Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Extremality of Gaussian quantum states

Definition
QAlgorithms_Papers_Independent_Support50_P14

by Goku · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

The objects the formalized statements of Extremality of Gaussian quantum states (arXiv:quant-ph/0509154) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • OscillatorKernel: PDF p.1: coefficients on the completed oscillator Hilbert space.
  • InterleavedQuadrature: PDF p.1: interleaved canonical coordinates, second coordinate 0=Q, 1=P.
  • OscillatorCovariance: PDF p.1: real symmetric covariance matrices are required in CompleteMoments.
  • oscillatorWeylKernel: PDF p.1, (1): bounded Weyl kernel exp(i xi.R), with hbar=1.
  • oscillatorCharacteristic: PDF p.1: trace-class/bounded pairing, multiplication before diagonal trace.
  • OscillatorCharacteristicSummable: Genuine convergence guards for the two iterated sums in the Weyl pairing.
  • meanForm: PDF p.1: xi dot d.
  • varianceForm: PDF p.1: the full covariance quadratic form.
  • photonEnergy: PDF p.1: extended mean photon number, on positive densities. The finite zero-point term does not affect finiteness.
  • CompleteMoments: PDF p.1, moments preceding (1): finite oscillator energy and all directional first/second derivatives of the actual characteristic function. Symmetry is explicit: directional quadratic forms alone cannot determine an antisymmetric part. For a density these derivatives are precisely the finite symmetric quadrature moments, including the subtraction of d d^T from the second moment.
  • GaussianAt: PDF p.1, (1): Gaussianity of this density with these actual moments. The density condition is supplied by State, never by a covariance-only surrogate.
  • InDomain: PDF p.2: no cap means all densities. A finite cap e bounds mean photon number by e times the number of modes. These compatible compact sublevels scale on tensor powers; the centering and passive copy mixers used in the proof and their single-copy marginals remain within the respective sublevels.
  • MultipartiteOscillatorFunctional: PDF pp.1--2: the functional is a family BEFORE choosing any input. Arbitrary finite mode labels and ownership maps support copies and regrouping. Values on arrays outside the specified density domain are unused.
  • oscillatorTensor: PDF p.2: the actual tensor product, regrouped into A+B oscillator modes.
  • marginalFirst: PDF p.2: trace out precisely the B modes, retaining all A coherences.
  • marginalSecond: PDF p.2: trace out precisely the A modes.
  • localKernel: PDF pp.1--3: independent bounded unitaries at all parties, allowing arbitrary mixing of all modes/copies held by that party. A party's Hilbert space is the completed l2 of its entire occupation basis.
  • StronglySuperadditive: PDF p.2, definition of strong super-additivity: BOTH the inequality on correlated inputs and equality on arbitrary product densities. Restricted-domain clauses are used only when all states in the clause belong to that domain.
  • LocallyInvariant: PDF pp.1--2: invariance under independent local bounded unitaries and ownership-preserving mode relabelings. No Gaussification or entropy conclusion is included among these purely algebraic hypotheses.
  • OscillatorTraceContinuous: PDF p.2: epsilon-delta continuity in full TRACE norm, relative to the normalized density domain.
  • TraceLowerSemicontinuous: PDF p.2: lower semicontinuity, with open strict superlevel sets in the relative trace-norm topology. This is distinct from continuity.
  • hadamardEntry: PDF p.2, (5): H=((1,1),(1,-1)) tensor m, in binary tensor order.
  • mixedArgument: PDF p.2, (5)--(7): the contravariant Weyl argument of the real passive mixer.
  • copiesKernel: PDF p.2, (3): actual independent copies of the given density, regrouped by mode.
  • copyMixerKernel: PDF p.2: identical copy mixer at each oscillator site.
  • insertCopy: Reinsert the retained copy and every discarded copy at their actual labels.
  • copyMarginal: PDF p.2: computed marginal, tracing out all copy labels except k.
  • positiveEntries: PDF p.2, (6)--(8): positive entries, either all or half the row.
  • counterexampleVector: PDF p.3, (iv): the explicit normalized Schmidt vector on TWO complete modes.
  • counterexampleState: PDF p.3, (iv): actual rank-one density coefficients, not its covariance.
  • negativityExampleCovariance: PDF p.3: the explicit state's covariance in the p.1 vacuum-I convention.
  • transposeFirst: PDF p.3: partial transpose of the first complete oscillator mode.
  • logarithmicNegativity: PDF p.3: E_N=log_2 ||rho^(T_A)||_1. Every use below concludes finiteness and norm>=1, before converting the extended norm to a real logarithm.
Definition code
import Mathlib
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P33
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P39
import Definitions.Def_QAlgorithms_Papers_Independent_Frontier30_P14
import Definitions.Def_QAlgorithms_Papers_Independent_Frontier18_P07
import Definitions.Def_QAlgorithms_Papers_Independent_Claimed50_P47
import Definitions.Def_QAlgorithms_Papers_Independent_Gap48_P20

/-!
Wolf--Giedke--Cirac, arXiv:quant-ph/0509154v1, PDF pp.1--4.
Full oscillator Hilbert spaces have countable occupation bases, without a cutoff.
Coordinates (mode,0),(mode,1) are interleaved Q,P; covariance is the full
anticommutator, vacuum I. No Gaussian-limit or entropy identity is an axiom of
any functional. Operational applications not yet represented are in HARD.md.
Compilation is pending the separate batch handoff.
-/

noncomputable section

open Filter
open scoped BigOperators ENNReal Topology

attribute [local instance] Classical.propDecidable Classical.decEq

namespace QAlgorithms.Papers.Independent.Support50.P14

open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33
  (State IsDensity IsTraceClass traceSize)
open QAlgorithms.Papers.Independent.Dependency50.P39 (displacementMatrix)
open QAlgorithms.Papers.Independent.Frontier30.P14
  (unitaryCoeff conjugateKernel boundedUnitaryConjugation)
open QAlgorithms.Papers.Independent.Frontier18.P07 (extendedEntropyNats)
open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation)
open QAlgorithms.Papers.Independent.Gap48.P20 (Copies)

/-!
Reuse audit (repair): Occupation is the accepted full oscillator label type;
Copies Bool m is the accepted register-power type specialized to binary labels.
No local aliases redeclare either one. The remaining descriptive names below
retain the previous bodies and distinguish incompatible accepted short names:
P32.Kernel is a classical box, P10.Phase uses grouped Fin n indices,
P04.Covariance fixes Fin n, Information100.P34.weyl is a finite clock/shift,
P39.characteristic takes a one-mode bounded operator, P19.CharacteristicSummable
fixes Fin n and one Weyl argument, P30.Functional has finite-dimensional inputs,
Information100.P21.tensor takes two vectors, Next46.P27.TraceContinuous maps INTO
one-mode operators, and Gap48.P13.counterexampleCovariance is a different complex
six-by-six matrix. ASSESSMENT.json records each exact declaration separately.
The defective P47.HasBipartiteMoments is never used.
-/

/-- PDF p.1: coefficients on the completed oscillator Hilbert space. -/
abbrev OscillatorKernel (I : Type) := Matrix (Occupation I) (Occupation I) ℂ

/-- PDF p.1: interleaved canonical coordinates, second coordinate 0=Q, 1=P. -/
abbrev InterleavedQuadrature (I : Type) := I × Fin 2

/-- PDF p.1: real symmetric covariance matrices are required in CompleteMoments. -/
abbrev OscillatorCovariance (I : Type) := Matrix (InterleavedQuadrature I) (InterleavedQuadrature I) ℝ

/-- PDF p.1, (1): bounded Weyl kernel exp(i xi.R), with hbar=1. -/
def oscillatorWeylKernel {I : Type} [Fintype I] (xi : InterleavedQuadrature I → ℝ) : OscillatorKernel I :=
  fun u v => ∏ j, displacementMatrix
    ((-(xi (j, 1) : ℂ) + Complex.I * (xi (j, 0) : ℂ)) / (Real.sqrt 2 : ℂ))
    (u j) (v j)

/-- PDF p.1: trace-class/bounded pairing, multiplication before diagonal trace. -/
def oscillatorCharacteristic {I : Type} [Fintype I] (rho : OscillatorKernel I) (xi : InterleavedQuadrature I → ℝ) : ℂ :=
  ∑' u, ∑' v, rho u v * oscillatorWeylKernel xi v u

/-- Genuine convergence guards for the two iterated sums in the Weyl pairing. -/
def OscillatorCharacteristicSummable {I : Type} [Fintype I] (rho : OscillatorKernel I) : Prop :=
  ∀ xi : InterleavedQuadrature I → ℝ,
    (∀ u, Summable (fun v => rho u v * oscillatorWeylKernel xi v u)) ∧
    Summable (fun u => ∑' v, rho u v * oscillatorWeylKernel xi v u)

/-- PDF p.1: xi dot d. -/
def meanForm {I : Type} [Fintype I] (d xi : InterleavedQuadrature I → ℝ) : ℝ :=
  ∑ a, xi a * d a

/-- PDF p.1: the full covariance quadratic form. -/
def varianceForm {I : Type} [Fintype I] (gamma : OscillatorCovariance I) (xi : InterleavedQuadrature I → ℝ) : ℝ :=
  ∑ a, ∑ b, xi a * gamma a b * xi b

/-- PDF p.1: extended mean photon number, on positive densities.
The finite zero-point term does not affect finiteness. -/
def photonEnergy {I : Type} [Fintype I] (rho : OscillatorKernel I) : ℝ≥0∞ :=
  ∑' u : Occupation I, ENNReal.ofReal (∑ j, (u j : ℝ)) * ENNReal.ofReal (rho u u).re

