Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Erasure cpp coding: exact integer-message fidelity and normal approximation

Open
QAlgorithms.Papers.Independent.Support50.P04.erasure_exact_and_expansion

by Goku · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let N:A→B\mathcal N:A\to BN:A→B be a CPTP channel on finite positive-dimensional complex Hilbert spaces with fixed orthonormal bases. Write R^(n;ε)\widehat R(n;\varepsilon)R(n;ε) and R^cpp(n;ε)\widehat R^{\mathrm{cpp}}(n;\varepsilon)Rcpp(n;ε) for the maximum unassisted and classical-post-processing assisted entanglement-transmission rates. A code has a positive integer message dimension MMM, rate log⁡2M/n\log_2 M/nlog2​M/n, and squared fidelity to the normalized maximally entangled target at least 1−ε1-\varepsilon1−ε. A cpp encoder may retain sender memory; its decoder is LOCC across that memory and the receiver outputs, with its final quantum output at the receiver and no preshared entanglement. For the pure-flag qubit erasure channel Eβ(X)=(1−β)X⊕βtr⁡(X)∣e⟩⟨e∣\mathcal E_\beta(X)=(1-\beta)X\oplus\beta\operatorname{tr}(X)|e\rangle\langle e|Eβ​(X)=(1−β)X⊕βtr(X)∣e⟩⟨e∣, with 0≤β≤10\le\beta\le10≤β≤1, every integer n,M≥1n,M\ge1n,M≥1 has optimal cpp squared fidelity

Fopt(n,M)=∑l=0n(nl)βl(1−β)n−lmin⁡{1,2n−lM}.F_{\mathrm{opt}}(n,M)=\sum_{l=0}^n\binom nl\beta^l(1-\beta)^{n-l}\min\left\{1,\frac{2^{n-l}}M\right\}.Fopt​(n,M)=l=0∑n​(ln​)βl(1−β)n−lmin{1,M2n−l​}.

An actual cpp code with this message dimension attains this value. Fix 0<ε<10<\varepsilon<10<ε<1. All logarithms are binary and Φ−1\Phi^{-1}Φ−1 is the standard Gaussian quantile. Remainder constants may depend on the fixed channel and error, and are chosen before the block length: there are C>0C>0C>0 and n0≥2n_0\ge2n0​≥2 such that the displayed bound holds for every integer n≥n0n\ge n_0n≥n0​. If 0<β<10<\beta<10<β<1, then

∣R^Eβcpp(n;ε)−1+β−β(1−β)/n Φ−1(ε)∣≤C/n.\left|\widehat R_{\mathcal E_\beta}^{\mathrm{cpp}}(n;\varepsilon)-1+\beta-\sqrt{\beta(1-\beta)/n}\,\Phi^{-1}(\varepsilon)\right|\le C/n.​REβ​cpp​(n;ε)−1+β−β(1−β)/n​Φ−1(ε)​≤C/n.

The finite-block formula retains both channel endpoints and integer-message rounding.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P04
Formal statement
noncomputable section

open scoped Matrix ComplexConjugate ComplexOrder

open QAlgorithms

namespace QAlgorithms.Papers.Independent.Support50.P04

variable {A B E : Type} [Fintype A] [DecidableEq A]

variable [Fintype B] [DecidableEq B] [Fintype E] [DecidableEq E]

