vector with coefficients cα(n)c_\alpha(n)cα​(n) if one exists, and zero otherwise. For a kernel XXX, let O(X)O(X)O(X) be a chosen bounded complex linear operator on HNH_{\mathbb N}HN​ whose basis matrix is XXX, if such an operator exists, and zero otherwise. Finally require

f=∫Cλπe−λ∣α∣2Re⁡⟨hα,O(Φ(Pα))hα⟩ dα.f=\int_{\mathbb C}\frac{\lambda}{\pi}e^{-\lambda|\alpha|^2} \operatorname{Re}\langle h_\alpha,O(\Phi(P_\alpha))h_\alpha\rangle\,d\alpha.f=∫C​πλ​e−λ∣α∣2Re⟨hα​,O(Φ(Pα​))hα​⟩dα.

The integral uses planar Lebesgue measure. No integrability hypothesis or restriction on λ,r\lambda,rλ,r is imposed in this definition; the integral is the total Bochner integral, which is zero for a nonintegrable integrand. The infinite sums likewise use the total sum operation, assigning zero to a nonsummable family.

QAlgorithms.Papers.Independent.Intakethird100.P89.twoModeMeanPhotonEnergy

For a state ρ\rhoρ indexed by N2\mathbb N^2N2, with coefficient matrix RRR, this defines the extended nonnegative real number

∑(a,b)∈N2max⁡ ⁣{0,a+b2Re⁡R(a,b),(a,b)}.\sum_{(a,b)\in\mathbb N^2} \max\!\left\{0,\frac{a+b}{2}\operatorname{Re}R_{(a,b),(a,b)}\right\}.(a,b)∈N2∑​max{0,2a+b​ReR(a,b),(a,b)​}.

Here being a state means that RRR has an expansion Ru,v=∑jxj(u)yj(v)‾R_{u,v}=\sum_j x_j(u)\overline{y_j(v)}Ru,v​=∑j​xj​(u)yj​(v)​ with xj,yj∈ℓ2(N2;C)x_j,y_j\in\ell^2(\mathbb N^2;\mathbb C)xj​,yj​∈ℓ2(N2;C) and ∑j∥xj∥∥yj∥<∞\sum_j\|x_j\|\|y_j\|<\infty∑j​∥xj​∥∥yj​∥<∞, all its finite quadratic forms are real and nonnegative, and its real diagonal is summable with sum 111. In particular its diagonal entries are nonnegative real numbers, so the displayed truncation does not change the terms. The value may be +∞+\infty+∞; the (0,0)(0,0)(0,0) term is zero.

Human review
  • Endorsed by Shuze Chen · Nov 1, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Nov 1, 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