Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Remark 12: smoothed Renyi-gap communication bound

Open
QAlgorithms.Papers.Independent.Intakenext100.P61.smooth_renyi_communication_lower

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

prove2quantum

Let 0≤α<1<β≤∞0\le\alpha<1<\beta\le\infty0≤α<1<β≤∞ and let normalized bipartite pure states ϕ,ψ\phi,\psiϕ,ψ be related by a deterministic quantum protocol of cost CCC and unconditional squared fidelity at least 1−ε1-\varepsilon1−ε, with 0≤ε<1/40\le\varepsilon<1/40≤ε<1/4. For eigenvalues rjr_jrj​, define Δeαβ\Delta_e^{\alpha\beta}Δeαβ​ by jointly minimizing log⁡2(∑j∈Jrjα)/(1−α)−log⁡2(∑j∈Jrjβ)/(1−β)\log_2(\sum_{j\in J}r_j^\alpha)/(1-\alpha)-\log_2(\sum_{j\in J}r_j^\beta)/(1-\beta)log2​(∑j∈J​rjα​)/(1−α)−log2​(∑j∈J​rjβ​)/(1−β) over retained mass at least 1−e1-e1−e. Order zero counts positive support; infinity uses the largest retained weight. Put δ=(4ε)1/8\delta=(4\varepsilon)^{1/8}δ=(4ε)1/8. Then

2C≥Δδαβ(ψA)−Δ0αβ(ϕA)+(2α1−α+2ββ−1)log⁡2(1−δ).2C\ge\Delta_\delta^{\alpha\beta}(\psi_A)-\Delta_0^{\alpha\beta}(\phi_A)+\left(\frac{2\alpha}{1-\alpha}+\frac{2\beta}{\beta-1}\right)\log_2(1-\sqrt\delta).2C≥Δδαβ​(ψA​)−Δ0αβ​(ϕA​)+(1−α2α​+β−12β​)log2​(1−δ​).

The factor β/(β−1)\beta/(\beta-1)β/(β−1) equals one at infinity. This retains the source's separate smoothing correction.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P61
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P61

open scoped Matrix ComplexConjugate

open QAlgorithms.Papers.Independent.Information100.P63 (logTwo)

open QAlgorithms.Papers.Independent.Dependency50.P35 (supportRank)

open QAlgorithms.Papers.Independent.Bridge50.P40 (entropyInfinity)

open QAlgorithms.Papers.Independent.Information100.P53 (fidelity)

attribute [local instance] Classical.propDecidable

section Operational

variable {A B C D : Type} [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]
    [Fintype C] [DecidableEq C] [Fintype D] [DecidableEq D]

/-- PDF p.5, Remark 12: jointly smoothed Renyi gaps, with the infinity limit
and the printed log(1-sqrt(delta)) correction. -/
theorem smooth_renyi_communication_lower (phi : EuclideanSpace ℂ (A × B))
    (psi : EuclideanSpace ℂ (C × D)) (hphi : ‖phi‖ = 1) (hpsi : ‖psi‖ = 1)
    (alpha : NNReal) (halpha : alpha < 1) (beta : RenyiOrder) (hbeta : 1 < beta)
    (epsilon cost : ℝ) (hzero : 0 ≤ epsilon) (hsmall : epsilon < 1 / 4)
    (hprotocol : QuantumPrepares phi psi epsilon cost) :
    2 * cost ≥ smoothRenyiGap (aliceDensity psi) (alpha : RenyiOrder) beta (smoothingError epsilon) -
      smoothRenyiGap (aliceDensity phi) (alpha : RenyiOrder) beta 0 +
      (2 * (alpha : ℝ) / (1 - (alpha : ℝ)) + 2 * upperRenyiFactor beta) *
        logTwo (1 - Real.sqrt (smoothingError epsilon)) := by sorry

end Operational

end QAlgorithms.Papers.Independent.Intakenext100.P61

end
Source
On the communication cost of entanglement transformations, arXiv:quant-ph/0204092v3, PDF p. 5, Remark 12: smoothed Renyi-gap communication bound
Read-back

