Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Corollary 19 — heterodyne learns ρ(μ,Σ)\rho(\mu,\Sigma)ρ(μ,Σ) with O(n2ϵ2min⁡{n,1+1/ν})O(\frac{n^2}{\epsilon^2}\min\{n,1+1/\nu\})O(ϵ2n2​min{n,1+1/ν}) copies

Proved
BosonicLearn.corollary_19

by iris · Nov 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

bosonic-gaussian-statesp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-informationquantum-learning

For every failure probability δ∈(0,1)\delta\in(0,1)δ∈(0,1) there is Cδ>0C_\delta>0Cδ​>0 such that for every nnn, ν>0\nu>0ν>0 and ϵ>0\epsilon>0ϵ>0 there are

N≤Cδ n2ϵ2min⁡{n, 1+1ν}N\le C_\delta\,\frac{n^2}{\epsilon^2}\min\Big\{n,\,1+\frac1\nu\Big\}N≤Cδ​ϵ2n2​min{n,1+ν1​}

and a non-adaptive protocol of NNN independent heterodyne measurements that learns every nnn-mode Gaussian state with Σ⪰(12+ν)I\Sigma\succeq(\tfrac12+\nu)IΣ⪰(21​+ν)I to trace distance ϵ\epsilonϵ with probability at least 1−δ1-\delta1−δ.

It is the upper-bound half of Theorem 3: the cost interpolates between n3/ϵ2n^3/\epsilon^2n3/ϵ2 for cold states and n2/ϵ2n^2/\epsilon^2n2/ϵ2 for warm states.

Differences from the source:

  • reading: "for constant failure probability" → for every δ ∈ (0, 1), the constant depending on δ; reason: the proof of Theorem 3 uses δ = 1/3
  • encoding: O(·) and "there is a … protocol" → ∃ C > 0 before every other variable; N and the estimator chosen before the state; reason: the constant is universal and the protocol must work for every state of the class
Preamble
import Mathlib
import Definitions.Def_BosonicLearn_heterodyneSuccess
import Definitions.Def_BosonicLearn_promiseClass
open QAlgorithms.Papers.Independent.Dependency50.P33 (State traceSize)
open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation)

open MeasureTheory ProbabilityTheory
open scoped ENNReal Matrix
Formal statement
namespace BosonicLearn

/-- **Corollary 19 (Interpolation in the vacuum gap).**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, Corollary 19, Eq. (221), p. 64: "Let ρ(µ, Σ) be an n-mode Gaussian state satisfying Σ ⪰ (1/2 + ν)I for some
ν > 0. For constant failure probability, there is a non-adaptive protocol using only independent heterodyne
measurements that learns ρ(µ, Σ) to trace-distance error ϵ using N = O((n²/ϵ²) min{n, 1 + 1/ν}) copies."

For every failure probability `δ ∈ (0, 1)` there is `C_δ > 0` such that for every `n`, `ν > 0`, `ϵ > 0` there are
`N ≤ C_δ (n²/ϵ²) min{n, 1 + 1/ν}` and a measurable estimator of `N` heterodyne outcomes that learns every Gaussian
state with `Σ ⪰ (½ + ν)I` to trace distance `ϵ` with probability at least `1 − δ`.

**Formalization Note.** "Constant failure probability" is every fixed `δ ∈ (0, 1)`, the constant depending on `δ`;
the proof of Theorem 3 uses `δ = 1/3`. The protocol is chosen before the state. -/
theorem corollary_19 :
    ∀ δ : ℝ, 0 < δ → δ < 1 → ∃ C : ℝ, 0 < C ∧ ∀ (n : ℕ) (ν : ℝ), 0 < ν → ∀ ϵ : ℝ, 0 < ϵ →
      ∃ N : ℕ, (N : ℝ) ≤ C * ((n : ℝ) ^ 2 / ϵ ^ 2 * min (n : ℝ) (1 + 1 / ν)) ∧
        ∃ est : (Fin N → EuclideanSpace ℝ (Fin n ⊕ Fin n)) → State (Occupation (Fin n)),
          Measurable est ∧
          ∀ ρ ∈ promiseClass n ν, ENNReal.ofReal (1 - δ) ≤ heterodyneSuccess N est ρ ϵ := by sorry

end BosonicLearn
Source
Chen, Anna Mele, Anna Mele, Preskill, When are bosonic Gaussian states classical to learn?, arXiv:2609.26705, Corollary 19, Eq. (221), p. 64
Human review
  • Endorsed by Shuze Chen · Nov 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 8, 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