/-- PDF p.1, moments preceding (1): finite oscillator energy and all directional
first/second derivatives of the actual characteristic function. Symmetry is
explicit: directional quadratic forms alone cannot determine an antisymmetric
part. For a density these derivatives are precisely the finite symmetric
quadrature moments, including the subtraction of d d^T from the second moment. -/
def CompleteMoments {I : Type} [Fintype I] (rho : OscillatorKernel I)
    (d : InterleavedQuadrature I → ℝ) (gamma : OscillatorCovariance I) : Prop :=
  photonEnergy rho < ⊤ ∧ gamma.transpose = gamma ∧ OscillatorCharacteristicSummable rho ∧
  ∀ xi : InterleavedQuadrature I → ℝ,
    HasDerivAt (fun t : ℝ => oscillatorCharacteristic rho (t • xi))
      (Complex.I * (meanForm d xi : ℂ)) 0 ∧
    HasDerivAt (deriv (fun t : ℝ => oscillatorCharacteristic rho (t • xi)))
      (((-varianceForm gamma xi / 2 - (meanForm d xi) ^ 2 : ℝ) : ℂ)) 0

/-- PDF p.1, (1): Gaussianity of this density with these actual moments.
The density condition is supplied by State, never by a covariance-only surrogate. -/
def GaussianAt {I : Type} [Fintype I] (rho : OscillatorKernel I)
    (d : InterleavedQuadrature I → ℝ) (gamma : OscillatorCovariance I) : Prop :=
  CompleteMoments rho d gamma ∧ ∀ xi : InterleavedQuadrature I → ℝ,
    oscillatorCharacteristic rho xi = Complex.exp
      (((-varianceForm gamma xi / 4 : ℝ) : ℂ) + Complex.I * (meanForm d xi : ℂ))

/-- PDF p.2: no cap means all densities. A finite cap e bounds mean photon
number by e times the number of modes. These compatible compact sublevels
scale on tensor powers; the centering and passive copy mixers used in the proof
and their single-copy marginals remain within the respective sublevels. -/
def InDomain {I : Type} [Fintype I] (cap : Option ℝ) (rho : OscillatorKernel I) : Prop :=
  IsDensity rho ∧ match cap with
    | none => True
    | some e => photonEnergy rho ≤ ENNReal.ofReal (e * Fintype.card I)

/-- PDF pp.1--2: the functional is a family BEFORE choosing any input.
Arbitrary finite mode labels and ownership maps support copies and regrouping.
Values on arrays outside the specified density domain are unused. -/
abbrev MultipartiteOscillatorFunctional (M : ℕ) :=
  (I : Type) → [Fintype I] → (I → Fin M) → OscillatorKernel I → ℝ

/-- PDF p.2: the actual tensor product, regrouped into A+B oscillator modes. -/
def oscillatorTensor {A B : Type} (rho : OscillatorKernel A) (sigma : OscillatorKernel B) : OscillatorKernel (A ⊕ B) :=
  fun u v => rho (u ∘ Sum.inl) (v ∘ Sum.inl) * sigma (u ∘ Sum.inr) (v ∘ Sum.inr)

/-- PDF p.2: trace out precisely the B modes, retaining all A coherences. -/
def marginalFirst {A B : Type} (rho : OscillatorKernel (A ⊕ B)) : OscillatorKernel A :=
  fun u v => ∑' w : Occupation B, rho (Sum.elim u w) (Sum.elim v w)

/-- PDF p.2: trace out precisely the A modes. -/
def marginalSecond {A B : Type} (rho : OscillatorKernel (A ⊕ B)) : OscillatorKernel B :=
  fun u v => ∑' w : Occupation A, rho (Sum.elim w u) (Sum.elim w v)

/-- PDF pp.1--3: independent bounded unitaries at all parties, allowing arbitrary
mixing of all modes/copies held by that party. A party's Hilbert space is the
completed l2 of its entire occupation basis. -/
def localKernel {M : ℕ} {I : Type} [Fintype I] (owner : I → Fin M)
    (U : (p : Fin M) → Hilbert (Occupation {j : I // owner j = p}) ≃ₗᵢ[ℂ]
      Hilbert (Occupation {j : I // owner j = p})) : OscillatorKernel I :=
  fun u v => ∏ p, unitaryCoeff (U p) (fun j => u j.val) (fun j => v j.val)

/-- PDF p.2, definition of strong super-additivity: BOTH the inequality on
correlated inputs and equality on arbitrary product densities. Restricted-domain
clauses are used only when all states in the clause belong to that domain. -/
def StronglySuperadditive {M : ℕ} (cap : Option ℝ) (f : MultipartiteOscillatorFunctional M) : Prop :=
  (∀ (A B : Type) [Fintype A] [Fintype B]
    (a : A → Fin M) (b : B → Fin M) (rho : OscillatorKernel (A ⊕ B)),
    InDomain cap rho → InDomain cap (marginalFirst rho) →
    InDomain cap (marginalSecond rho) →
    f A a (marginalFirst rho) + f B b (marginalSecond rho) ≤
      f (A ⊕ B) (Sum.elim a b) rho) ∧
  (∀ (A B : Type) [Fintype A] [Fintype B]
    (a : A → Fin M) (b : B → Fin M) (rho : OscillatorKernel A) (sigma : OscillatorKernel B),
    InDomain cap rho → InDomain cap sigma → InDomain cap (oscillatorTensor rho sigma) →
    f (A ⊕ B) (Sum.elim a b) (oscillatorTensor rho sigma) = f A a rho + f B b sigma)

/-- PDF pp.1--2: invariance under independent local bounded unitaries and
ownership-preserving mode relabelings. No Gaussification or entropy conclusion
is included among these purely algebraic hypotheses. -/
def LocallyInvariant {M : ℕ} (cap : Option ℝ) (f : MultipartiteOscillatorFunctional M) : Prop :=
  (∀ (I : Type) [Fintype I] (owner : I → Fin M) (rho : OscillatorKernel I)
    (U : (p : Fin M) → Hilbert (Occupation {j : I // owner j = p}) ≃ₗᵢ[ℂ]
      Hilbert (Occupation {j : I // owner j = p})),
    InDomain cap rho → InDomain cap (conjugateKernel (localKernel owner U) rho) →
    f I owner (conjugateKernel (localKernel owner U) rho) = f I owner rho) ∧
  (∀ (I J : Type) [Fintype I] [Fintype J] (e : I ≃ J)
    (owner : J → Fin M) (rho : OscillatorKernel I), InDomain cap rho →
    InDomain cap (fun u v => rho (u ∘ e) (v ∘ e)) →
    f J owner (fun u v => rho (u ∘ e) (v ∘ e)) = f I (owner ∘ e) rho)

/-- PDF p.2: epsilon-delta continuity in full TRACE norm, relative to the
normalized density domain. -/
def OscillatorTraceContinuous {M : ℕ} (cap : Option ℝ) (f : MultipartiteOscillatorFunctional M) : Prop :=
  ∀ (I : Type) [Fintype I] (owner : I → Fin M) (rho : OscillatorKernel I),
    InDomain cap rho → ∀ epsilon : ℝ, 0 < epsilon →
    ∃ delta : ℝ, 0 < delta ∧ ∀ sigma : OscillatorKernel I, InDomain cap sigma →
      traceSize (sigma - rho) < ENNReal.ofReal delta →
      |f I owner sigma - f I owner rho| < epsilon

/-- PDF p.2: lower semicontinuity, with open strict superlevel sets in the
relative trace-norm topology. This is distinct from continuity. -/
def TraceLowerSemicontinuous {M : ℕ} (cap : Option ℝ) (f : MultipartiteOscillatorFunctional M) : Prop :=
  ∀ (I : Type) [Fintype I] (owner : I → Fin M) (rho : OscillatorKernel I),
    InDomain cap rho → ∀ a : ℝ, a < f I owner rho →
    ∃ delta : ℝ, 0 < delta ∧ ∀ sigma : OscillatorKernel I, InDomain cap sigma →
      traceSize (sigma - rho) < ENNReal.ofReal delta → a < f I owner sigma

/-- PDF p.2, (5): H=((1,1),(1,-1)) tensor m, in binary tensor order. -/
def hadamardEntry (m : ℕ) (k l : Copies Bool m) : ℝ :=
  (-1 : ℝ) ^ (∑ j : Fin m, if k j = true ∧ l j = true then 1 else 0)

/-- PDF p.2, (5)--(7): the contravariant Weyl argument of the real passive mixer. -/
def mixedArgument (m : ℕ) (xi : InterleavedQuadrature (Copies Bool m) → ℝ) : InterleavedQuadrature (Copies Bool m) → ℝ :=
  fun l => ∑ k : Copies Bool m, hadamardEntry m k l.1 /
    Real.sqrt ((2 : ℝ) ^ m) * xi (k, l.2)

/-- PDF p.2, (3): actual independent copies of the given density, regrouped by mode. -/
def copiesKernel {I : Type} (m : ℕ) (rho : OscillatorKernel I) : OscillatorKernel (I × Copies Bool m) :=
  fun u v => ∏ k : Copies Bool m, rho (fun j => u (j, k)) (fun j => v (j, k))

