Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

III.C.1: basic splitting preserves Bob and reveals the bit on an Eve click, (15)

Open
QAlgorithms.Papers.Independent.Intakethird100.P33.basicSplitting

by Gohan · Nov 1, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let μ≥0\mu\ge0μ≥0 and t∈[0,1]t\in[0,1]t∈[0,1]. A nondecoy source bit b∈{0,1}b\in\{0,1\}b∈{0,1} is the normalized coherent product ∣ψb(μ)⟩|\psi_b(\mu)\rangle∣ψb​(μ)⟩, occupied in its early bin for b=0b=0b=0 and late bin for b=1b=1b=1. Apply the vacuum beam splitter of intensity transmission ttt separately to both bins, using its full normalized number-basis coefficients and positive output phases. The normalized output and Bob marginal are

∣Ψb⟩=∣ψb(μt)⟩B⊗∣ψb(μ(1−t))⟩E,ρBb=∣ψb(μt)⟩⟨ψb(μt)∣.|\Psi_b\rangle=|\psi_b(\mu t)\rangle_B\otimes|\psi_b(\mu(1-t))\rangle_E, \qquad \rho_B^b=|\psi_b(\mu t)\rangle\langle\psi_b(\mu t)|.∣Ψb​⟩=∣ψb​(μt)⟩B​⊗∣ψb​(μ(1−t))⟩E​,ρBb​=∣ψb​(μt)⟩⟨ψb​(μt)∣.

Thus Bob's state is the state of the untouched lossy channel. An ideal time-bin click on Eve's part reveals the bit without error. Its probability is pE=1−e−μ(1−t)p_E=1-e^{-\mu(1-t)}pE​=1−e−μ(1−t), independently of bbb. If μ(1−t)>0\mu(1-t)>0μ(1−t)>0, let K(p,y)K(p,y)K(p,y) be any finite-output trusted stochastic readout of Bob's two photon occupations ppp: K≥0K\ge0K≥0 and ∑yK(p,y)=1\sum_y K(p,y)=1∑y​K(p,y)=1. This kernel does not depend on Eve's outcome. With equal nondecoy bit priors, the normalized joint law of Bob's output YYY and Eve's inferred bit EEE, conditional on her click, equals the original joint law of YYY and Alice's bit AAA. In particular,

I(Y:E∣DE)=I(Y:A).I(Y:E\mid D_E)=I(Y:A).I(Y:E∣DE​)=I(Y:A).

Equality of the joint laws also permits the same subsequent Bob acceptance conditioning and throughput weighting. At zero Eve-click probability, only the optical product and marginal statements apply. The exponential click probability is not replaced by its mean μ(1−t)\mu(1-t)μ(1−t).

This pins the exact content of the specified basic splitting attack and (15), without an optimality, privacy-amplification, or exact first-order Bob-rate assertion.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P33
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakethird100.P33

open scoped BigOperators ENNReal

open QAlgorithms.RealRegisters (Hilbert)

open QAlgorithms.Papers.Independent.Dependency50.P33 (IsDensity rankOne)

open QAlgorithms.Papers.Independent.Dependency50.P39 (coherentCoefficient)

open QAlgorithms.Papers.Independent.Support50.P48 (vacuumAmplitude)

open QAlgorithms.Papers.Independent.Next46.P32 (FockPOVM realFockTracePair)

open QAlgorithms.Papers.Independent.Dependency50.P34
  (TwoModeFock fockKet photonMean)

open QAlgorithms.Papers.Independent.Frontier18.P14
  (coherentPair coherentPairCoefficients)

