Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 4, Eq.(7): Gaussian channel synthesis

Proved
QAlgorithms.Papers.Independent.Bridge50.P16.gaussianSynthesis

by Gohan · Oct 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let Φ1,Φ2\Phi_1,\Phi_2Φ1​,Φ2​ be Gaussian CPTP channels on n≥1n\ge1n≥1 bosonic modes, with real gain matrices TjT_jTj​, symmetric noise matrices NjN_jNj​, and displacement vectors djd_jdj​. Use interleaved quadratures with [qk,pl]=iδkl[q_k,p_l]=i\delta_{kl}[qk​,pl​]=iδkl​ and vacuum covariance I/2I/2I/2. Then their composition is the full Gaussian channel with specification

T=T2T1,N=T2N1T2T+N2,d=T2d1+d2.T=T_2T_1,\qquad N=T_2N_1T_2^{\mathsf T}+N_2,\qquad d=T_2d_1+d_2.T=T2​T1​,N=T2​N1​T2T​+N2​,d=T2​d1​+d2​.

The specification is the all-input Weyl characteristic-function law, not only a rule for Gaussian input moments. This gives the channel composition used in subsequent loss/amplifier identities.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Bridge50_P16
Formal statement
noncomputable section

open Filter MeasureTheory

open scoped BigOperators ENNReal NNReal Matrix ComplexOrder

namespace QAlgorithms.Papers.Independent.Bridge50.P16

attribute [local instance] Classical.propDecidable

attribute [local instance] Classical.decEq

open QAlgorithms.Papers.Independent.Frontier30.P08 (interleavedSymplecticForm)

open QAlgorithms.Papers.Independent.Dependency50.P30 (thermalEntropyBits)

open QAlgorithms.Papers.Independent.Frontier18.P18 (operatorCoeff boundedOperator)

open QAlgorithms.Papers.Independent.Information100.P34 (choi)

open QAlgorithms.Papers.Independent.Dependency50.P32 (CPTP applyChoi)

/-- Theorem 4, PDF p.3 Eq.(7); known result credited to [45] Eq.(26).
All three affine Gaussian parameters are retained on the full multimode channel. -/
theorem gaussianSynthesis {n : ℕ} (hn : 0 < n)
    (E1 E2 : T.Operation (Fin n → ℕ) (Fin n → ℕ))
    (M1 M2 V1 V2 : Matrix (PhaseIndex n) (PhaseIndex n) ℝ)
    (d1 d2 : PhaseIndex n → ℝ)
    (h1 : GaussianSpecification E1 M1 V1 d1)
    (h2 : GaussianSpecification E2 M2 V2 d2) :
    GaussianSpecification (E2 ∘ E1) (M2 * M1)
      (M2 * V1 * M2.transpose + V2) (M2.mulVec d1 + d2) := by sorry

end QAlgorithms.Papers.Independent.Bridge50.P16

end
Source
Improved quantum capacity bounds of Gaussian loss channels and achievable rates with Gottesman-Kitaev-Preskill codes, arXiv:1801.07271v2, PDF p. 3, Theorem 4, Eq.(7): Gaussian channel synthesis
Read-back

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

For every positive natural number nnn, put I={0,…,n−1}×{0,1}I=\{0,\ldots,n-1\}\times\{0,1\}I={0,…,n−1}×{0,1}, A=N{0,…,n−1}A=\mathbb N^{\{0,\ldots,n-1\}}A=N{0,…,n−1}, and let Ω\OmegaΩ be the real matrix indexed by III with Ω(r,s),(t,u)=0\Omega_{(r,s),(t,u)}=0Ω(r,s),(t,u)​=0 if r≠tr\ne tr=t or s=us=us=u, 111 if r=t,s=0,u=1r=t,s=0,u=1r=t,s=0,u=1, and −1-1−1 if r=t,s=1,u=0r=t,s=1,u=0r=t,s=1,u=0. The statement concerns arbitrary functions E1,E2E_1,E_2E1​,E2​ from complex matrices indexed by AAA to such matrices, real matrices M1,M2,V1,V2M_1,M_2,V_1,V_2M1​,M2​,V1​,V2​ indexed by III, and real vectors d1,d2d_1,d_2d1​,d2​ indexed by III. To specify its hypotheses and conclusion, call a matrix XXX indexed by any set JJJ trace class when there exist sequences ur,vr∈ℓ2(J;C)u_r,v_r\in\ell^2(J;\mathbb C)ur​,vr​∈ℓ2(J;C) such that ∑r∥ur∥∥vr∥<∞\sum_r\|u_r\|\|v_r\|<\infty∑r​∥ur​∥∥vr​∥<∞ and Xij=∑rur(i)vr(j)‾X_{ij}=\sum_r u_r(i)\overline{v_r(j)}Xij​=∑r​ur​(i)vr​(j)​ for every i,ji,ji,j; call it a density matrix when it is trace class, every finite quadratic form ∑i,j∈Fzi‾Xijzj\sum_{i,j\in F}\overline{z_i}X_{ij}z_j∑i,j∈F​zi​​Xij​zj​, for finite F⊆JF\subseteq JF⊆J and arbitrary z:J→Cz:J\to\mathbb Cz:J→C, has zero imaginary part and nonnegative real part, and its real diagonal is summable with sum 111. A function EEE is a channel here when it preserves trace-class matrices, is additive and complex-homogeneous on them, and for every natural number rrr, applying EEE to the AAA-indexed blocks of any density matrix indexed by A×{0,…,r−1}A\times\{0,\ldots,r-1\}A×{0,…,r−1} produces a density matrix; explicitly, this blockwise application has entry