/-- PDF p.2: identical copy mixer at each oscillator site. -/
def copyMixerKernel {I : Type} [Fintype I] (m : ℕ)
    (U : Hilbert (Occupation (Copies Bool m)) ≃ₗᵢ[ℂ] Hilbert (Occupation (Copies Bool m))) :
    OscillatorKernel (I × Copies Bool m) :=
  fun u v => ∏ j, unitaryCoeff U (fun k => u (j, k)) (fun k => v (j, k))

/-- Reinsert the retained copy and every discarded copy at their actual labels. -/
def insertCopy {I : Type} {m : ℕ} (k : Copies Bool m) (u : Occupation I)
    (rest : Occupation (I × {l : Copies Bool m // l ≠ k})) : Occupation (I × Copies Bool m) :=
  fun p => if h : p.2 = k then u p.1 else rest (p.1, ⟨p.2, h⟩)

/-- PDF p.2: computed marginal, tracing out all copy labels except k. -/
def copyMarginal {I : Type} {m : ℕ} (rho : OscillatorKernel (I × Copies Bool m)) (k : Copies Bool m) : OscillatorKernel I :=
  fun u v => ∑' rest : Occupation (I × {l : Copies Bool m // l ≠ k}),
    rho (insertCopy k u rest) (insertCopy k v rest)

/-- PDF p.2, (6)--(8): positive entries, either all or half the row. -/
def positiveEntries (m : ℕ) (k : Copies Bool m) : ℕ :=
  if k = (fun _ => false) then 2 ^ m else 2 ^ m / 2

/-- PDF p.3, (iv): the explicit normalized Schmidt vector on TWO complete modes. -/
def counterexampleVector (u : Occupation (Fin 2)) : ℂ :=
  if u = (fun _ => 0) then (Real.sqrt 15 / 4 : ℝ)
  else if u = (fun _ => 1) then (1 / 4 : ℝ) else 0

/-- PDF p.3, (iv): actual rank-one density coefficients, not its covariance. -/
def counterexampleState : OscillatorKernel (Fin 2) :=
  fun u v => counterexampleVector u * star (counterexampleVector v)

/-- PDF p.3: the explicit state's covariance in the p.1 vacuum-I convention. -/
def negativityExampleCovariance : OscillatorCovariance (Fin 2) :=
  fun a b => if a.1 = b.1 then (if a.2 = b.2 then 9 / 8 else 0)
    else if a.2 = b.2 then (if a.2 = 0 then Real.sqrt 15 / 8 else -Real.sqrt 15 / 8)
    else 0

/-- PDF p.3: partial transpose of the first complete oscillator mode. -/
def transposeFirst (rho : OscillatorKernel (Fin 2)) : Matrix (ℕ × ℕ) (ℕ × ℕ) ℂ :=
  fun u v => rho ![v.1, u.2] ![u.1, v.2]

/-- PDF p.3: E_N=log_2 ||rho^(T_A)||_1. Every use below concludes finiteness
and norm>=1, before converting the extended norm to a real logarithm. -/
def logarithmicNegativity (rho : OscillatorKernel (Fin 2)) : ℝ :=
  Real.logb 2 (traceSize (transposeFirst rho)).toReal

/-! Sanity examples pin definitions only; source results remain unproved. -/

example {I : Type} [Fintype I] (gamma : OscillatorCovariance I) : varianceForm gamma 0 = 0 := by
  simp [varianceForm]

example {I : Type} [Fintype I] (rho : OscillatorKernel I) :
    InDomain none rho ↔ IsDensity rho := by simp [InDomain]

example {A B : Type} (rho : OscillatorKernel A) (sigma : OscillatorKernel B) (u v : Occupation (A ⊕ B)) :
    oscillatorTensor rho sigma u v = rho (u ∘ Sum.inl) (v ∘ Sum.inl) *
      sigma (u ∘ Sum.inr) (v ∘ Sum.inr) := rfl

example {I : Type} {m : ℕ} (k : Copies Bool m) (u : Occupation I)
    (rest : Occupation (I × {l : Copies Bool m // l ≠ k})) (j : I) :
    insertCopy k u rest (j, k) = u j := by simp [insertCopy]

example : hadamardEntry 0 (fun i => Fin.elim0 i) (fun i => Fin.elim0 i) = 1 := by
  simp [hadamardEntry]

example (m : ℕ) : positiveEntries m (fun _ => false) = 2 ^ m := by
  simp [positiveEntries]

example : negativityExampleCovariance (0, 0) (1, 0) = Real.sqrt 15 / 8 := by
  norm_num [negativityExampleCovariance]

example : negativityExampleCovariance (0, 1) (1, 1) = -Real.sqrt 15 / 8 := by
  norm_num [negativityExampleCovariance]

example (rho : OscillatorKernel (Fin 2)) : transposeFirst rho (0, 1) (1, 0) = rho ![1, 1] ![0, 0] := rfl

end QAlgorithms.Papers.Independent.Support50.P14
Source
Extremality of Gaussian quantum states, arXiv:quant-ph/0509154v1, https://arxiv.org/abs/quant-ph/0509154v1
Read-back

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

QAlgorithms.Papers.Independent.Support50.P14.CompleteMoments

For any finite type III, arbitrary complex occupation kernel ρ\rhoρ, real function d:I×{0,1}→Rd:I\times\{0,1\}\to\mathbb Rd:I×{0,1}→R, and real matrix γ\gammaγ on I×{0,1}I\times\{0,1\}I×{0,1}, this property requires finite energy

E(ρ)=∑u:I→N(∑j∈Iu(j))max⁡{Re⁡ρ(u,u),0}<∞,\mathcal E(\rho)=\sum_{u:I\to\mathbb N} \left(\sum_{j\in I}u(j)\right)\max\{\operatorname{Re}\rho(u,u),0\}<\infty,E(ρ)=u:I→N∑​​j∈I∑​u(j)​max{Reρ(u,u),0}<∞,

symmetry γT=γ\gamma^{\mathsf T}=\gammaγT=γ, and the following summability and derivative conditions. For every real ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, put

zj=−ξ(j,1)+i ξ(j,0)2,Dmn(z)=e−∣z∣2/2m! n!∑r=0min⁡(m,n)zm−r(−z‾)n−rr! (m−r)! (n−r)!,χρ(ξ)=∑u∑vρ(u,v)∏j∈IDv(j),u(j)(zj).z_j=\frac{-\xi(j,1)+\mathrm i\,\xi(j,0)}{\sqrt2}, \quad D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{r=0}^{\min(m,n)} \frac{z^{m-r}(-\overline z)^{n-r}}{r!\,(m-r)!\,(n-r)!}, \quad \chi_\rho(\xi)=\sum_u\sum_v\rho(u,v)\prod_{j\in I}D_{v(j),u(j)}(z_j).zj​=2​−ξ(j,1)+iξ(j,0)​,Dmn​(z)=e−∣z∣2/2m!n!​r=0∑min(m,n)​r!(m−r)!(n−r)!zm−r(−z)n−r​,χρ​(ξ)=u∑​v∑​ρ(u,v)j∈I∏​Dv(j),u(j)​(zj​).

For each uuu, the inner family indexed by vvv must be summable, and the family indexed by uuu of these inner sums must also be summable. Writing md(ξ)=∑aξ(a)d(a)m_d(\xi)=\sum_a\xi(a)d(a)md​(ξ)=∑a​ξ(a)d(a), vγ(ξ)=∑a∑bξ(a)γ(a,b)ξ(b)v_\gamma(\xi)=\sum_a\sum_b\xi(a)\gamma(a,b)\xi(b)vγ​(ξ)=∑a​∑b​ξ(a)γ(a,b)ξ(b), and Fξ(t)=χρ(tξ)F_\xi(t)=\chi_\rho(t\xi)Fξ​(t)=χρ​(tξ) for real ttt, the required derivatives at 000 are

Fξ′(0)=i md(ξ),(Fξ′)′(0)=−12vγ(ξ)−md(ξ)2.F_\xi'(0)=\mathrm i\,m_d(\xi), \qquad (F_\xi')'(0)=-\frac12v_\gamma(\xi)-m_d(\xi)^2.Fξ′​(0)=imd​(ξ),(Fξ′​)′(0)=−21​vγ​(ξ)−md​(ξ)2.

In the second assertion, the derivative function uses the total derivative convention, taking value 000 wherever the first derivative does not exist. No density, normalization, or positivity condition on ρ\rhoρ, and no positivity condition on γ\gammaγ, is explicitly included. Empty III is allowed; in that case the property holds for every kernel and the unique d,γd,\gammad,γ.