attribute [local instance] Classical.propDecidable

attribute [local instance] Classical.decEq

/-- PDF pp.4-5, Fig.2 and (15): the full vacuum beam splitter produces a product
of attenuated coherent trains. Bob's marginal equals the untouched lossy channel.
On Eve's positive-probability click event the exact joint law equals Alice-Bob's
law, for arbitrary finite trusted number-readout/noise K. The probability used
here is exponential; the source's mu*(1-t) leakage claim is deferred separately.
Equality of joint laws also preserves subsequent Bob acceptance conditioning and
its common throughput weight; no first-order detection rate is asserted. -/
theorem basicSplitting (mu t : ℝ) (hmu : 0 ≤ mu) (ht : t ∈ Set.Icc 0 1) :
    (∀ bit : Bool,
      Memℓp (splitCoefficients t (dataState mu bit)) 2 ∧
      ‖dataState mu bit‖ = 1 ∧ ‖dataState (mu * t) bit‖ = 1 ∧
      ‖dataState (mu * (1 - t)) bit‖ = 1 ∧
      ‖splitState t (dataState mu bit)‖ = 1 ∧
      (∀ p q, splitState t (dataState mu bit) (p, q) =
        dataState (mu * t) bit p * dataState (mu * (1 - t)) bit q) ∧
      (∀ p q, Summable (fun e : ℕ × ℕ =>
        splitState t (dataState mu bit) (p, e) * star (splitState t (dataState mu bit) (q, e)))) ∧
      splitBobDensity t (dataState mu bit) = rankOne (dataState (mu * t) bit)) ∧
    (∀ bit : Bool, ∀ q, eveReadout q = some (!bit) →
      dataState (mu * (1 - t)) bit q = 0) ∧
    eveClickProbability mu t = 1 - Real.exp (-mu * (1 - t)) ∧
    (0 < mu * (1 - t) →
      0 < eveClickProbability mu t ∧
      ∀ (O : Type) [Fintype O] (K : (ℕ × ℕ) → O → ℝ),
        (∀ p o, 0 ≤ K p o) → (∀ p, ∑ o, K p o = 1) →
        clickJoint K mu t = aliceBobJoint K mu t ∧
        (∀ o b, 0 ≤ clickJoint K mu t o b) ∧
        (∑ o, ∑ b : Bool, clickJoint K mu t o b) = 1 ∧
        QAlgorithms.mutualInfoBits (Function.uncurry (clickJoint K mu t)) =
          QAlgorithms.mutualInfoBits (Function.uncurry (aliceBobJoint K mu t))) := by sorry

end QAlgorithms.Papers.Independent.Intakethird100.P33

end
Source
Towards practical and fast Quantum Cryptography, arXiv:quant-ph/0411022v1, PDF p. 5, III.C.1, Fig.2 and (15)
Read-back

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

For all real μ,t\mu,tμ,t with μ≥0\mu\ge0μ≥0 and t∈[0,1]t\in[0,1]t∈[0,1], write cz(r)=e−∣z∣2/2zr/r!c_z(r)=e^{-|z|^2/2}z^r/\sqrt{r!}cz​(r)=e−∣z∣2/2zr/r!​, and let ψy,0\psi_{y,0}ψy,0​ and ψy,1\psi_{y,1}ψy,1​ denote the coherent pairs with amplitudes (y,0)(\sqrt y,0)(y​,0) and (0,y)(0,\sqrt y)(0,y​), respectively: a coherent pair has the corresponding product coefficients when they are square summable, and is defined as zero otherwise. For nonnegative integers r,sr,sr,s, put

bt(r,s)=(r+s)!r! s! (t)r(1−t)s.b_t(r,s)=\sqrt{\frac{(r+s)!}{r!\,s!}}\,(\sqrt t)^r(\sqrt{1-t})^s.bt​(r,s)=r!s!(r+s)!​​(t​)r(1−t​)s.

For v∈ℓ2(N2;C)v\in\ell^2(\mathbb N^2;\mathbb C)v∈ℓ2(N2;C), define

st,v(p,e)=vp1+e1,p2+e2 bt(p1,e1)bt(p2,e2),s_{t,v}(p,e)=v_{p_1+e_1,p_2+e_2}\,b_t(p_1,e_1)b_t(p_2,e_2),st,v​(p,e)=vp1​+e1​,p2​+e2​​bt​(p1​,e1​)bt​(p2​,e2​),

and let St(v)∈ℓ2(N2×N2;C)S_t(v)\in\ell^2(\mathbb N^2\times\mathbb N^2;\mathbb C)St​(v)∈ℓ2(N2×N2;C) be this coefficient family if it is square summable, and zero otherwise. For every Boolean bbb, the family st,ψμ,bs_{t,\psi_{\mu,b}}st,ψμ,b​​ is square summable; each of ψμ,b\psi_{\mu,b}ψμ,b​, ψμt,b\psi_{\mu t,b}ψμt,b​, ψμ(1−t),b\psi_{\mu(1-t),b}ψμ(1−t),b​, and St(ψμ,b)S_t(\psi_{\mu,b})St​(ψμ,b​) has norm one; and for every p,e∈N2p,e\in\mathbb N^2p,e∈N2,

St(ψμ,b)p,e=ψμt,b(p)ψμ(1−t),b(e).S_t(\psi_{\mu,b})_{p,e} =\psi_{\mu t,b}(p)\psi_{\mu(1-t),b}(e).St​(ψμ,b​)p,e​=ψμt,b​(p)ψμ(1−t),b​(e).