(E⊗idr)(X)(a,s),(b,t)=E((X(u,s),(v,t))u,v∈A)ab.(E\otimes\mathrm{id}_r)(X)_{(a,s),(b,t)} =E\bigl((X_{(u,s),(v,t)})_{u,v\in A}\bigr)_{ab}.(E⊗idr​)(X)(a,s),(b,t)​=E((X(u,s),(v,t)​)u,v∈A​)ab​.

For a complex matrix XXX indexed by AAA, define

χX(ξ)=∑a∈A∑b∈AXab∏r=0n−1D ⁣(ξ(r,0)+iξ(r,1)2)br,ar,\chi_X(\xi)= \sum_{a\in A}\sum_{b\in A}X_{ab} \prod_{r=0}^{n-1}D\!\left( \frac{\xi_{(r,0)}+i\xi_{(r,1)}}{\sqrt2}\right)_{b_r,a_r},χX​(ξ)=a∈A∑​b∈A∑​Xab​r=0∏n−1​D(2​ξ(r,0)​+iξ(r,1)​​)br​,ar​​,

where

D(z)pq=e−∣z∣2/2p! q!∑j=0min⁡(p,q)zp−j(−z‾)q−jj!(p−j)!(q−j)!.D(z)_{pq} =e^{-|z|^2/2}\sqrt{p!\,q!} \sum_{j=0}^{\min(p,q)} \frac{z^{p-j}(-\overline z)^{q-j}}{j!(p-j)!(q-j)!}.D(z)pq​=e−∣z∣2/2p!q!​j=0∑min(p,q)​j!(p−j)!(q−j)!zp−j(−z)q−j​.

For each s=1,2s=1,2s=1,2, assume that EsE_sEs​ is a channel in the just-specified sense, VsT=VsV_s^T=V_sVsT​=Vs​,

Vs+i2(Ω−MsΩMsT)⪰0,V_s+\frac i2(\Omega-M_s\Omega M_s^T)\succeq0,Vs​+2i​(Ω−Ms​ΩMsT​)⪰0,

and, for every trace-class XXX and every real vector ξ\xiξ indexed by III,

χEs(X)(ξ)=χX(ΩMsTΩTξ)exp⁡ ⁣(−12ξTΩVsΩTξ+i dsTΩξ).\chi_{E_s(X)}(\xi)= \chi_X(\Omega M_s^T\Omega^T\xi) \exp\!\left(-\frac12\xi^T\Omega V_s\Omega^T\xi +i\,d_s^T\Omega\xi\right).χEs​(X)​(ξ)=χX​(ΩMsT​ΩTξ)exp(−21​ξTΩVs​ΩTξ+idsT​Ωξ).

Then the same complete list of properties holds for E2∘E1E_2\circ E_1E2​∘E1​ with

M=M2M1,V=M2V1M2T+V2,d=M2d1+d2.M=M_2M_1,\qquad V=M_2V_1M_2^T+V_2,\qquad d=M_2d_1+d_2.M=M2​M1​,V=M2​V1​M2T​+V2​,d=M2​d1​+d2​.

In particular, the conclusion asserts that this composition is a channel, VVV is symmetric, V+i2(Ω−MΩMT)V+\tfrac i2(\Omega-M\Omega M^T)V+2i​(Ω−MΩMT) is positive semidefinite, and its characteristic expression obeys the displayed transformation with these M,V,dM,V,dM,V,d for every trace-class input. All infinite sums in the characteristic expression are the total infinite-sum operation; their summability is not separately assumed or asserted. The finite ancillary dimension rrr includes 000, for which there are no density matrices to test.

Human review
  • Endorsed by Shuze Chen · Oct 29, 2026

    Confirmed by the moderator at approval.

  • 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