QAlgorithms.Papers.Independent.Support50.P14.GaussianAt

For any finite type III, arbitrary complex occupation kernel ρ\rhoρ, real function d:I×{0,1}→Rd:I\times\{0,1\}\to\mathbb Rd:I×{0,1}→R, and real matrix γ\gammaγ on that index set, this property requires all of the following. The energy

∑u:I→N(∑j∈Iu(j))max⁡{Re⁡ρ(u,u),0}\sum_{u:I\to\mathbb N}\left(\sum_{j\in I}u(j)\right) \max\{\operatorname{Re}\rho(u,u),0\}u:I→N∑​​j∈I∑​u(j)​max{Reρ(u,u),0}

is finite, γT=γ\gamma^{\mathsf T}=\gammaγT=γ, and, for every real ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, each inner family indexed by vvv and the outer family of its sums in

χρ(ξ)=∑u∑vρ(u,v)∏j∈IDv(j),u(j)(zj),zj=−ξ(j,1)+i ξ(j,0)2,Dmn(z)=e−∣z∣2/2m! n!∑r=0min⁡(m,n)zm−r(−z‾)n−rr! (m−r)! (n−r)!\chi_\rho(\xi)=\sum_u\sum_v\rho(u,v)\prod_{j\in I}D_{v(j),u(j)}(z_j), \quad z_j=\frac{-\xi(j,1)+\mathrm i\,\xi(j,0)}{\sqrt2}, \quad D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{r=0}^{\min(m,n)} \frac{z^{m-r}(-\overline z)^{n-r}}{r!\,(m-r)!\,(n-r)!}χρ​(ξ)=u∑​v∑​ρ(u,v)j∈I∏​Dv(j),u(j)​(zj​),zj​=2​−ξ(j,1)+iξ(j,0)​,Dmn​(z)=e−∣z∣2/2m!n!​r=0∑min(m,n)​r!(m−r)!(n−r)!zm−r(−z)n−r​

are summable. With md(ξ)=∑aξ(a)d(a)m_d(\xi)=\sum_a\xi(a)d(a)md​(ξ)=∑a​ξ(a)d(a), vγ(ξ)=∑a∑bξ(a)γ(a,b)ξ(b)v_\gamma(\xi)=\sum_a\sum_b\xi(a)\gamma(a,b)\xi(b)vγ​(ξ)=∑a​∑b​ξ(a)γ(a,b)ξ(b), and Fξ(t)=χρ(tξ)F_\xi(t)=\chi_\rho(t\xi)Fξ​(t)=χρ​(tξ), it additionally requires

Fξ′(0)=i md(ξ),(Fξ′)′(0)=−12vγ(ξ)−md(ξ)2,χρ(ξ)=exp⁡ ⁣(−14vγ(ξ)+i md(ξ)).F_\xi'(0)=\mathrm i\,m_d(\xi), \qquad (F_\xi')'(0)=-\frac12v_\gamma(\xi)-m_d(\xi)^2, \qquad \chi_\rho(\xi)=\exp\!\left(-\frac14v_\gamma(\xi)+\mathrm i\,m_d(\xi)\right).Fξ′​(0)=imd​(ξ),(Fξ′​)′(0)=−21​vγ​(ξ)−md​(ξ)2,χρ​(ξ)=exp(−41​vγ​(ξ)+imd​(ξ)).

The derivative function in the second condition takes value 000 wherever a derivative does not exist. No density or positivity condition on ρ\rhoρ, or positivity condition on γ\gammaγ, is explicitly required. For empty III, these requirements reduce to the sole entry of ρ\rhoρ being 111.

QAlgorithms.Papers.Independent.Support50.P14.InDomain

For any finite type III, an optional real cap, and arbitrary complex kernel ρ\rhoρ indexed by occupation functions I→NI\to\mathbb NI→N, this property first requires ρ\rhoρ to satisfy the following density conditions: there exist sequences an,bn∈ℓ2(I→N;C)a_n,b_n\in\ell^2(I\to\mathbb N;\mathbb C)an​,bn​∈ℓ2(I→N;C), indexed by n∈Nn\in\mathbb Nn∈N, such that ∑n∥an∥∥bn∥\sum_n\|a_n\|\|b_n\|∑n​∥an​∥∥bn​∥ is summable and ρ(u,v)=∑nan(u)bn(v)‾\rho(u,v)=\sum_n a_n(u)\overline{b_n(v)}ρ(u,v)=∑n​an​(u)bn​(v)​ for every u,vu,vu,v; for every finite set sss of occupation functions and every complex function zzz on occupation functions, the number ∑u∈s∑v∈sz(u)‾ρ(u,v)z(v)\sum_{u\in s}\sum_{v\in s}\overline{z(u)}\rho(u,v)z(v)∑u∈s​∑v∈s​z(u)​ρ(u,v)z(v) has imaginary part 000 and nonnegative real part; and the real diagonal family is summable with ∑uRe⁡ρ(u,u)=1\sum_u\operatorname{Re}\rho(u,u)=1∑u​Reρ(u,u)=1. If no cap is supplied, there is no further requirement. If the supplied cap is e∈Re\in\mathbb Re∈R, the additional requirement is

∑u:I→N(∑j∈Iu(j))max⁡{Re⁡ρ(u,u),0}≤max⁡{e ∣I∣,0},\sum_{u:I\to\mathbb N} \left(\sum_{j\in I}u(j)\right)\max\{\operatorname{Re}\rho(u,u),0\} \le \max\{e\,|I|,0\},u:I→N∑​​j∈I∑​u(j)​max{Reρ(u,u),0}≤max{e∣I∣,0},

interpreted in [0,∞][0,\infty][0,∞]. The cap eee is not assumed nonnegative: a negative cap gives a right-hand side of 000. Without a cap, finite photon energy is not required. Empty III is allowed; then the density conditions force the sole entry to be 111, and every supplied cap gives the energy bound 0≤00\le00≤0.

QAlgorithms.Papers.Independent.Support50.P14.InterleavedQuadrature

For any type III, the interleaved quadrature index set is I×{0,1}I\times\{0,1\}I×{0,1}. Thus each j∈Ij\in Ij∈I has two indices, (j,0)(j,0)(j,0) and (j,1)(j,1)(j,1). No finiteness or nonemptiness assumption is imposed; when III is empty, this index set is empty.

QAlgorithms.Papers.Independent.Support50.P14.LocallyInvariant

For a natural number MMM, an optional real cap ccc, and a real-valued functional f(I,o,T)f(I,o,T)f(I,o,T) on finitely enumerated mode types III, maps o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and complex matrices on NI\mathbb N^INI, this predicate requires invariance under the following two operations whenever both the original and resulting matrices are admissible. Here a matrix TTT on an index set LLL is a density when it admits Tij=∑rxr(i)yr(j)‾T_{ij}=\sum_r x_r(i)\overline{y_r(j)}Tij​=∑r​xr​(i)yr​(j)​ with xr,yr∈ℓ2(L;C)x_r,y_r\in\ell^2(L;\mathbb C)xr​,yr​∈ℓ2(L;C) and ∑r∥xr∥∥yr∥<∞\sum_r\|x_r\|\|y_r\|<\infty∑r​∥xr​∥∥yr​∥<∞, every quadratic form ∑i,j∈Fzi‾Tijzj\sum_{i,j\in F}\overline{z_i}T_{ij}z_j∑i,j∈F​zi​​Tij​zj​ is real and nonnegative for every finite F⊆LF\subseteq LF⊆L and every z:L→Cz:L\to\mathbb Cz:L→C, and its real diagonal entries are summable with sum 111. Admissibility additionally requires, when ccc specifies e∈Re\in\mathbb Re∈R, that ∑u∈NI(∑j∈Iuj)max⁡{Re⁡Tu,u,0}≤max⁡{e∣I∣,0}\sum_{u\in\mathbb N^I}(\sum_{j\in I}u_j)\max\{\operatorname{Re}T_{u,u},0\}\le\max\{e|I|,0\}∑u∈NI​(∑j∈I​uj​)max{ReTu,u​,0}≤max{e∣I∣,0} in [0,∞][0,\infty][0,∞]; an absent cap adds no condition. First, for every I,o,ρI,o,\rhoI,o,ρ and every family of surjective complex linear isometries UpU_pUp​ of ℓ2(NIp;C)\ell^2(\mathbb N^{I_p};\mathbb C)ℓ2(NIp​;C), where Ip={j:o(j)=p}I_p=\{j:o(j)=p\}Ip​={j:o(j)=p}, put Ku,v=∏p=0M−1(Upev∣Ip)(u∣Ip)K_{u,v}=\prod_{p=0}^{M-1}(U_p e_{v|_{I_p}})(u|_{I_p})Ku,v​=∏p=0M−1​(Up​ev∣Ip​​​)(u∣Ip​​), with ewe_wew​ the coordinate unit vector, and define

ρu,v′=∑a∈NIKu,a(∑b∈NIρa,bKv,b‾).\rho'_{u,v}=\sum_{a\in\mathbb N^I}K_{u,a} \left(\sum_{b\in\mathbb N^I}\rho_{a,b}\overline{K_{v,b}}\right).ρu,v′​=a∈NI∑​Ku,a​(b∈NI∑​ρa,b​Kv,b​​).

If ρ,ρ′\rho,\rho'ρ,ρ′ are admissible, then f(I,o,ρ′)=f(I,o,ρ)f(I,o,\rho')=f(I,o,\rho)f(I,o,ρ′)=f(I,o,ρ). Second, for every pair of finitely enumerated types I,JI,JI,J, bijection e:I→Je:I\to Je:I→J, ownership map o:J→{0,…,M−1}o:J\to\{0,\ldots,M-1\}o:J→{0,…,M−1}, and matrix ρ\rhoρ on NI\mathbb N^INI, define ρu,ve=ρu∘e,v∘e\rho^e_{u,v}=\rho_{u\circ e,v\circ e}ρu,ve​=ρu∘e,v∘e​ for u,v:J→Nu,v:J\to\mathbb Nu,v:J→N; if ρ,ρe\rho,\rho^eρ,ρe are admissible, then f(J,o,ρe)=f(I,o∘e,ρ)f(J,o,\rho^e)=f(I,o\circ e,\rho)f(J,o,ρe)=f(I,o∘e,ρ). The nested complex sums retain their indicated order and are zero individually when nonsummable. Empty mode types and parties owning no modes are included; if M=0M=0M=0, ownership maps require empty mode types and the product defining KKK is 111. A negative specified cap gives the energy bound zero.

