Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Common high-SNR limit and the factor-d good-regime threshold shift

Open
QAlgorithms.Papers.Independent.Intakenext100.P96.good_regime_limit

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Fix an integer d≥1d\ge1d≥1, a reflection probability 0<η<10<\eta<10<η<1, and a unit single-photon input ∣ψ⟩|\psi\rangle∣ψ⟩. For each sufficiently small positive bbb, let ρ0,ρ1,σ0,σ1\rho_0,\rho_1,\sigma_0,\sigma_1ρ0​,ρ1​,σ0​,σ1​ be the retained unentangled and entangled pairs described by (1.0),(1.1),(8.0),(8.1), using the perfectly retained resource ∣Φd⟩=d−1/2∑k∣k⟩S∣k⟩A|\Phi_d\rangle=d^{-1/2}\sum_k|k\rangle_S|k\rangle_A∣Φd​⟩=d−1/2∑k​∣k⟩S​∣k⟩A​. In particular, ρ1=(1−η)ρ0+η∣ψ⟩⟨ψ∣\rho_1=(1-\eta)\rho_0+\eta|\psi\rangle\langle\psi|ρ1​=(1−η)ρ0​+η∣ψ⟩⟨ψ∣, σ0=ρ0⊗IA/d\sigma_0=\rho_0\otimes I_A/dσ0​=ρ0​⊗IA​/d and σ1=(1−η)σ0+η∣Φd⟩⟨Φd∣\sigma_1=(1-\eta)\sigma_0+\eta|\Phi_d\rangle\langle\Phi_d|σ1​=(1−η)σ0​+η∣Φd​⟩⟨Φd​∣.

For states a0,a1a_0,a_1a0​,a1​, write Q(a0,a1)=inf⁡0<s<1Tr⁡(a01−sa1s)Q(a_0,a_1)=\inf_{0<s<1}\operatorname{Tr}(a_0^{1-s}a_1^s)Q(a0​,a1​)=inf0<s<1​Tr(a01−s​a1s​), including one-sided limits at singular endpoints. Then

lim⁡b↓0Q(ρ0,ρ1)=lim⁡b↓0Q(σ0,σ1)=1−η.\lim_{b\downarrow0}Q(\rho_0,\rho_1) =\lim_{b\downarrow0}Q(\sigma_0,\sigma_1)=1-\eta.b↓0lim​Q(ρ0​,ρ1​)=b↓0lim​Q(σ0​,σ1​)=1−η.

The source's good-regime inequalities are, for every b>0b>0b>0,

η/b>1  ⟺  b<η,ηd/b>1  ⟺  b<dη.\eta/b>1\iff b<\eta,\qquad \eta d/b>1\iff b<d\eta.η/b>1⟺b<η,ηd/b>1⟺b<dη.

Thus the good regime has a noise threshold larger by a factor ddd. This specifies a fixed-ddd, fixed-η\etaη leading limit, with db<1db<1db<1 eventually, rather than a finite-ratio approximation or a claimed remainder.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P96
Formal statement
noncomputable section

open Filter

open scoped BigOperators Topology ComplexConjugate ComplexOrder Matrix.Norms.L2Operator