/-- arXiv:1504.04617v2, PDF pp.5–6,18–19, Result 2, Eqs.(8)–(9),(69)–(70).
Printed Eq.(8) is an inversion shorthand; changed to exact fixed-INTEGER-M
optimal fidelity and attainment. Integer rounding need not saturate epsilon.
Only the Gaussian clause excludes beta=0,1; the exact clause includes them. -/
theorem erasure_exact_and_expansion (beta : ℝ) (hb0 : 0 ≤ beta) (hb1 : beta ≤ 1) :
    (∀ n M : ℕ, 0 < n → 0 < M →
      optimalCppFidelity (QAlgorithms.Papers.Independent.Renew46.P36.flaggedErasure (A := Fin 2) beta) n M = (erasureFidelityFormula beta n M : EReal) ∧
      ∃ C : CppCode (Fin n → Fin 2) (Fin n → Option (Fin 2)),
        C.M = M ∧ cppFidelity (piSuperop (fun _ : Fin n => (QAlgorithms.Papers.Independent.Renew46.P36.flaggedErasure (A := Fin 2) beta))) C = erasureFidelityFormula beta n M) ∧
    (0 < beta → beta < 1 → ∀ epsilon : ℝ, 0 < epsilon → epsilon < 1 →
      constantRemainderExpansion (fun n => cppBoundary (QAlgorithms.Papers.Independent.Renew46.P36.flaggedErasure (A := Fin 2) beta) n epsilon)
        (1-beta) (beta*(1-beta)) epsilon) := by sorry

end QAlgorithms.Papers.Independent.Support50.P04

end
Source
Quantum Coding with Finite Resources, arXiv:1504.04617v2, PDF p. 6, Result 2, Eqs. (8)–(9), (69)–(70)
Read-back

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

For every real β\betaβ with 0≤β≤10\le\beta\le10≤β≤1, let Eβ\mathcal E_\betaEβ​ map 2×22\times22×2 matrices to matrices indexed by {⊥,0,1}\{\bot,0,1\}{⊥,0,1}, with ordinary 2×22\times22×2 block (1−β)X(1-\beta)X(1−β)X, entry (⊥,⊥)(\bot,\bot)(⊥,⊥) equal to βtr⁡X\beta\operatorname{tr}XβtrX, and all cross entries zero. A code for a channel N\mathcal NN consists of positive integers M,qM,qM,q, a channel A:M→(input×q)\mathcal A:M\to(\text{input}\times q)A:M→(input×q), and a channel D:(q×output)→M\mathcal D:(q\times\text{output})\to MD:(q×output)→M; channels are finite Kraus sums X↦∑jEjXEj†X\mapsto\sum_jE_jXE_j^\daggerX↦∑j​Ej​XEj†​ with ∑jEj†Ej=I\sum_jE_j^\dagger E_j=I∑j​Ej†​Ej​=I. The decoder must satisfy this additional receiver rule: there are positive integers a,ba,ba,b, isometries W:q→aW:q\to aW:q→a, V:output→bV:\text{output}\to bV:output→b, a map L\mathcal LL on a×ba\times ba×b, and a channel F:b→M\mathcal F:b\to MF:b→M such that D(X)=F(tr⁡aL((W⊗V)X(W⊗V)†))\mathcal D(X)=\mathcal F(\operatorname{tr}_a\mathcal L((W\otimes V)X(W\otimes V)^\dagger))D(X)=F(tra​L((W⊗V)X(W⊗V)†)). The required property of L\mathcal LL is the following countable-history limit condition. An instrument has completely positive complex-linear branches, its branch outputs are summable for every matrix, and their sum preserves trace; complete positivity means positivity after amplification by every finite identity register. A local round has branches that are tensor products of two local maps, with one party’s maps forming an instrument and the other party’s map a channel for each outcome. A raw history instrument starts with a local round and extends by applying an outcome-dependent local round to each previous branch, encoding each pair of successive history labels by the fixed pairing of natural numbers. There must be a sequence KjK_jKj​ of raw history instruments, each having a positive finite number of rounds and each extending its predecessor by one such round, and single-outcome instruments AjA_jAj​ obtained from KjK_jKj​ by coarse-graining their outcomes, such that

sup⁡ω≥0, tr⁡ω=1∥(id⊗(Aj−L))(ω)∥1⟶0,\sup_{\omega\ge0,\ \operatorname{tr}\omega=1} \bigl\|(\mathrm{id}\otimes(A_j-\mathcal L))(\omega)\bigr\|_1\longrightarrow0,ω≥0, trω=1sup​​(id⊗(Aj​−L))(ω)​1​⟶0,