QAlgorithms.Papers.Independent.Support50.P14.MultipartiteOscillatorFunctional

For any natural number MMM, a multipartite oscillator functional assigns a real number f(I,o,ρ)f(I,o,\rho)f(I,o,ρ) to every type III equipped with a finite enumeration, every map o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and every complex matrix ρ\rhoρ indexed by NI×NI\mathbb N^I\times\mathbb N^INI×NI, where NI\mathbb N^INI is the set of occupation assignments I→NI\to\mathbb NI→N. No positivity, normalization, energy bound, or regularity is required of either the matrix or the functional. Empty III is included; then NI\mathbb N^INI has one element, so such a matrix has one complex entry. The value M=0M=0M=0 is included, in which case an ownership map exists only when III is empty.

QAlgorithms.Papers.Independent.Support50.P14.OscillatorCharacteristicSummable

For any finite type III and arbitrary complex kernel ρ\rhoρ indexed by occupation functions I→NI\to\mathbb NI→N, this property means that, for every real ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, both of the following summability conditions hold: for every occupation function uuu, the complex family

v⟼ρ(u,v)∏j∈IDv(j),u(j)(zj)v\longmapsto \rho(u,v)\prod_{j\in I}D_{v(j),u(j)}(z_j)v⟼ρ(u,v)j∈I∏​Dv(j),u(j)​(zj​)

is summable, and the family

u⟼∑v:I→Nρ(u,v)∏j∈IDv(j),u(j)(zj)u\longmapsto\sum_{v:I\to\mathbb N}\rho(u,v)\prod_{j\in I}D_{v(j),u(j)}(z_j)u⟼v:I→N∑​ρ(u,v)j∈I∏​Dv(j),u(j)​(zj​)

is summable, where

zj=−ξ(j,1)+i ξ(j,0)2,Dmn(z)=e−∣z∣2/2m! n!∑r=0min⁡(m,n)zm−r(−z‾)n−rr! (m−r)! (n−r)!.z_j=\frac{-\xi(j,1)+\mathrm i\,\xi(j,0)}{\sqrt2}, \qquad D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{r=0}^{\min(m,n)} \frac{z^{m-r}(-\overline z)^{n-r}}{r!\,(m-r)!\,(n-r)!}.zj​=2​−ξ(j,1)+iξ(j,0)​,Dmn​(z)=e−∣z∣2/2m!n!​r=0∑min(m,n)​r!(m−r)!(n−r)!zm−r(−z)n−r​.

These are conditions on the inner families and the family of their sums; no separate summability condition on the family indexed jointly by (u,v)(u,v)(u,v) is included. No density condition is imposed. For empty III, both required families have singleton index sets, so the property holds for every kernel.

QAlgorithms.Papers.Independent.Support50.P14.OscillatorCovariance

For any type III, an oscillator covariance is an arbitrary real matrix γ\gammaγ indexed by I×{0,1}I\times\{0,1\}I×{0,1} in both coordinates, equivalently a function γ:(I×{0,1})×(I×{0,1})→R\gamma:(I\times\{0,1\})\times(I\times\{0,1\})\to\mathbb Rγ:(I×{0,1})×(I×{0,1})→R. This abbreviation imposes no symmetry, positivity, or other covariance condition, and does not require III to be finite or nonempty. For empty III, it is the unique matrix with no entries.

QAlgorithms.Papers.Independent.Support50.P14.OscillatorKernel

For any type III, an oscillator kernel is a complex matrix indexed in both coordinates by occupation functions u:I→Nu:I\to\mathbb Nu:I→N; explicitly, it is an arbitrary function ρ:(I→N)×(I→N)→C\rho:(I\to\mathbb N)\times(I\to\mathbb N)\to\mathbb Cρ:(I→N)×(I→N)→C. No finiteness assumption on III, positivity, normalization, symmetry, boundedness, or summability is imposed. If III is empty, there is exactly one occupation function, so such a kernel has one complex entry.

QAlgorithms.Papers.Independent.Support50.P14.OscillatorTraceContinuous

For every natural number MMM, optional real cap ccc, and real-valued functional f(I,o,T)f(I,o,T)f(I,o,T) on finitely enumerated mode types III, ownership maps o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and arbitrary complex matrices on NI\mathbb N^INI, this predicate means the following. A matrix TTT on an index set LLL is a density here when it has an expansion Tij=∑rxr(i)yr(j)‾T_{ij}=\sum_r x_r(i)\overline{y_r(j)}Tij​=∑r​xr​(i)yr​(j)​ with xr,yr∈ℓ2(L;C)x_r,y_r\in\ell^2(L;\mathbb C)xr​,yr​∈ℓ2(L;C) and ∑r∥xr∥∥yr∥<∞\sum_r\|x_r\|\|y_r\|<\infty∑r​∥xr​∥∥yr​∥<∞, all finite quadratic forms ∑i,j∈Fzi‾Tijzj\sum_{i,j\in F}\overline{z_i}T_{ij}z_j∑i,j∈F​zi​​Tij​zj​ are real and nonnegative for every finite F⊆LF\subseteq LF⊆L and every z:L→Cz:L\to\mathbb Cz:L→C, and its real diagonal entries are summable with sum 111. It is admissible if it is a density and, when ccc specifies e∈Re\in\mathbb Re∈R, its energy ∑u∈NI(∑j∈Iuj)max⁡{Re⁡Tu,u,0}\sum_{u\in\mathbb N^I}(\sum_{j\in I}u_j)\max\{\operatorname{Re}T_{u,u},0\}∑u∈NI​(∑j∈I​uj​)max{ReTu,u​,0} is at most max⁡{e∣I∣,0}\max\{e|I|,0\}max{e∣I∣,0} in [0,∞][0,\infty][0,∞]; an absent cap adds no energy condition. Define τ(X)∈[0,∞]\tau(X)\in[0,\infty]τ(X)∈[0,∞] to be the infimum of ∑r∥xr∥∥yr∥\sum_r\|x_r\|\|y_r\|∑r​∥xr​∥∥yr​∥ over all expansions of XXX of the preceding form with summable norm products, with infimum +∞+\infty+∞ if no expansion exists. Then, for every I,oI,oI,o, every admissible ρ\rhoρ, and every real ε>0\varepsilon>0ε>0, there is a real δ>0\delta>0δ>0 such that every admissible σ\sigmaσ satisfying τ(σ−ρ)<δ\tau(\sigma-\rho)<\deltaτ(σ−ρ)<δ also satisfies

∣f(I,o,σ)−f(I,o,ρ)∣<ε.|f(I,o,\sigma)-f(I,o,\rho)|<\varepsilon.∣f(I,o,σ)−f(I,o,ρ)∣<ε.

Empty mode types are included; for M=0M=0M=0, only empty mode types admit ownership maps. A negative specified cap gives the energy bound zero. The number δ\deltaδ may depend on III, its finite enumeration, ooo, ρ\rhoρ, and ε\varepsilonε.

QAlgorithms.Papers.Independent.Support50.P14.StronglySuperadditive

For a natural number MMM, an optional real cap ccc, and a real-valued functional f(I,o,T)f(I,o,T)f(I,o,T) defined for every finitely enumerated mode type III, ownership map o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and arbitrary complex matrix TTT on NI\mathbb N^INI, this predicate requires both the marginal inequality and product equality below. For clarity, a matrix TTT on an index set LLL is a density here precisely when it has an expansion Tij=∑r=0∞xr(i)yr(j)‾T_{ij}=\sum_{r=0}^{\infty}x_r(i)\overline{y_r(j)}Tij​=∑r=0∞​xr​(i)yr​(j)​ with xr,yr∈ℓ2(L;C)x_r,y_r\in\ell^2(L;\mathbb C)xr​,yr​∈ℓ2(L;C) and ∑r∥xr∥∥yr∥<∞\sum_r\|x_r\|\|y_r\|<\infty∑r​∥xr​∥∥yr​∥<∞, every finite quadratic form ∑i,j∈Fzi‾Tijzj\sum_{i,j\in F}\overline{z_i}T_{ij}z_j∑i,j∈F​zi​​Tij​zj​ is real and nonnegative for every finite F⊆LF\subseteq LF⊆L and every z:L→Cz:L\to\mathbb Cz:L→C, and the real diagonal entries are summable with sum 111. An oscillator matrix TTT on NI\mathbb N^INI is admissible when it is such a density and, if ccc specifies a real number eee, its energy EI(T)=∑u∈NI(∑j∈Iuj)max⁡{Re⁡Tu,u,0}\mathcal E_I(T)=\sum_{u\in\mathbb N^I}(\sum_{j\in I}u_j)\max\{\operatorname{Re}T_{u,u},0\}EI​(T)=∑u∈NI​(∑j∈I​uj​)max{ReTu,u​,0}, evaluated in [0,∞][0,\infty][0,∞], satisfies EI(T)≤max⁡{e∣I∣,0}\mathcal E_I(T)\le\max\{e|I|,0\}EI​(T)≤max{e∣I∣,0}; an absent cap adds no energy condition. The first requirement is that, for every pair of finitely enumerated types A,BA,BA,B, ownership maps a,ba,ba,b, and matrix TTT on occupations of A⊔BA\sqcup BA⊔B, if TTT and both matrices (TA)u,v=∑w∈NBTu⊔w,v⊔w(T_A)_{u,v}=\sum_{w\in\mathbb N^B}T_{u\sqcup w,v\sqcup w}(TA​)u,v​=∑w∈NB​Tu⊔w,v⊔w​ and (TB)u,v=∑w∈NATw⊔u,w⊔v(T_B)_{u,v}=\sum_{w\in\mathbb N^A}T_{w\sqcup u,w\sqcup v}(TB​)u,v​=∑w∈NA​Tw⊔u,w⊔v​ are admissible, then

f(A,a,TA)+f(B,b,TB)≤f(A⊔B,a⊔b,T).f(A,a,T_A)+f(B,b,T_B)\le f(A\sqcup B,a\sqcup b,T).f(A,a,TA​)+f(B,b,TB​)≤f(A⊔B,a⊔b,T).

