Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section III, (27)–(29): Gaussian spectral p-purity formula

Open
QAlgorithms.Papers.Independent.Support50.P28.gaussian_spectral_purity

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

prove2quantum

Let n≥1n\ge1n≥1 and let ρ\rhoρ be an actual normalized centered Gaussian density with real symmetric physical doubled covariance γ\gammaγ. Let ν1,…,νn≥1\nu_1,\ldots,\nu_n\ge1ν1​,…,νn​≥1 be its symplectic eigenvalues, counted once per mode, identified by a Williamson congruence. For real p>1p>1p>1, define fp(x)=(x+1)p−(x−1)pf_p(x)=(x+1)^p-(x-1)^pfp​(x)=(x+1)p−(x−1)p and Fp(γ)=∏j=1nfp(νj)F_p(\gamma)=\prod_{j=1}^n f_p(\nu_j)Fp​(γ)=∏j=1n​fp​(νj​). Then

Tr⁡ρp=∏j=1n2pfp(νj)=2pnFp(γ),μ(ρ):=Tr⁡ρ2=1∏jνj=1det⁡γ.\operatorname{Tr}\rho^p=\prod_{j=1}^n\frac{2^p}{f_p(\nu_j)}=\frac{2^{pn}}{F_p(\gamma)},\qquad \mu(\rho):=\operatorname{Tr}\rho^2=\frac1{\prod_j\nu_j}=\frac1{\sqrt{\det\gamma}}.Trρp=j=1∏n​fp​(νj​)2p​=Fp​(γ)2pn​,μ(ρ):=Trρ2=∏j​νj​1​=detγ​1​.

This computes spectral trace powers from covariance data. The ordinary purity is the square of the Schatten 222-norm. The trace power is independently defined by the actual density's spectral functional calculus. The bounded operator used in that calculus is required, in the theorem's conclusion, to realize the original density coefficients.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P28
Formal statement
noncomputable section

open scoped BigOperators ENNReal ComplexOrder Matrix

namespace QAlgorithms.Papers.Independent.Support50.P28

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms.RealRegisters (Hilbert)

open QAlgorithms.Papers.Independent.Dependency50.P33
  (State IsDensity IsTraceClassChannel rankOne)

open QAlgorithms.Papers.Independent.Claimed50.P47
  (Occupation gaussianStateCharacteristic singleCharacteristic weylKernel)

open QAlgorithms.Papers.Independent.Claimed50.P30 (phiMinus)

/-- PDF p.4, (27)--(29): spectral trace power and ordinary purity of a centered
Gaussian density. Physical normalization follows the proof, correcting (6),(7).
A Williamson spectrum identifies the symplectic eigenvalues without assuming a
normal-state decomposition or any trace-power formula. -/
theorem gaussian_spectral_purity {n : ℕ} (hn : 0 < n)
    (rho : State (Occupation (Fin n))) (G : OscillatorCovariance (Fin n))
    (hG : HasCenteredGaussianCovariance rho G) (nu : Fin n → ℝ)
    (hnu : WilliamsonSpectrum G nu) (p : ℝ) (hp : 1 < p) :
    oscillatorCoefficients (realizeOscillatorCoeff rho.coefficients) = rho.coefficients ∧
    tracePower p rho.coefficients =
      ENNReal.ofReal (∏ i, Real.rpow 2 p / fp p (nu i)) ∧
    tracePower p rho.coefficients =
      ENNReal.ofReal (Real.rpow 2 (p * n) / spectralFp p nu) ∧
    tracePower 2 rho.coefficients = ENNReal.ofReal (1 / ∏ i, nu i) ∧
    tracePower 2 rho.coefficients = ENNReal.ofReal (1 / Real.sqrt G.det) := by sorry

end QAlgorithms.Papers.Independent.Support50.P28

end
Source
Multiplicativity of maximal output purities of Gaussian channels under Gaussian inputs, arXiv:quant-ph/0406065v1, PDF p. 4, Section III, (27)–(29): Gaussian spectral p-purity formula
Read-back

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

Let n>0n>0n>0, let L=N{0,…,n−1}L=\mathbb N^{\{0,\ldots,n-1\}}L=N{0,…,n−1}, and use the coordinate basis eue_ueu​ of ℓ2(L;C)\ell^2(L;\mathbb C)ℓ2(L;C). Let ρ=(ρuv)u,v∈L\rho=(\rho_{uv})_{u,v\in L}ρ=(ρuv​)u,v∈L​ be a state in the following sense: there are sequences aj,bj∈ℓ2(L;C)a_j,b_j\in\ell^2(L;\mathbb C)aj​,bj​∈ℓ2(L;C) such that ∑j∥aj∥∥bj∥<∞\sum_j\|a_j\|\|b_j\|<\infty∑j​∥aj​∥∥bj​∥<∞ and ρuv=∑jaj(u)bj(v)‾\rho_{uv}=\sum_j a_j(u)\overline{b_j(v)}ρuv​=∑j​aj​(u)bj​(v)​; every finite quadratic form ∑u,v∈szu‾ρuvzv\sum_{u,v\in s}\overline{z_u}\rho_{uv}z_v∑u,v∈s​zu​​ρuv​zv​ is real and nonnegative; and the real diagonal entries are summable with sum 111. Write JJJ for the real matrix with one block (01−10)\begin{pmatrix}0&1\\-1&0\end{pmatrix}(0−1​10​) per mode. Suppose GGG is a real symmetric 2n×2n2n\times2n2n×2n matrix with G+iJG+\mathrm iJG+iJ positive semidefinite, and that ρ\rhoρ has the centered Gaussian characteristic specified as follows. For ξ∈R2n\xi\in\mathbb R^{2n}ξ∈R2n, put zi=−(ξi,0+iξi,1)/2z_i=-(\xi_{i,0}+\mathrm i\xi_{i,1})/\sqrt2zi​=−(ξi,0​+iξi,1​)/2​ and