namespace QAlgorithms.Papers.Independent.Intakenext100.P96

attribute [local instance] Classical.propDecidable Classical.decEq

/-- PDF pp.4,8, repeated label (11) on p.8: with d, eta and the input fixed,
the high-SNR limit b tending to zero gives 1-eta for both factors. The good
regime's noise threshold is multiplied by d. Printed "d/b>1" for the
unentangled case on p.8 is corrected to "eta/b>1", consistently with p.3.
This states no O((db)^2) remainder. -/
theorem good_regime_limit (d : ℕ) (hd : 0 < d) (eta : ℝ)
    (heta0 : 0 < eta) (heta1 : eta < 1)
    (psi : EuclideanSpace ℂ (Fin d)) (hpsi : IsState psi) :
    Tendsto (fun b : ℝ => illuminationChernoffFactor (background d b) (signalPresent b eta psi))
      (nhdsWithin 0 (Set.Ioi 0)) (nhds (1 - eta)) ∧
    Tendsto (fun b : ℝ => illuminationChernoffFactor (pairAbsent d b) (pairPresent d b eta))
      (nhdsWithin 0 (Set.Ioi 0)) (nhds (1 - eta)) ∧
    (∀ b : ℝ, 0 < b →
      (1 < eta / b ↔ b < eta) ∧
      (1 < eta * (d : ℝ) / b ↔ b < (d : ℝ) * eta)) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P96

end
Source
Quantum Illumination, arXiv:0803.2022v2, PDF p. 8, Good-regime leading Chernoff limit; repeated label (11) on p.8
Read-back

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

For every natural number d>0d>0d>0, every real η\etaη with 0<η<10<\eta<10<η<1, and every vector ψ∈Cd\psi\in\mathbb C^dψ∈Cd with Euclidean norm exactly 111, let Bb=diag⁡(1−db,b,…,b)B_b=\operatorname{diag}(1-db,b,\ldots,b)Bb​=diag(1−db,b,…,b) on the space with basis e∗,e0,…,ed−1e_\ast,e_0,\ldots,e_{d-1}e∗​,e0​,…,ed−1​, let ψ~\widetilde\psiψ​ have coordinate 000 at ∗\ast∗ and coordinates ψi\psi_iψi​ at iii, and set

Sb=(1−η)Bb+η ψ~ψ~∗,Ω=1d∑i=0d−1ei⊗ei,R0,b=Bb⊗Idd,R1,b=(1−η)R0,b+η ΩΩ∗.S_b=(1-\eta)B_b+\eta\,\widetilde\psi\widetilde\psi^\ast,\qquad \Omega=\frac1{\sqrt d}\sum_{i=0}^{d-1}e_i\otimes e_i,\qquad R_{0,b}=B_b\otimes\frac{I_d}{d},\qquad R_{1,b}=(1-\eta)R_{0,b}+\eta\,\Omega\Omega^\ast .Sb​=(1−η)Bb​+ηψ​ψ​∗,Ω=d​1​i=0∑d−1​ei​⊗ei​,R0,b​=Bb​⊗dId​​,R1,b​=(1−η)R0,b​+ηΩΩ∗.

Define C(A,C)=inf⁡0<s<1Re⁡tr⁡(CsA1−s)\mathcal C(A,C)=\inf_{0<s<1}\operatorname{Re}\operatorname{tr}(C^sA^{1-s})C(A,C)=inf0<s<1​Retr(CsA1−s), with matrix powers given by continuous functional calculus of the corresponding real power functions. The declaration asserts

lim⁡b→0b>0C(Bb,Sb)=1−η,lim⁡b→0b>0C(R0,b,R1,b)=1−η,\lim_{\substack{b\to0\\b>0}}\mathcal C(B_b,S_b)=1-\eta, \qquad \lim_{\substack{b\to0\\b>0}}\mathcal C(R_{0,b},R_{1,b})=1-\eta,b→0b>0​lim​C(Bb​,Sb​)=1−η,b→0b>0​lim​C(R0,b​,R1,b​)=1−η,

and, for every real b>0b>0b>0, both equivalences

1<ηb ⟺ b<η,1<ηdb ⟺ b<dη.1<\frac{\eta}{b}\ \Longleftrightarrow\ b<\eta, \qquad 1<\frac{\eta d}{b}\ \Longleftrightarrow\ b<d\eta .1<bη​ ⟺ b<η,1<bηd​ ⟺ b<dη.

The limits approach zero through positive real values of bbb; the statement imposes no additional restriction db≤1db\le1db≤1 on that approach or on the universally quantified positive bbb. It includes d=1d=1d=1.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 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