The second requirement is that, for every such A,B,a,bA,B,a,bA,B,a,b and every pair of matrices ρ,σ\rho,\sigmaρ,σ, if ρ,σ\rho,\sigmaρ,σ and the matrix (ρ⊗σ)u,v=ρu∣A,v∣Aσu∣B,v∣B(\rho\otimes\sigma)_{u,v}=\rho_{u|_A,v|_A}\sigma_{u|_B,v|_B}(ρ⊗σ)u,v​=ρu∣A​,v∣A​​σu∣B​,v∣B​​ are admissible, then

f(A⊔B,a⊔b,ρ⊗σ)=f(A,a,ρ)+f(B,b,σ).f(A\sqcup B,a\sqcup b,\rho\otimes\sigma)=f(A,a,\rho)+f(B,b,\sigma).f(A⊔B,a⊔b,ρ⊗σ)=f(A,a,ρ)+f(B,b,σ).

The marginal sums are unconditional complex sums, with value zero for nonsummable families. Empty mode types are included; a negative specified cap imposes the energy bound zero. When M=0M=0M=0, ownership maps exist only for empty mode types. Each requirement is conditional on all three stated admissibility hypotheses.

QAlgorithms.Papers.Independent.Support50.P14.TraceLowerSemicontinuous

For every natural number MMM, optional real cap ccc, and real-valued functional f(I,o,T)f(I,o,T)f(I,o,T) on finitely enumerated mode types III, ownership maps o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and arbitrary complex matrices on NI\mathbb N^INI, this predicate means the following. A matrix TTT on an index set LLL is a density here when Tij=∑rxr(i)yr(j)‾T_{ij}=\sum_r x_r(i)\overline{y_r(j)}Tij​=∑r​xr​(i)yr​(j)​ for some xr,yr∈ℓ2(L;C)x_r,y_r\in\ell^2(L;\mathbb C)xr​,yr​∈ℓ2(L;C) with ∑r∥xr∥∥yr∥<∞\sum_r\|x_r\|\|y_r\|<\infty∑r​∥xr​∥∥yr​∥<∞, every finite quadratic form ∑i,j∈Fzi‾Tijzj\sum_{i,j\in F}\overline{z_i}T_{ij}z_j∑i,j∈F​zi​​Tij​zj​ is real and nonnegative for every finite F⊆LF\subseteq LF⊆L and every z:L→Cz:L\to\mathbb Cz:L→C, and its real diagonal entries are summable with sum 111. It is admissible if it is a density and, when ccc specifies e∈Re\in\mathbb Re∈R, its energy ∑u∈NI(∑j∈Iuj)max⁡{Re⁡Tu,u,0}\sum_{u\in\mathbb N^I}(\sum_{j\in I}u_j)\max\{\operatorname{Re}T_{u,u},0\}∑u∈NI​(∑j∈I​uj​)max{ReTu,u​,0} is at most max⁡{e∣I∣,0}\max\{e|I|,0\}max{e∣I∣,0} in [0,∞][0,\infty][0,∞]; an absent cap adds no energy condition. Let τ(X)∈[0,∞]\tau(X)\in[0,\infty]τ(X)∈[0,∞] be the infimum of ∑r∥xr∥∥yr∥\sum_r\|x_r\|\|y_r\|∑r​∥xr​∥∥yr​∥ over all such expansions of XXX with summable norm products, and let it be +∞+\infty+∞ if no expansion exists. Then, for every I,oI,oI,o, every admissible ρ\rhoρ, and every real a<f(I,o,ρ)a<f(I,o,\rho)a<f(I,o,ρ), there exists a real δ>0\delta>0δ>0 such that every admissible σ\sigmaσ satisfies

τ(σ−ρ)<δ⟹a<f(I,o,σ).\tau(\sigma-\rho)<\delta\quad\Longrightarrow\quad a<f(I,o,\sigma).τ(σ−ρ)<δ⟹a<f(I,o,σ).

Empty mode types are included; when M=0M=0M=0, ownership maps exist only for empty mode types. A negative specified cap gives the energy bound zero. The number δ\deltaδ may depend on III, its finite enumeration, ooo, ρ\rhoρ, and aaa.

QAlgorithms.Papers.Independent.Support50.P14.copiesKernel

For every type III, with no finiteness assumption, every natural number mmm, and every complex matrix ρ\rhoρ indexed by NI×NI\mathbb N^I\times\mathbb N^INI×NI, let CmC_mCm​ be the set of Boolean strings of length mmm. This definition gives a matrix indexed by occupation assignments u,v:I×Cm→Nu,v:I\times C_m\to\mathbb Nu,v:I×Cm​→N, with entries

ρu,v[m]=∏k∈Cmρ j↦u(j,k), j↦v(j,k).\rho^{[m]}_{u,v} =\prod_{k\in C_m} \rho_{\,j\mapsto u(j,k),\,j\mapsto v(j,k)}.ρu,v[m]​=k∈Cm​∏​ρj↦u(j,k),j↦v(j,k)​.

The product has 2m2^m2m factors, and no density, normalization, or summability assumptions are imposed on ρ\rhoρ. For m=0m=0m=0, there is one empty Boolean string, so the product has one factor. If III is empty, the resulting matrix has one entry equal to the single entry of ρ\rhoρ raised to the power 2m2^m2m.

QAlgorithms.Papers.Independent.Support50.P14.copyMarginal

For any type III, any natural number mmm, any complex matrix ρ\rhoρ indexed by occupation functions I×Cm→NI\times C_m\to\mathbb NI×Cm​→N, and any k∈Cm={0,1}{0,…,m−1}k\in C_m=\{0,1\}^{\{0,\ldots,m-1\}}k∈Cm​={0,1}{0,…,m−1}, this defines a complex matrix on occupation functions u,v:I→Nu,v:I\to\mathbb Nu,v:I→N by

(ρk)u,v=∑r:I×(Cm∖{k})→Nρinsert⁡k(u,r), insert⁡k(v,r),(\rho_k)_{u,v} = \sum_{r:I\times(C_m\setminus\{k\})\to\mathbb N} \rho_{\operatorname{insert}_k(u,r),\,\operatorname{insert}_k(v,r)},(ρk​)u,v​=r:I×(Cm​∖{k})→N∑​ρinsertk​(u,r),insertk​(v,r)​,

where insert⁡k(u,r)(j,l)\operatorname{insert}_k(u,r)(j,l)insertk​(u,r)(j,l) is u(j)u(j)u(j) if l=kl=kl=k, and r(j,l)r(j,l)r(j,l) otherwise. No density, positivity, trace-class, or summability hypothesis is imposed on ρ\rhoρ, and III need not be finite; the displayed sum is the totalized sum over its index type, which is 000 when its summands are not summable. For m=0m=0m=0, there is exactly one rest occupation function, so the sum has one term.

QAlgorithms.Papers.Independent.Support50.P14.copyMixerKernel

For any finite type III, any natural number mmm, and any complex linear isometric equivalence UUU of ℓ2(NCm;C)\ell^2(\mathbb N^{C_m};\mathbb C)ℓ2(NCm​;C) with itself, where Cm={0,1}{0,…,m−1}C_m=\{0,1\}^{\{0,\ldots,m-1\}}Cm​={0,1}{0,…,m−1}, this defines a complex matrix KKK indexed by occupation functions u,v:I×Cm→Nu,v:I\times C_m\to\mathbb Nu,v:I×Cm​→N, with entries

Ku,v=∏j∈I(Uevj)(uj),uj(k)=u(j,k),vj(k)=v(j,k),K_{u,v}=\prod_{j\in I}\bigl(Ue_{v_j}\bigr)(u_j), \qquad u_j(k)=u(j,k),\quad v_j(k)=v(j,k),Ku,v​=j∈I∏​(Uevj​​)(uj​),uj​(k)=u(j,k),vj​(k)=v(j,k),

where ewe_wew​ is the function equal to 111 at the occupation function www and 000 elsewhere. No nonemptiness assumption is made on III: if III is empty, this product is 111. The definition includes m=0m=0m=0, when CmC_mCm​ has one element.

QAlgorithms.Papers.Independent.Support50.P14.counterexampleState

This defines a complex matrix ρ\rhoρ indexed by occupation functions u,v:{0,1}→Nu,v:\{0,1\}\to\mathbb Nu,v:{0,1}→N, with ρu,v=c(u)c(v)‾\rho_{u,v}=c(u)\overline{c(v)}ρu,v​=c(u)c(v)​, where ccc is 15/4\sqrt{15}/415​/4 at the occupation (0,0)(0,0)(0,0), 1/41/41/4 at (1,1)(1,1)(1,1), and 000 elsewhere. Equivalently, its only nonzero entries are

ρ(0,0),(0,0)=1516,ρ(1,1),(1,1)=116,ρ(0,0),(1,1)=ρ(1,1),(0,0)=1516.\rho_{(0,0),(0,0)}=\frac{15}{16},\qquad \rho_{(1,1),(1,1)}=\frac1{16},\qquad \rho_{(0,0),(1,1)}=\rho_{(1,1),(0,0)}=\frac{\sqrt{15}}{16}.ρ(0,0),(0,0)​=1615​,ρ(1,1),(1,1)​=161​,ρ(0,0),(1,1)​=ρ(1,1),(0,0)​=1615​​.

