Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

If Σ\SigmaΣ is a valid covariance matrix and Σ′⪰Σ\Sigma'\succeq\SigmaΣ′⪰Σ then ρ(μ′,Σ′)\rho(\mu',\Sigma')ρ(μ′,Σ′) exists for every μ′\mu'μ′

Proved
BosonicLearn.gaussianState_exists_of_le

by Senrui Chen · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

bosonic-gaussian-statesquantum-informationquantum-learning

Enlarging the covariance of a Gaussian state, and displacing it, gives a Gaussian state.

Let ρ(μ,Σ)\rho(\mu,\Sigma)ρ(μ,Σ) be an nnn-mode Gaussian state. Let Σ′\Sigma'Σ′ be a real 2n×2n2n\times2n2n×2n matrix with Σ′−Σ\Sigma'-\SigmaΣ′−Σ positive semidefinite, and μ′∈R2n\mu'\in\mathbb R^{2n}μ′∈R2n. Then there is an nnn-mode Gaussian state with mean μ′\mu'μ′ and covariance matrix Σ′\Sigma'Σ′.

Why it holds. Σ′+i2Ω=(Σ+i2Ω)+(Σ′−Σ)⪰0\Sigma'+\tfrac{\mathrm i}2\Omega=(\Sigma+\tfrac{\mathrm i}2\Omega)+(\Sigma'-\Sigma)\succeq0Σ′+2i​Ω=(Σ+2i​Ω)+(Σ′−Σ)⪰0, so Σ′\Sigma'Σ′ is a valid covariance matrix. Concretely ρ(μ′,Σ′)=∫Dx ρ(μ,Σ) Dx† dγ(x)\rho(\mu',\Sigma')=\int D_x\,\rho(\mu,\Sigma)\,D_x^\dagger\,\mathrm d\gamma(x)ρ(μ′,Σ′)=∫Dx​ρ(μ,Σ)Dx†​dγ(x) with DxD_xDx​ the displacement operators and γ\gammaγ the Gaussian measure on R2n\mathbb R^{2n}R2n with mean μ′−μ\mu'-\muμ′−μ and covariance Σ′−Σ\Sigma'-\SigmaΣ′−Σ.

Role. In the proof of Theorem 8 this gives, on the good event where Σ~⪰Σ\tilde\Sigma\succeq\SigmaΣ~⪰Σ, the estimate ρ(μ^,Σ~)\rho(\hat\mu,\tilde\Sigma)ρ(μ^​,Σ~), the intermediate state ρ(μ^,Σ)\rho(\hat\mu,\Sigma)ρ(μ^​,Σ) and the centred states ρ(0,Σ~)\rho(0,\tilde\Sigma)ρ(0,Σ~), ρ(0,Σ)\rho(0,\Sigma)ρ(0,Σ) to which Theorem 7 applies.

Formalization Note IsGaussianState ρ μ S is the mission's predicate for ρ=ρ(μ,Σ)\rho=\rho(\mu,\Sigma)ρ=ρ(μ,Σ): the characteristic function of ρ\rhoρ is exp⁡(iξ⋅μ−12ξ⊤Σξ)\exp(\mathrm i\xi\cdot\mu-\tfrac12\xi^\top\Sigma\xi)exp(iξ⋅μ−21​ξ⊤Σξ), together with finite energy, symmetry of Σ\SigmaΣ and the first and second moment identities; quadratures are in grouped order (q1,…,qn,p1,…,pn)(q_1,\dots,q_n,p_1,\dots,p_n)(q1​,…,qn​,p1​,…,pn​) and the vacuum has Σ=121\Sigma=\tfrac12\mathbf 1Σ=21​1. States are State (Occupation (Fin n)), density operators on ℓ2(Nn)\ell^2(\mathbb N^n)ℓ2(Nn). The hypothesis is Matrix.PosSemidef (S' − S), which contains the symmetry of S' − S.

Preamble
import Mathlib
import Definitions.Def_BosonicLearn_traceDist
import Definitions.Def_BosonicLearn_IsGaussianState

open QAlgorithms.Papers.Independent.Dependency50.P33 (State traceSize)
open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation)
open MeasureTheory
open scoped ENNReal Matrix
Formal statement
namespace BosonicLearn

/-- **Gaussian states exist above a valid covariance matrix, with any mean.**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, §7.2, proof of Theorem 8, p. 62: "Σ̃ + (i/2)Ω = (Σ + (i/2)Ω) + (Σ̃ − Σ) ⪰ 0, because Σ is a valid covariance
matrix and Σ̃ − Σ ⪰ 0. Thus Σ̃ is physical".

If `Σ` is the covariance matrix of some `n`-mode Gaussian state and `Σ′ ⪰ Σ` (real symmetric), then for every
`µ′ ∈ ℝ²ⁿ` there is a Gaussian state `ρ(µ′, Σ′)`.

**Formalization Note.** `Σ′ ⪰ Σ` is positive semidefiniteness of `Σ′ − Σ`. Physically `ρ(µ′, Σ′)` is obtained
from `ρ(µ, Σ)` by a displacement and classical Gaussian noise of covariance `Σ′ − Σ`. With `Σ′ = Σ` the statement
gives all displacements of a Gaussian state. -/
theorem gaussianState_exists_of_le {n : ℕ} (ρ : State (Occupation (Fin n))) (μ : Fin n ⊕ Fin n → ℝ)
    (S : Matrix (Fin n ⊕ Fin n) (Fin n ⊕ Fin n) ℝ) (h : IsGaussianState ρ μ S)
    (μ' : Fin n ⊕ Fin n → ℝ) (S' : Matrix (Fin n ⊕ Fin n) (Fin n ⊕ Fin n) ℝ)
    (hS' : (S' - S).PosSemidef) :
    ∃ ρ' : State (Occupation (Fin n)), IsGaussianState ρ' μ' S' := by sorry

end BosonicLearn
Source
Chen, Anna Mele, Anna Mele, Preskill, When are bosonic Gaussian states classical to learn?, arXiv:2609.26705, p. 62 ("Σ̃ + (i/2)Ω = (Σ + (i/2)Ω) + (Σ̃ − Σ) ⪰ 0 … Thus Σ̃ is physical", proof of Theorem 8)

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