Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem II.2: For 0<lambda<1, |log2 N(n,lambda)/n−sum_i C(Wi)/n|→0; C(Wi)=sup_P[H(PWi)−H(Wi|P)].

Open
QAlgorithms.Papers.Independent.Renew46.P43.nonstationaryCoding

by Goku · Oct 22, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Fix a nonzero finite output system of dimension ddd. The nonstationary channels WiW_iWi​ are nonempty measurable state sets; C(Wi)=sup⁡PI(P;Wi)C(W_i)=\sup_PI(P;W_i)C(Wi​)=supP​I(P;Wi​) optimizes over ALL probability laws on that set. A code has temporal product codewords, a collective block POVM with optional failure, positive length and a nonempty finite message set. N(n,λ)N(n,\lambda)N(n,λ) optimizes actual codes under maximum error, with 0<λ<10<\lambda<10<λ<1. In the maximal-code statements, 0<τ<10<\tau<10<τ<1, the laws PiP_iPi​ are arbitrary, and the measurable event AAA has product mass at least τ\tauτ; constants K′,δ>0K',\delta>0K′,δ>0 are chosen before the channel, laws, event and length. In the converse ϵ>0\epsilon>0ϵ>0 and the threshold is uniform over channel sequences. The hybrid version uses intrinsic output trace and algebra-valued decoding.

∣log⁡2N(n,λ)n−1n∑i=1nC(Wi)∣⟶0.\left|\frac{\log_2N(n,\lambda)}n-\frac1n\sum_{i=1}^nC(W_i)\right|\longrightarrow0.​nlog2​N(n,λ)​−n1​i=1∑n​C(Wi​)​⟶0.

This is the source result with its stated operational or information-theoretic domain.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Renew46_P43
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Renew46.P43

open scoped BigOperators Matrix ComplexConjugate ComplexOrder ENNReal Matrix.Norms.L2Operator

open Filter MeasureTheory

open QAlgorithms.Papers.Independent.Information100.P08 (IsPure)

open QAlgorithms.Papers.Independent.Information100.P06 (Below)

open QAlgorithms.Papers.Independent.Information100.P60 (copies relativeBits)

open QAlgorithms.Papers.Independent.Dependency50.P19 (count both)

open QAlgorithms.Papers.Independent.Information100.P22 (coherentInformation)

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms.Papers.Independent.Claimed50.P40 (InBlocks)

attribute [local instance] QAlgorithms.Papers.Independent.Renew46.P43.matrixMeasurableSpace

/-- Theorem II.2, PDF p.24: absolute difference between operational rate and the
mean single-use capacities tends to zero for every fixed error in (0,1). -/
theorem nonstationaryCoding (d : ℕ) (hd : 0 < d) (W : ℕ → StateChannel (fullAlgebra d hd))
    (lambda : ℝ) (hlambda : 0 < lambda) (hsmall : lambda < 1) :
    Tendsto (fun n : ℕ => |Real.logb 2 (nonstationaryN W (n+1) lambda) / (n+1 : ℝ) -
      (∑ i : Fin (n+1), (W i.val).capacity) / (n+1 : ℝ)|) atTop (nhds 0) := by sorry

end QAlgorithms.Papers.Independent.Renew46.P43

end
Source
Coding Theorems of Quantum Information Theory, arXiv:quant-ph/9907077v1, PDF p. 24, Theorem II.2
Read-back

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

For every natural number d>0d>0d>0, every sequence W=(Wi)i∈NW=(\mathcal W_i)_{i\in\mathbb N}W=(Wi​)i∈N​ of nonempty measurable sets of positive semidefinite, trace-one complex d×dd\times dd×d matrices, and every real number 0<λ<10<\lambda<10<λ<1, the following limit holds:

lim⁡n→∞∣log⁡2NW(n+1,λ)n+1−1n+1∑i=0nC(Wi)∣=0.\lim_{n\to\infty} \left| \frac{\log_2 N_W(n+1,\lambda)}{n+1} -\frac{1}{n+1}\sum_{i=0}^{n}C(\mathcal W_i) \right|=0.n→∞lim​​n+1log2​NW​(n+1,λ)​−n+11​i=0∑n​C(Wi​)​=0.

Here all block labels are 000, so the block restrictions impose no additional zero-entry conditions. For a positive length ℓ\ellℓ, NW(ℓ,λ)N_W(\ell,\lambda)NW​(ℓ,λ) is the natural-number supremum of positive message counts MMM for which one can choose matrices ρm,i∈Wi\rho_{m,i}\in\mathcal W_iρm,i​∈Wi​, 0≤i<ℓ0\le i<\ell0≤i<ℓ, and positive semidefinite decoding matrices Dnone,Dsome(0),…,Dsome(M−1)D_{\mathrm{none}},D_{\mathrm{some}(0)},\ldots,D_{\mathrm{some}(M-1)}Dnone​,Dsome(0)​,…,Dsome(M−1)​ summing to the identity on the word space, such that max⁡m[1−Re⁡tr⁡((⨂iρm,i)Dsome(m))]≤λ\max_m[1-\operatorname{Re}\operatorname{tr}((\bigotimes_i\rho_{m,i})D_{\mathrm{some}(m)})]\le\lambdamaxm​[1−Retr((⨂i​ρm,i​)Dsome(m)​)]≤λ; the tensor product has entries ∏iρm,i(ai,bi)\prod_i\rho_{m,i}(a_i,b_i)∏i​ρm,i​(ai​,bi​), and the natural-number supremum is 000 for an empty or unbounded set. The quantity C(W)C(\mathcal W)C(W) is the real supremum, over all probability measures PPP on the set W\mathcal WW, of

(∫W−D2(ρ∥ρˉP) dP(ρ))toReal,ρˉP=∫Wρ dP(ρ).\left(\int_{\mathcal W}^{-}D_2(\rho\Vert\bar\rho_P)\,dP(\rho)\right)_{\mathrm{toReal}}, \qquad \bar\rho_P=\int_{\mathcal W}\rho\,dP(\rho).(∫W−​D2​(ρ∥ρˉ​P​)dP(ρ))toReal​,ρˉ​P​=∫W​ρdP(ρ).

The first integral is an extended nonnegative integral, the matrix average is a Bochner integral, and conversion to a real number sends +∞+\infty+∞ to 000. The extended quantity D2(ρ∥σ)D_2(\rho\Vert\sigma)D2​(ρ∥σ) is 000 if either matrix fails to be positive semidefinite with trace 111; otherwise it is +∞+\infty+∞ when ker⁡σ⊈ker⁡ρ\ker\sigma\not\subseteq\ker\rhokerσ⊆kerρ, and when ker⁡σ⊆ker⁡ρ\ker\sigma\subseteq\ker\rhokerσ⊆kerρ it equals

max⁡ ⁣{Re⁡tr⁡ ⁣(f(ρ)−ρ g(σ)),0}log⁡2,f(x)=xlog⁡x,g(x)=log⁡x,\frac{\max\!\left\{\operatorname{Re}\operatorname{tr}\!\left(f(\rho)-\rho\,g(\sigma)\right),0\right\}}{\log 2}, \qquad f(x)=x\log x,\quad g(x)=\log x,log2max{Retr(f(ρ)−ρg(σ)),0}​,f(x)=xlogx,g(x)=logx,

with these matrix functions defined by continuous functional calculus and the real logarithm taking value 000 at 000. The logarithm in the rate is likewise the total real base-two logarithm, whose value at 000 is 000.

Human review
  • Endorsed by Shuze Chen · Oct 22, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 22, 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