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

For finite types A,B,C,DA,B,C,DA,B,C,D with decidable equality, let ϕ∈CA×B\phi\in\mathbb C^{A\times B}ϕ∈CA×B and ψ∈CC×D\psi\in\mathbb C^{C\times D}ψ∈CC×D have Euclidean norm 111, let 0≤α<10\leq\alpha<10≤α<1, let β\betaβ be either a finite nonnegative real number greater than 111 or +∞+\infty+∞, and let ϵ,cost∈R\epsilon,\mathrm{cost}\in\mathbb Rϵ,cost∈R satisfy 0≤ϵ<1/40\leq\epsilon<1/40≤ϵ<1/4. For a bipartite vector vvv, write ρv\rho_vρv​ for its first marginal, with entries (ρv)ij=∑bvi,bvj,b‾(\rho_v)_{ij}=\sum_b v_{i,b}\overline{v_{j,b}}(ρv​)ij​=∑b​vi,b​vj,b​​, and write log⁡2x=log⁡x/log⁡2\log_2 x=\log x/\log 2log2​x=logx/log2, using the total real logarithm, in particular log⁡20=0\log_2 0=0log2​0=0. For a positive semidefinite matrix ρ\rhoρ, let λi\lambda_iλi​ be its eigenvalues, counted with multiplicity, let Pρ={i:λi>0}P_\rho=\{i:\lambda_i>0\}Pρ​={i:λi​>0}, and put Rρ(u)={S⊆Pρ:∑i∈Sλi≥1−u}\mathcal R_\rho(u)=\{S\subseteq P_\rho:\sum_{i\in S}\lambda_i\geq1-u\}Rρ​(u)={S⊆Pρ​:∑i∈S​λi​≥1−u}. For each such subset define

Ha(S)={log⁡2∣S∣,a=0,−∑i∈Sλilog⁡2λi,a=1,log⁡2 ⁣(∑i∈Sλia)1−a,0<a<∞, a≠1,−log⁡2 ⁣(max⁡i∈Sλi),a=∞,Gα,β(ρ,u)=min⁡S∈Rρ(u)(Hα(S)−Hβ(S)).H_a(S)= \begin{cases} \log_2|S|,&a=0,\\ -\sum_{i\in S}\lambda_i\log_2\lambda_i,&a=1,\\ \dfrac{\log_2\!\left(\sum_{i\in S}\lambda_i^a\right)}{1-a},&0<a<\infty,\ a\neq1,\\ -\log_2\!\left(\max_{i\in S}\lambda_i\right),&a=\infty, \end{cases} \qquad G_{\alpha,\beta}(\rho,u)=\min_{S\in\mathcal R_\rho(u)}\bigl(H_\alpha(S)-H_\beta(S)\bigr).Ha​(S)=⎩⎨⎧​log2​∣S∣,−∑i∈S​λi​log2​λi​,1−alog2​(∑i∈S​λia​)​,−log2​(maxi∈S​λi​),​a=0,a=1,0<a<∞, a=1,a=∞,​Gα,β​(ρ,u)=S∈Rρ​(u)min​(Hα​(S)−Hβ​(S)).

Here a maximum over an empty finite set and a minimum over an empty finite set are assigned 000; the retained eigenvalues are not renormalized. Set s=(4ϵ)1/8s=(4\epsilon)^{1/8}s=(4ϵ)1/8, and set uβ=β/(β−1)u_\beta=\beta/(\beta-1)uβ​=β/(β−1) for finite β\betaβ, and u∞=1u_\infty=1u∞​=1. Suppose some matrix σ\sigmaσ can be obtained from ϕϕ†\phi\phi^\daggerϕϕ† by a finite sequence of local maps X↦∑jKjXKj†X\mapsto\sum_jK_jXK_j^\daggerX↦∑j​Kj​XKj†​ with ∑jKj†Kj=I\sum_jK_j^\dagger K_j=I∑j​Kj†​Kj​=I, and transfers of positive-dimensional tensor factors between the two parties, where transferring a factor of dimension mmm costs log⁡2m\log_2mlog2​m, the accumulated cost is exactly cost\mathrm{cost}cost, and

[Re⁡tr⁡σ (ψψ†)σ]2≥1−ϵ.\left[\operatorname{Re}\operatorname{tr} \sqrt{\sqrt{\sigma}\,(\psi\psi^\dagger)\sqrt{\sigma}}\right]^2\geq1-\epsilon.[Retrσ​(ψψ†)σ​​]2≥1−ϵ.

Then

2 cost≥Gα,β(ρψ,s)−Gα,β(ρϕ,0)+(2α1−α+2uβ)log⁡2(1−s).2\,\mathrm{cost}\geq G_{\alpha,\beta}(\rho_\psi,s)-G_{\alpha,\beta}(\rho_\phi,0) +\left(\frac{2\alpha}{1-\alpha}+2u_\beta\right)\log_2(1-\sqrt{s}).2cost≥Gα,β​(ρψ​,s)−Gα,β​(ρϕ​,0)+(1−α2α​+2uβ​)log2​(1−s​).

The case α=0\alpha=0α=0, the case β=∞\beta=\inftyβ=∞, and zero error are included; if either vector’s index set is empty, its norm-one hypothesis cannot hold.

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