Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Supplementary Note 6, Claim: estimating the singlet fraction

Open
QAlgorithms.Papers.Independent.Intakenext100.P87.singlet_sampling

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

prove2quantum

Let 1≤m≤N1\le m\le N1≤m≤N and let ρ\rhoρ be any normalized state of NNN pairs, allowing arbitrary correlations among the pairs and with an external reference. Uniformly select mmm distinct positions, and independently choose one uniform axis from x,y,zx,y,zx,y,z for each selected pair. Let KKK count antiparallel outcomes and let FsingletF_{\rm singlet}Fsinglet​ be the mean probability of a local singlet projection over all NNN pairs. The source's estimator satisfies

F^=3K−m2m,E[F^]=Fsinglet.\widehat F=\frac{3K-m}{2m},\qquad\mathbb E[\widehat F]=F_{\rm singlet}.F=2m3K−m​,E[F]=Fsinglet​.

Moreover, for every fixed sample and axis choice, the entire coarse outcome law is unchanged by complete Bell premeasurement. This permits classical sampling analysis. No confidence-width formula or clipping of negative estimates is introduced.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P87
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P87

open scoped Matrix ComplexConjugate

open QAlgorithms.Measurements

/-- PDF pp.11,20--21, Supplementary Note 6 Claim: the unclipped estimator is unbiased;
its full coarse sampling law equals the Bell-premeasured law for arbitrary correlated inputs. -/
theorem singlet_sampling (N m : ℕ) (hm : 0 < m) (hNm : m ≤ N)
    (rho : Matrix (PairRegister N) (PairRegister N) ℂ) (hrho : IsDensityMatrix rho) :
    meanSingletEstimate rho m = meanLocalSingletWeight rho ∧
    ∀ (s : Fin m ↪ Fin N) (axes : Fin m → Fin 3) (anti : Qubits m),
      pairAxisSampleLaw (bellPinch rho) s axes anti = pairAxisSampleLaw rho s axes anti := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P87

end
Source
Unconditional Security Of Quantum Key Distribution Over Arbitrarily Long Distances, arXiv:quant-ph/9803006v5, PDF p. 21, Supplementary Note 6, Claim: estimating the singlet fraction
Read-back

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

For every pair of natural numbers N,mN,mN,m satisfying 0<m≤N0<m\leq N0<m≤N, and every complex matrix ρ\rhoρ on ΩN={0,1}N×{0,1}N\Omega_N=\{0,1\}^N\times\{0,1\}^NΩN​={0,1}N×{0,1}N that is positive semidefinite and has trace 111, the theorem asserts both

∑s,a,bLρ(s,a,b) 3∣{j:bj=true}∣−m2m∣{s:{0,…,m−1}↪{0,…,N−1}}∣ 3m=1N∑i=0N−1Re⁡tr⁡(Ei(vv†)ρ)andLB(ρ)(s,a,b)=Lρ(s,a,b) for every s,a,b.\frac{\displaystyle\sum_{s,a,b} L_\rho(s,a,b)\, \frac{3|\{j:b_j=\mathrm{true}\}|-m}{2m}} {\displaystyle|\{s:\{0,\ldots,m-1\}\hookrightarrow\{0,\ldots,N-1\}\}|\,3^m} = \frac1N\sum_{i=0}^{N-1} \operatorname{Re}\operatorname{tr}(E_i(vv^\dagger)\rho) \quad\text{and}\quad L_{\mathcal B(\rho)}(s,a,b)=L_\rho(s,a,b) \ \text{for every }s,a,b.∣{s:{0,…,m−1}↪{0,…,N−1}}∣3ms,a,b∑​Lρ​(s,a,b)2m3∣{j:bj​=true}∣−m​​=N1​i=0∑N−1​Retr(Ei​(vv†)ρ)andLB(ρ)​(s,a,b)=Lρ​(s,a,b) for every s,a,b.

The sum and final quantifiers range over all injections s:{0,…,m−1}↪{0,…,N−1}s:\{0,\ldots,m-1\}\hookrightarrow\{0,\ldots,N-1\}s:{0,…,m−1}↪{0,…,N−1}, all a∈{0,1,2}ma\in\{0,1,2\}^ma∈{0,1,2}m, and all Boolean strings b∈{0,1}mb\in\{0,1\}^mb∈{0,1}m. Here v=(∣01⟩−∣10⟩)/2v=(|01\rangle-|10\rangle)/\sqrt2v=(∣01⟩−∣10⟩)/2​, where 000 means false and 111 means true; Lτ(s,a,b)=Re⁡tr⁡(Ks,a,bτKs,a,b†)L_\tau(s,a,b)=\operatorname{Re}\operatorname{tr}(K_{s,a,b}\tau K_{s,a,b}^\dagger)Lτ​(s,a,b)=Retr(Ks,a,b​τKs,a,b†​), where Ks,a,b=Es(m−1)(Pa(m−1),b(m−1))⋯Es(0)(Pa(0),b(0))K_{s,a,b}=E_{s(m-1)}(P_{a(m-1),b(m-1)})\cdots E_{s(0)}(P_{a(0),b(0)})Ks,a,b​=Es(m−1)​(Pa(m−1),b(m−1)​)⋯Es(0)​(Pa(0),b(0)​) and Pk,c=12(I4+(−1)1{c=true}Ak⊗Ak)P_{k,c}=\tfrac12(I_4+(-1)^{\mathbf1_{\{c=\mathrm{true}\}}}A_k\otimes A_k)Pk,c​=21​(I4​+(−1)1{c=true}​Ak​⊗Ak​). The matrices are A0=X=(0110)A_0=X=\left(\begin{smallmatrix}0&1\\1&0\end{smallmatrix}\right)A0​=X=(01​10​), A1=Y=(0−ii0)A_1=Y=\left(\begin{smallmatrix}0&-\mathrm i\\\mathrm i&0\end{smallmatrix}\right)A1​=Y=(0i​−i0​), and A2=Z=(100−1)A_2=Z=\left(\begin{smallmatrix}1&0\\0&-1\end{smallmatrix}\right)A2​=Z=(10​0−1​). The embedding Ei(P)E_i(P)Ei​(P) uses PPP on the iii-th pair of bits and has zero entries unless the row and column agree at all other positions. Finally, B(ρ)=∑u,w∈{0,1}NBu,wρBu,w\mathcal B(\rho)=\sum_{u,w\in\{0,1\}^N}B_{u,w}\rho B_{u,w}B(ρ)=∑u,w∈{0,1}N​Bu,w​ρBu,w​, where Bu,w=∣βu,w⟩⟨βu,w∣B_{u,w}=|\beta_{u,w}\rangle\langle\beta_{u,w}|Bu,w​=∣βu,w​⟩⟨βu,w​∣, ∣βu,w⟩=(I⊗⨂iRui,wi) 2−N/2∑z∈{0,1}N∣z,z⟩|\beta_{u,w}\rangle=(I\otimes\bigotimes_i R_{u_i,w_i})\,2^{-N/2}\sum_{z\in\{0,1\}^N}|z,z\rangle∣βu,w​⟩=(I⊗⨂i​Rui​,wi​​)2−N/2∑z∈{0,1}N​∣z,z⟩, and R0,0=I2R_{0,0}=I_2R0,0​=I2​, R0,1=XR_{0,1}=XR0,1​=X, R1,0=ZR_{1,0}=ZR1,0​=Z, R1,1=YR_{1,1}=YR1,1​=Y. The hypotheses exclude m=0m=0m=0 and N=0N=0N=0.

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