Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

The hard family ρRX\rho_{R_X}ρRX​​, X∈GsuppX\in G_{\rm supp}X∈Gsupp​, consists of passive Gaussian states with Σ⪰(12+ν)I\Sigma\succeq(\tfrac12+\nu)IΣ⪰(21​+ν)I

Proved
BosonicLearn.fugacityFamily_promiseClass

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

bosonic-gaussian-statesquantum-informationquantum-learning

The states used for the single-copy lower bound belong to the class that the learner is promised.

Let 0<ν≤10<\nu\le10<ν≤1, 0<ξ<1/320<\xi<1/320<ξ<1/32, and let XXX be a traceless Hermitian n×nn\times nn×n matrix with ∥X∥op≤4\|X\|_{\mathrm{op}}\le4∥X∥op​≤4 (that is, X∈GsuppX\in G_{\mathrm{supp}}X∈Gsupp​). Put

RX=r (1+ξX),r=ν(1−4ξ)(1+ν).R_X=r\,(\mathbf 1+\xi X),\qquad r=\frac{\nu}{(1-4\xi)(1+\nu)} .RX​=r(1+ξX),r=(1−4ξ)(1+ν)ν​.

Then ρRX=det⁡(1−RX) Γ(RX)\rho_{R_X}=\det(\mathbf 1-R_X)\,\Gamma(R_X)ρRX​​=det(1−RX​)Γ(RX​) is a density operator on the nnn-mode Fock space; it is a passive Gaussian state, and its covariance matrix satisfies Σ⪰(12+ν)1\Sigma\succeq(\tfrac12+\nu)\mathbf 1Σ⪰(21​+ν)1.

Why it holds. Since ∥X∥op≤4\|X\|_{\mathrm{op}}\le4∥X∥op​≤4, r(1−4ξ)1⪯RX⪯r(1+4ξ)1≺1r(1-4\xi)\mathbf 1\preceq R_X\preceq r(1+4\xi)\mathbf 1\prec\mathbf 1r(1−4ξ)1⪯RX​⪯r(1+4ξ)1≺1 and r(1−4ξ)=ν/(1+ν)r(1-4\xi)=\nu/(1+\nu)r(1−4ξ)=ν/(1+ν). By Lemma 2, ρR\rho_RρR​ is a passive Gaussian state for 0⪯R≺10\preceq R\prec\mathbf 10⪯R≺1; its covariance has eigenvalues 12+λ/(1−λ)\tfrac12+\lambda/(1-\lambda)21​+λ/(1−λ) with λ\lambdaλ the eigenvalues of RRR, and λ≥ν/(1+ν)\lambda\ge\nu/(1+\nu)λ≥ν/(1+ν) gives λ/(1−λ)≥ν\lambda/(1-\lambda)\ge\nuλ/(1−λ)≥ν.

Role. A learner for the promise class in particular learns every ρRX\rho_{R_X}ρRX​​, X∈GsuppX\in G_{\mathrm{supp}}X∈Gsupp​; this is the first step of the proof of Theorem 5 from Proposition 5.

Formalization Note The conclusion is the existence of a State with coefficient array fugacityCoeff (fugacityFamily ν ξ X) which satisfies IsPassive and lies in promiseClass n ν. Γ(R)\Gamma(R)Γ(R) is gammaFock, GsuppG_{\mathrm{supp}}Gsupp​ is Gsupp n ⊆ Herm0 n with the operator norm schattenNorm ⊤. The upper bound Σ⪯(12+2ν)1\Sigma\preceq(\tfrac12+2\nu)\mathbf 1Σ⪯(21​+2ν)1 of the paper is not needed and not stated.

Preamble
import Mathlib
import Definitions.Def_BosonicLearn_fugacityCoeff
import Definitions.Def_BosonicLearn_fugacityFamily
import Definitions.Def_BosonicLearn_Gsupp
import Definitions.Def_BosonicLearn_IsPassive
import Definitions.Def_BosonicLearn_promiseClass

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

/-- **The hard family consists of passive Gaussian states of the promise class.**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, §5.1, Eqs. (30)–(31), p. 30: "Define the following family of fugacity matrices: R_X := r(I + ξX) for
X ∈ Herm_0(n) … r := ν/((1 − 4ξ)(1 + ν))", and proof of Theorem 5, p. 40: "Note this family satisfies
(1/2 + ν)I ≤ Σ ≤ (1/2 + 2ν)I as has been shown above."

For `0 < ν ≤ 1`, `0 < ξ < 1/32` and `X ∈ G_supp` (traceless Hermitian, `‖X‖_op ≤ 4`), the array
`ρ_{R_X} = det(I − R_X) Γ(R_X)`, `R_X = r(I + ξX)`, is the coefficient array of a density operator which is a
passive Gaussian state with covariance `Σ ⪰ (½ + ν) I`.

**Formalization Note.** The conclusion gives a `State` with these coefficients, `IsPassive`, and membership in
`promiseClass n ν`. Since `‖X‖_op ≤ 4`, `r(1 − 4ξ) I ⪯ R_X ⪯ r(1 + 4ξ) I ≺ I` and `r(1 − 4ξ) = ν/(1 + ν)`; a
passive state with fugacity `R` has covariance eigenvalues `½ + λ/(1 − λ)`, `λ` the eigenvalues of `R`
(Lemma 2). -/
theorem fugacityFamily_promiseClass (n : ℕ) (ν ξ : ℝ) (hν0 : 0 < ν) (hν1 : ν ≤ 1) (hξ0 : 0 < ξ)
    (hξ1 : ξ < 1 / 32) (X : Herm0 n) (hX : X ∈ Gsupp n) :
    ∃ ρ : State (Occupation (Fin n)), ρ.coefficients = fugacityCoeff (fugacityFamily ν ξ X) ∧
      IsPassive ρ ∧ ρ ∈ promiseClass n ν := by sorry

end BosonicLearn
Source
Chen, Anna Mele, Anna Mele, Preskill, When are bosonic Gaussian states classical to learn?, arXiv:2609.26705, §5.1, Eqs. (30)–(31), p. 30 (the family R_X = r(I + ξX)); proof of Theorem 5, p. 40 ("this family satisfies (1/2 + ν)I ≤ Σ ≤ (1/2 + 2ν)I"); Lemma 2, p. 20

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