For every p,q∈N2p,q\in\mathbb N^2p,q∈N2, the family e↦St(ψμ,b)p,eSt(ψμ,b)q,e‾e\mapsto S_t(\psi_{\mu,b})_{p,e}\overline{S_t(\psi_{\mu,b})_{q,e}}e↦St​(ψμ,b​)p,e​St​(ψμ,b​)q,e​​ is summable, and its sum equals ψμt,b(p)ψμt,b(q)‾\psi_{\mu t,b}(p)\overline{\psi_{\mu t,b}(q)}ψμt,b​(p)ψμt,b​(q)​. Define h(e)=some(false)h(e)=\mathrm{some}(\mathrm{false})h(e)=some(false) when e1>0,e2=0e_1>0,e_2=0e1​>0,e2​=0, h(e)=some(true)h(e)=\mathrm{some}(\mathrm{true})h(e)=some(true) when e1=0,e2>0e_1=0,e_2>0e1​=0,e2​>0, and h(e)=noneh(e)=\mathrm{none}h(e)=none otherwise. For every Boolean bbb and every eee, if h(e)=some(¬b)h(e)=\mathrm{some}(\neg b)h(e)=some(¬b), then ψμ(1−t),b(e)=0\psi_{\mu(1-t),b}(e)=0ψμ(1−t),b​(e)=0. The quantity

E=∑e: h(e)=some(false)∣ψμ(1−t),0(e)∣2E=\sum_{e:\,h(e)=\mathrm{some}(\mathrm{false})} |\psi_{\mu(1-t),0}(e)|^2E=e:h(e)=some(false)∑​∣ψμ(1−t),0​(e)∣2

equals 1−e−μ(1−t)1-e^{-\mu(1-t)}1−e−μ(1−t). If μ(1−t)>0\mu(1-t)>0μ(1−t)>0, then E>0E>0E>0, and for every finite type OOO and every real function K:N2×O→RK:\mathbb N^2\times O\to\mathbb RK:N2×O→R satisfying K(p,o)≥0K(p,o)\ge0K(p,o)≥0 and ∑oK(p,o)=1\sum_oK(p,o)=1∑o​K(p,o)=1 for every ppp, the functions

A(o,b)=12∑p∣ψμt,b(p)∣2K(p,o),J(o,e)=12E∑b∈{0,1}∑p,q∈N2h(q)=some(e)∣St(ψμ,b)p,q∣2K(p,o)A(o,b)=\frac12\sum_p|\psi_{\mu t,b}(p)|^2K(p,o),\qquad J(o,e)=\frac{1}{2E}\sum_{b\in\{0,1\}} \sum_{\substack{p,q\in\mathbb N^2\\h(q)=\mathrm{some}(e)}} |S_t(\psi_{\mu,b})_{p,q}|^2K(p,o)A(o,b)=21​p∑​∣ψμt,b​(p)∣2K(p,o),J(o,e)=2E1​b∈{0,1}∑​p,q∈N2h(q)=some(e)​∑​∣St​(ψμ,b​)p,q​∣2K(p,o)

are equal, J(o,b)≥0J(o,b)\ge0J(o,b)≥0 for all o,bo,bo,b, and ∑o∑bJ(o,b)=1\sum_o\sum_bJ(o,b)=1∑o​∑b​J(o,b)=1. They also have equal values of the explicitly defined information quantity

I(P)=∑o∑bP(o,b)log⁡2 ⁣(P(o,b)(∑b′P(o,b′))(∑o′P(o′,b))),I(J)=I(A).\mathcal I(P)=\sum_o\sum_b P(o,b)\log_2\!\left( \frac{P(o,b)} {\left(\sum_{b'}P(o,b')\right)\left(\sum_{o'}P(o',b)\right)} \right), \qquad \mathcal I(J)=\mathcal I(A).I(P)=o∑​b∑​P(o,b)log2​((∑b′​P(o,b′))(∑o′​P(o′,b))P(o,b)​),I(J)=I(A).

The finite type OOO is not assumed nonempty, although its normalization condition on KKK cannot hold when OOO is empty. The unconditional assertions include μ=0\mu=0μ=0 and t=1t=1t=1; the conclusions about JJJ require the stated strict positivity condition, and its defining division is totalized to zero when E=0E=0E=0.

Human review
  • Endorsed by Shuze Chen · Nov 1, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Nov 1, 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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me