QAlgorithms.Papers.Independent.Support50.P14.counterexampleVector

This defines a complex-valued function on occupation functions u:{0,1}→Nu:\{0,1\}\to\mathbb Nu:{0,1}→N by

c(u)={15/4,u(0)=u(1)=0,1/4,u(0)=u(1)=1,0,otherwise,c(u)= \begin{cases} \sqrt{15}/4,&u(0)=u(1)=0,\\ 1/4,&u(0)=u(1)=1,\\ 0,&\text{otherwise}, \end{cases}c(u)=⎩⎨⎧​15​/4,1/4,0,​u(0)=u(1)=0,u(0)=u(1)=1,otherwise,​

with the displayed real values regarded as complex numbers. In particular, the two exceptional occupations are the constant-zero and constant-one functions.

QAlgorithms.Papers.Independent.Support50.P14.hadamardEntry

For every natural number mmm and every pair of Boolean strings k,l:{0,…,m−1}→{false,true}k,l:\{0,\ldots,m-1\}\to\{\mathrm{false},\mathrm{true}\}k,l:{0,…,m−1}→{false,true}, this real-valued entry is

Hm(k,l)=(−1) ∣{j:k(j)=true and l(j)=true}∣.H_m(k,l)=(-1)^{\,|\{j: k(j)=\mathrm{true}\ \text{and}\ l(j)=\mathrm{true}\}|}.Hm​(k,l)=(−1)∣{j:k(j)=true and l(j)=true}∣.

Thus its exponent is the natural-number count of positions where both strings are true. The case m=0m=0m=0 is included: there is one empty string, the count is zero, and the entry is 111.

QAlgorithms.Papers.Independent.Support50.P14.insertCopy

For any type III, any natural number mmm, any binary string k∈Cm={0,1}{0,…,m−1}k\in C_m=\{0,1\}^{\{0,\ldots,m-1\}}k∈Cm​={0,1}{0,…,m−1}, any occupation function u:I→Nu:I\to\mathbb Nu:I→N, and any occupation function r:I×(Cm∖{k})→Nr:I\times(C_m\setminus\{k\})\to\mathbb Nr:I×(Cm​∖{k})→N, this defines an occupation function on I×CmI\times C_mI×Cm​ by

insert⁡k(u,r)(j,l)={u(j),l=k,r(j,l),l≠k.\operatorname{insert}_k(u,r)(j,l)= \begin{cases} u(j),&l=k,\\ r(j,l),&l\ne k. \end{cases}insertk​(u,r)(j,l)={u(j),r(j,l),​l=k,l=k.​

There is no finiteness or nonemptiness assumption on III. For m=0m=0m=0, kkk is the unique empty binary string, the domain of rrr is empty, and the resulting function takes value u(j)u(j)u(j) on the unique copy.

QAlgorithms.Papers.Independent.Support50.P14.localKernel

For every natural number MMM, every finitely enumerated type III, every ownership map o:I→{0,…,M−1}o:I\to\{0,\ldots,M-1\}o:I→{0,…,M−1}, and every family of surjective complex linear isometries UpU_pUp​ of ℓ2(NIp;C)\ell^2(\mathbb N^{I_p};\mathbb C)ℓ2(NIp​;C), where Ip={j∈I:o(j)=p}I_p=\{j\in I:o(j)=p\}Ip​={j∈I:o(j)=p}, the local kernel is the complex matrix on NI\mathbb N^INI defined, for all occupation assignments u,vu,vu,v, by

Ku,v=∏p=0M−1(Upev∣Ip)(u∣Ip),K_{u,v}=\prod_{p=0}^{M-1}\bigl(U_p e_{v|_{I_p}}\bigr)(u|_{I_p}),Ku,v​=p=0∏M−1​(Up​ev∣Ip​​​)(u∣Ip​​),

where ewe_wew​ is the square-summable vector equal to 111 at occupation assignment www and 000 elsewhere. Parties with Ip=∅I_p=\varnothingIp​=∅ are included: their occupation space is a singleton and their Hilbert space is one-dimensional. If M=0M=0M=0, the ownership map requires I=∅I=\varnothingI=∅, and the empty product gives the single kernel entry 111.

QAlgorithms.Papers.Independent.Support50.P14.logarithmicNegativity

For every complex matrix ρ\rhoρ indexed by occupations {0,1}→N\{0,1\}\to\mathbb N{0,1}→N, form T(a,b),(c,d)=ρ(c,b),(a,d)T_{(a,b),(c,d)}=\rho_{(c,b),(a,d)}T(a,b),(c,d)​=ρ(c,b),(a,d)​. This defines the real number

log⁡(N(T)real)log⁡2,N(T)=inf⁡ar,br∈ℓ2(N2;C)∑r∥ar∥2∥br∥2<∞Tp,q=∑r=0∞ar(p)br(q)‾ ∀p,q∑r=0∞∥ar∥2∥br∥2.\frac{\log\bigl(\mathcal N(T)_{\mathrm{real}}\bigr)}{\log 2}, \qquad \mathcal N(T)= \inf_{\substack{a_r,b_r\in\ell^2(\mathbb N^2;\mathbb C)\\ \sum_r\|a_r\|_2\|b_r\|_2<\infty\\ T_{p,q}=\sum_{r=0}^{\infty}a_r(p)\overline{b_r(q)}\ \forall p,q}} \sum_{r=0}^{\infty}\|a_r\|_2\|b_r\|_2.log2log(N(T)real​)​,N(T)=ar​,br​∈ℓ2(N2;C)∑r​∥ar​∥2​∥br​∥2​<∞Tp,q​=∑r=0∞​ar​(p)br​(q)​ ∀p,q​inf​r=0∑∞​∥ar​∥2​∥br​∥2​.

The infimum is taken in [0,∞][0,\infty][0,∞], and is +∞+\infty+∞ if there is no such expansion. The conversion N(T)real\mathcal N(T)_{\mathrm{real}}N(T)real​ equals N(T)\mathcal N(T)N(T) when finite and equals 000 when it is infinite. The logarithm is totalized with log⁡0=0\log 0=0log0=0; thus the defined value is 000 if N(T)\mathcal N(T)N(T) is either 000 or +∞+\infty+∞. No density or trace-class hypothesis is required.

QAlgorithms.Papers.Independent.Support50.P14.marginalFirst

For arbitrary types A,BA,BA,B, with no finiteness assumptions, and any complex matrix ρ\rhoρ indexed by occupation assignments of A⊔BA\sqcup BA⊔B, the first marginal is the complex matrix indexed by NA×NA\mathbb N^A\times\mathbb N^ANA×NA whose entries, for every u,v:A→Nu,v:A\to\mathbb Nu,v:A→N, are

(ρA)u,v=∑w∈NBρu⊔w, v⊔w,(\rho_A)_{u,v} =\sum_{w\in\mathbb N^B}\rho_{u\sqcup w,\,v\sqcup w},(ρA​)u,v​=w∈NB∑​ρu⊔w,v⊔w​,

where u⊔wu\sqcup wu⊔w is the assignment equal to uuu on AAA and www on BBB. No density or summability hypothesis is required; the sum is the unconditional complex sum, defined to be zero when the indexed family is not summable. If BBB is empty, the sum has one term, corresponding to its unique empty occupation assignment.

QAlgorithms.Papers.Independent.Support50.P14.marginalSecond

For arbitrary types A,BA,BA,B, with no finiteness assumptions, and any complex matrix ρ\rhoρ indexed by occupation assignments of A⊔BA\sqcup BA⊔B, the second marginal is the complex matrix indexed by NB×NB\mathbb N^B\times\mathbb N^BNB×NB whose entries, for every u,v:B→Nu,v:B\to\mathbb Nu,v:B→N, are

(ρB)u,v=∑w∈NAρw⊔u, w⊔v,(\rho_B)_{u,v} =\sum_{w\in\mathbb N^A}\rho_{w\sqcup u,\,w\sqcup v},(ρB​)u,v​=w∈NA∑​ρw⊔u,w⊔v​,

where w⊔uw\sqcup uw⊔u is the assignment equal to www on AAA and uuu on BBB. No density or summability hypothesis is required; the sum is the unconditional complex sum, defined to be zero when the indexed family is not summable. If AAA is empty, the sum has one term, corresponding to its unique empty occupation assignment.

QAlgorithms.Papers.Independent.Support50.P14.meanForm

For any finite type III and arbitrary real functions d,ξ:I×{0,1}→Rd,\xi:I\times\{0,1\}\to\mathbb Rd,ξ:I×{0,1}→R, the mean form is

md(ξ)=∑a∈I×{0,1}ξ(a)d(a).m_d(\xi)=\sum_{a\in I\times\{0,1\}}\xi(a)d(a).md​(ξ)=a∈I×{0,1}∑​ξ(a)d(a).

There are no additional conditions on either function. When III is empty, this sum equals 000.