Dab(z)=e−∣z∣2/2a! b!∑ℓ=0min⁡(a,b)za−ℓ(−z‾)b−ℓℓ! (a−ℓ)! (b−ℓ)!.D_{ab}(z)=e^{-|z|^2/2}\sqrt{a!\,b!} \sum_{\ell=0}^{\min(a,b)} \frac{z^{a-\ell}(-\overline z)^{b-\ell}} {\ell!\,(a-\ell)!\,(b-\ell)!}.Dab​(z)=e−∣z∣2/2a!b!​ℓ=0∑min(a,b)​ℓ!(a−ℓ)!(b−ℓ)!za−ℓ(−z)b−ℓ​.

For every ξ\xiξ, each inner series and then the outer series in

χρ(ξ)=∑u∑vρuv∏iDvi,ui(zi)\chi_\rho(\xi)=\sum_u\sum_v\rho_{uv}\prod_iD_{v_i,u_i}(z_i)χρ​(ξ)=u∑​v∑​ρuv​i∏​Dvi​,ui​​(zi​)

are summable, and χρ(ξ)=exp⁡(−ξTJTGJξ/4)\chi_\rho(\xi)=\exp(-\xi^{\mathsf T}J^{\mathsf T}GJ\xi/4)χρ​(ξ)=exp(−ξTJTGJξ/4). Suppose also that νi≥1\nu_i\ge1νi​≥1 for every mode and that there exists a real matrix SSS satisfying STJS=JS^{\mathsf T}JS=JSTJS=J and STGS=diag⁡(νi,νi)iS^{\mathsf T}GS=\operatorname{diag}(\nu_i,\nu_i)_iSTGS=diag(νi​,νi​)i​. For any coefficient matrix AAA, let R(A)R(A)R(A) be a chosen bounded complex linear operator with ⟨eu,R(A)ev⟩=Auv\langle e_u,R(A)e_v\rangle=A_{uv}⟨eu​,R(A)ev​⟩=Auv​, if such an operator exists, and let R(A)=0R(A)=0R(A)=0 otherwise; define

Tq(A)=∑uofReal⁡ ⁣(Re⁡⟨eu,R(A)qeu⟩)∈[0,∞],T_q(A)=\sum_u \operatorname{ofReal}\!\left(\operatorname{Re} \langle e_u,R(A)^q e_u\rangle\right)\in[0,\infty],Tq​(A)=u∑​ofReal(Re⟨eu​,R(A)qeu​⟩)∈[0,∞],

where the power is the continuous functional calculus power and ofReal⁡(t)=max⁡(t,0)\operatorname{ofReal}(t)=\max(t,0)ofReal(t)=max(t,0), viewed in the extended nonnegative reals. Then, for every real p>1p>1p>1,

⟨eu,R(ρ)ev⟩=ρuvfor all u,v,Tp(ρ)=ofReal⁡ ⁣(∏i=0n−12p(νi+1)p−(νi−1)p)=ofReal⁡ ⁣(2pn∏i=0n−1((νi+1)p−(νi−1)p)),T2(ρ)=ofReal⁡ ⁣(1∏iνi)=ofReal⁡ ⁣(1det⁡G).\begin{aligned} \langle e_u,R(\rho)e_v\rangle&=\rho_{uv}\quad\text{for all }u,v,\\ T_p(\rho) &=\operatorname{ofReal}\!\left(\prod_{i=0}^{n-1} \frac{2^p}{(\nu_i+1)^p-(\nu_i-1)^p}\right)\\ &=\operatorname{ofReal}\!\left( \frac{2^{pn}}{\prod_{i=0}^{n-1}((\nu_i+1)^p-(\nu_i-1)^p)}\right),\\ T_2(\rho)&=\operatorname{ofReal}\!\left(\frac1{\prod_i\nu_i}\right) =\operatorname{ofReal}\!\left(\frac1{\sqrt{\det G}}\right). \end{aligned}⟨eu​,R(ρ)ev​⟩Tp​(ρ)T2​(ρ)​=ρuv​for all u,v,=ofReal(i=0∏n−1​(νi​+1)p−(νi​−1)p2p​)=ofReal(∏i=0n−1​((νi​+1)p−(νi​−1)p)2pn​),=ofReal(∏i​νi​1​)=ofReal(detG​1​).​

The exponents in these scalar formulas are real powers; values νi=1\nu_i=1νi​=1 are included, while n=0n=0n=0 and p≤1p\le1p≤1 are excluded by the hypotheses.

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).

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