where the identity register has the same dimension as a×ba\times ba×b, and coarse-graining means summing branches over fibers of a function on outcome labels. For such a code, define T=D∘(idq⊗N)∘swap∘A\mathcal T=\mathcal D\circ(\mathrm{id}_q\otimes\mathcal N)\circ\mathrm{swap}\circ\mathcal AT=D∘(idq​⊗N)∘swap∘A, where the swap puts the retained qqq-register first, and define F(N,C)=Re⁡tr⁡[ΦM(id⊗T)(ΦM)]F(\mathcal N,C)=\operatorname{Re}\operatorname{tr}[\Phi_M(\mathrm{id}\otimes\mathcal T)(\Phi_M)]F(N,C)=Retr[ΦM​(id⊗T)(ΦM​)], with ΦM=M−1∑i,j∣i,i⟩⟨j,j∣\Phi_M=M^{-1}\sum_{i,j}|i,i\rangle\langle j,j|ΦM​=M−1∑i,j​∣i,i⟩⟨j,j∣. The theorem asserts that for every natural n>0n>0n>0 and M>0M>0M>0, the supremum in the extended reals of these fidelities over all such codes for Eβ⊗n\mathcal E_\beta^{\otimes n}Eβ⊗n​ having logical dimension exactly MMM equals the real number

fβ,n,M=∑ℓ=0n(nℓ)βℓ(1−β)n−ℓmin⁡ ⁣(1,2n−ℓM)f_{\beta,n,M}=\sum_{\ell=0}^n\binom n\ell\beta^\ell(1-\beta)^{n-\ell} \min\!\left(1,\frac{2^{n-\ell}}M\right)fβ,n,M​=ℓ=0∑n​(ℓn​)βℓ(1−β)n−ℓmin(1,M2n−ℓ​)

embedded in the extended reals, and that an individual such code attains this number. Moreover, if 0<β<10<\beta<10<β<1, then for every 0<ϵ<10<\epsilon<10<ϵ<1, writing

Bn=sup⁡{R∈R:∃C, R≤log⁡2C.Mn, 1−ϵ≤F(Eβ⊗n,C)}B_n=\sup\left\{R\in\mathbb R:\exists C,\ R\le\frac{\log_2 C.M}{n},\ 1-\epsilon\le F(\mathcal E_\beta^{\otimes n},C)\right\}Bn​=sup{R∈R:∃C, R≤nlog2​C.M​, 1−ϵ≤F(Eβ⊗n​,C)}

as an extended-real supremum, there exist C0>0C_0>0C0​>0 and a natural n0≥2n_0\ge2n0​≥2 such that every n≥n0n\ge n_0n≥n0​ satisfies

1−β+β(1−β)n zϵ−C0n≤Bn≤1−β+β(1−β)n zϵ+C0n.1-\beta+\sqrt{\frac{\beta(1-\beta)}n}\,z_\epsilon-\frac{C_0}n \le B_n\le 1-\beta+\sqrt{\frac{\beta(1-\beta)}n}\,z_\epsilon+\frac{C_0}n.1−β+nβ(1−β)​​zϵ​−nC0​​≤Bn​≤1−β+nβ(1−β)​​zϵ​+nC0​​.

Here log⁡2x=ln⁡x/ln⁡2\log_2x=\ln x/\ln2log2​x=lnx/ln2, zϵz_\epsilonzϵ​ is the infimum of the real zzz for which the standard normal integral up to zzz is at least ϵ\epsilonϵ, and ∥X∥1=Re⁡tr⁡X†X\|X\|_1=\operatorname{Re}\operatorname{tr}\sqrt{X^\dagger X}∥X∥1​=RetrX†X​. The exact finite-block assertion includes β=0,1\beta=0,1β=0,1, with zeroth powers equal to 111; the expansion requires the strict interior conditions.

Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 24, 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