QAlgorithms.Papers.Independent.Support50.P14.mixedArgument

For every natural number mmm, let CmC_mCm​ be the set of Boolean strings of length mmm. Given any real-valued function ξ:Cm×{0,1}→R\xi:C_m\times\{0,1\}\to\mathbb Rξ:Cm​×{0,1}→R, its mixed argument is the function on the same set defined, for every l∈Cml\in C_ml∈Cm​ and q∈{0,1}q\in\{0,1\}q∈{0,1}, by

ξl,q′=∑k∈Cm(−1) ∣{j:k(j)=true and l(j)=true}∣2m ξk,q.\xi'_{{l},q} =\sum_{k\in C_m} \frac{(-1)^{\,|\{j:k(j)=\mathrm{true}\ \text{and}\ l(j)=\mathrm{true}\}|}} {\sqrt{2^m}}\,\xi_{k,q}.ξl,q′​=k∈Cm​∑​2m​(−1)∣{j:k(j)=true and l(j)=true}∣​ξk,q​.

There are no restrictions on ξ\xiξ, and the second coordinate qqq is unchanged by the mixing. The denominator is positive for every mmm. For m=0m=0m=0, CmC_mCm​ contains one empty string and the resulting function equals ξ\xiξ.

QAlgorithms.Papers.Independent.Support50.P14.negativityExampleCovariance

This defines a real matrix indexed by pairs (j,a)∈{0,1}×{0,1}(j,a)\in\{0,1\}\times\{0,1\}(j,a)∈{0,1}×{0,1}. In the ordered basis (0,0),(0,1),(1,0),(1,1)(0,0),(0,1),(1,0),(1,1)(0,0),(0,1),(1,0),(1,1), it is

(9/8015/8009/80−15/815/809/800−15/809/8).\begin{pmatrix} 9/8&0&\sqrt{15}/8&0\\ 0&9/8&0&-\sqrt{15}/8\\ \sqrt{15}/8&0&9/8&0\\ 0&-\sqrt{15}/8&0&9/8 \end{pmatrix}.​9/8015​/80​09/80−15​/8​15​/809/80​0−15​/809/8​​.

QAlgorithms.Papers.Independent.Support50.P14.oscillatorCharacteristic

For any finite type III, arbitrary complex kernel ρ\rhoρ on occupation functions I→NI\to\mathbb NI→N, and real function ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, the characteristic is the iterated sum

χρ(ξ)=∑u:I→N(∑v:I→Nρ(u,v)∏j∈IDv(j),u(j)(zj)),zj=−ξ(j,1)+i ξ(j,0)2,Dmn(z)=e−∣z∣2/2m! n!∑r=0min⁡(m,n)zm−r(−z‾)n−rr! (m−r)! (n−r)!.\chi_\rho(\xi)= \sum_{u:I\to\mathbb N}\left( \sum_{v:I\to\mathbb N}\rho(u,v) \prod_{j\in I}D_{v(j),u(j)}(z_j)\right), \quad z_j=\frac{-\xi(j,1)+\mathrm i\,\xi(j,0)}{\sqrt2}, \quad D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{r=0}^{\min(m,n)} \frac{z^{m-r}(-\overline z)^{n-r}}{r!\,(m-r)!\,(n-r)!}.χρ​(ξ)=u:I→N∑​​v:I→N∑​ρ(u,v)j∈I∏​Dv(j),u(j)​(zj​)​,zj​=2​−ξ(j,1)+iξ(j,0)​,Dmn​(z)=e−∣z∣2/2m!n!​r=0∑min(m,n)​r!(m−r)!(n−r)!zm−r(−z)n−r​.

The inner sum is taken first. Neither density conditions nor summability are assumed: each infinite complex sum is defined to be 000 when its summand family is not summable. If III is empty, the product is 111 and both occupation index sets are singletons, so the characteristic is the sole entry of ρ\rhoρ, independent of ξ\xiξ.

QAlgorithms.Papers.Independent.Support50.P14.oscillatorTensor

For arbitrary types A,BA,BA,B, with no finiteness assumptions, and arbitrary complex matrices ρ\rhoρ indexed by NA×NA\mathbb N^A\times\mathbb N^ANA×NA and σ\sigmaσ indexed by NB×NB\mathbb N^B\times\mathbb N^BNB×NB, their oscillator tensor is the matrix on occupation assignments of the disjoint union A⊔BA\sqcup BA⊔B defined, for every u,v:A⊔B→Nu,v:A\sqcup B\to\mathbb Nu,v:A⊔B→N, by

(ρ⊗σ)u,v=ρu∣A,v∣A σu∣B,v∣B.(\rho\otimes\sigma)_{u,v} =\rho_{u|_A,v|_A}\,\sigma_{u|_B,v|_B}.(ρ⊗σ)u,v​=ρu∣A​,v∣A​​σu∣B​,v∣B​​.

No density or summability conditions are imposed. Either type may be empty; its occupation space then consists of the unique empty assignment, and its matrix contributes its single complex entry as a scalar factor.

QAlgorithms.Papers.Independent.Support50.P14.oscillatorWeylKernel

For any finite type III and real function ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, this definition assigns a complex kernel WξW_\xiWξ​ on occupation functions u,v:I→Nu,v:I\to\mathbb Nu,v:I→N by

Wξ(u,v)=∏j∈IDu(j),v(j) ⁣(−ξ(j,1)+i ξ(j,0)2),Dmn(z)=e−∣z∣2/2m! n!∑r=0min⁡(m,n)zm−r(−z‾)n−rr! (m−r)! (n−r)!.W_\xi(u,v)=\prod_{j\in I}D_{u(j),v(j)}\!\left(\frac{-\xi(j,1)+\mathrm i\,\xi(j,0)}{\sqrt2}\right), \qquad D_{mn}(z)=e^{-|z|^2/2}\sqrt{m!\,n!} \sum_{r=0}^{\min(m,n)} \frac{z^{m-r}(-\overline z)^{n-r}}{r!\,(m-r)!\,(n-r)!}.Wξ​(u,v)=j∈I∏​Du(j),v(j)​(2​−ξ(j,1)+iξ(j,0)​),Dmn​(z)=e−∣z∣2/2m!n!​r=0∑min(m,n)​r!(m−r)!(n−r)!zm−r(−z)n−r​.

Here real factors are regarded as complex numbers, and m,nm,nm,n may be zero, with zeroth powers equal to 111, including 000^000. If III is empty, the product equals 111 at its sole pair of occupation functions.

QAlgorithms.Papers.Independent.Support50.P14.photonEnergy

For any finite type III and arbitrary complex kernel ρ\rhoρ indexed by occupation functions u:I→Nu:I\to\mathbb Nu:I→N, the photon energy is the extended nonnegative real number

E(ρ)=∑u:I→N(∑j∈Iu(j))max⁡{Re⁡ρ(u,u),0},\mathcal E(\rho)= \sum_{u:I\to\mathbb N} \left(\sum_{j\in I}u(j)\right) \max\{\operatorname{Re}\rho(u,u),0\},E(ρ)=u:I→N∑​​j∈I∑​u(j)​max{Reρ(u,u),0},

with all terms interpreted in [0,∞][0,\infty][0,∞]. Thus negative real parts of diagonal entries contribute 000; no positivity, normalization, or convergence hypothesis on ρ\rhoρ is required, and the result may be +∞+\infty+∞. If III is empty, the sole occupation function has total occupation 000, and the energy is 000.

QAlgorithms.Papers.Independent.Support50.P14.positiveEntries

For every natural number mmm and binary string k:{0,…,m−1}→{false,true}k:\{0,\ldots,m-1\}\to\{\mathrm{false},\mathrm{true}\}k:{0,…,m−1}→{false,true}, this defines the natural number

pm(k)={2m,k is identically false,⌊2m/2⌋,otherwise.p_m(k)= \begin{cases} 2^m,&k\text{ is identically false},\\ \left\lfloor 2^m/2\right\rfloor,&\text{otherwise}. \end{cases}pm​(k)={2m,⌊2m/2⌋,​k is identically false,otherwise.​

The division in the second branch is natural-number division. When m=0m=0m=0, the unique string is identically false, so the value is 111; the second branch cannot occur. When m≥1m\ge1m≥1, the second branch equals 2m−12^{m-1}2m−1.

QAlgorithms.Papers.Independent.Support50.P14.transposeFirst

For every complex matrix ρ\rhoρ indexed by occupation functions {0,1}→N\{0,1\}\to\mathbb N{0,1}→N, this defines a complex matrix TTT indexed by pairs of natural numbers by

T(a,b),(c,d)=ρ(c,b),(a,d).T_{(a,b),(c,d)}=\rho_{(c,b),(a,d)}.T(a,b),(c,d)​=ρ(c,b),(a,d)​.

Here a pair (x,y)(x,y)(x,y) used as an index of ρ\rhoρ denotes the occupation function taking values xxx and yyy at 000 and 111, respectively. There are no density, positivity, or summability assumptions on ρ\rhoρ.

QAlgorithms.Papers.Independent.Support50.P14.varianceForm

For any finite type III, arbitrary real matrix γ\gammaγ indexed by I×{0,1}I\times\{0,1\}I×{0,1}, and real function ξ:I×{0,1}→R\xi:I\times\{0,1\}\to\mathbb Rξ:I×{0,1}→R, the variance form is

vγ(ξ)=∑a∈I×{0,1}∑b∈I×{0,1}ξ(a)γ(a,b)ξ(b).v_\gamma(\xi)= \sum_{a\in I\times\{0,1\}}\sum_{b\in I\times\{0,1\}} \xi(a)\gamma(a,b)\xi(b).vγ​(ξ)=a∈I×{0,1}∑​b∈I×{0,1}∑​ξ(a)γ(a,b)ξ(b).

No symmetry or positivity assumption on γ\gammaγ is imposed, so the definition does not require this value to be nonnegative. When III is empty, it equals 000.

Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 24, 2026

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

  • Endorsed by Gohan · Oct 29, 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