Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Channel twirling gives a depolarizing channel with the same entanglement fidelity

Open
QAlgorithms.Papers.Independent.Intakenext100.P18.channel_twirl_depolarizing

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

prove2quantum

Let d≥1d\ge1d≥1 and let Λ\LambdaΛ be any completely positive trace-preserving map on ddd-dimensional matrices. Put F(Λ)=Tr⁡[P+(id⁡⊗Λ)(P+)]F(\Lambda)=\operatorname{Tr}[P_+(\operatorname{id}\otimes\Lambda)(P_+)]F(Λ)=Tr[P+​(id⊗Λ)(P+​)] for the fixed normalized plus Bell projector. With probability Haar measure on U(d)U(d)U(d), define

T(Λ)(X)=∫U†Λ(UXU†)U dU.\mathcal T(\Lambda)(X)=\int U^\dagger\Lambda(UXU^\dagger)U\,dU.T(Λ)(X)=∫U†Λ(UXU†)UdU.

This is a channel and F(T(Λ))=F(Λ)F(\mathcal T(\Lambda))=F(\Lambda)F(T(Λ))=F(Λ). If d≥2d\ge2d≥2, it equals the depolarizing map Dp(X)=pX+(1−p)Tr⁡(X)I/dD_p(X)=pX+(1-p)\operatorname{Tr}(X)I/dDp​(X)=pX+(1−p)Tr(X)I/d, where

p=d2F(Λ)−1d2−1∈[−1d2−1,1].p=\frac{d^2F(\Lambda)-1}{d^2-1}\in\left[-\frac1{d^2-1},1\right].p=d2−1d2F(Λ)−1​∈[−d2−11​,1].

At d=1d=1d=1 it is the identity channel, represented by p=1p=1p=1. This extends the preliminary convex-mixture parameter range as required by the lemma's every-channel quantifier. A traceless unitary channel can have F=0F=0F=0.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P18
Formal statement
noncomputable section

open QAlgorithms MeasureTheory

open scoped Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator

namespace QAlgorithms.Papers.Independent.Intakenext100.P18

open QAlgorithms.Papers.Independent.Claimed50.P28 (maximallyEntangledDensity)

open QAlgorithms.Papers.Independent.Support50.P11 (choiState teleportWeyl bellKet)

open QAlgorithms.Papers.Independent.Renew46.P40 (initialJoint)

/-- PDF pp.6--7, Lemma 1, (21),(25). Source correction: arbitrary channels
require the full CP depolarizing retention range, including negative values.
Dimension one is split before division by d^2-1. -/
theorem channel_twirl_depolarizing (d : ℕ) (hd : 0 < d)
    (E : Superop (Fin d) (Fin d)) (hE : IsChannel E) :
    IsChannel (QuditFidelity.haarTwirl d hd E) ∧
    QuditFidelity.entanglementFidelity d hd (QuditFidelity.haarTwirl d hd E) =
      QuditFidelity.entanglementFidelity d hd E ∧
    (d = 1 → QuditFidelity.haarTwirl d hd E = retentionChannel d hd 1) ∧
    (2 ≤ d →
      let p := ((d : ℝ)^2 * QuditFidelity.entanglementFidelity d hd E - 1) / ((d : ℝ)^2 - 1);
      -(1 / ((d : ℝ)^2 - 1)) ≤ p ∧ p ≤ 1 ∧
      QuditFidelity.haarTwirl d hd E = retentionChannel d hd p) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P18

end
Source
General teleportation channel, singlet fraction and quasi-distillation, arXiv:quant-ph/9807091v2, PDF p. 6, §III.D, Lemma 1, eq.(21): channel twirling preserves entanglement fidelity
Read-back

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

For every positive integer ddd and every map E\mathcal EE on complex d×dd\times dd×d matrices admitting a finite Kraus representation E(X)=∑ℓKℓXKℓ†\mathcal E(X)=\sum_\ell K_\ell X K_\ell^\daggerE(X)=∑ℓ​Kℓ​XKℓ†​ for all XXX, with ∑ℓKℓ†Kℓ=Id\sum_\ell K_\ell^\dagger K_\ell=I_d∑ℓ​Kℓ†​Kℓ​=Id​, define

H(E)(X)=∫U†E(UXU†)U dU,Rp(X)=pX+1−pdtr⁡(X)Id,\mathcal H(\mathcal E)(X)=\int U^\dagger\mathcal E(UXU^\dagger)U\,dU, \qquad \mathcal R_p(X)=pX+\frac{1-p}{d}\operatorname{tr}(X)I_d,H(E)(X)=∫U†E(UXU†)UdU,Rp​(X)=pX+d1−p​tr(X)Id​,

where dUdUdU is normalized Haar measure on the unitary group. Let Φd\Phi_dΦd​ be the outer product of d−1/2∑j∣j,j⟩d^{-1/2}\sum_j|j,j\rangled−1/2∑j​∣j,j⟩ with itself, and define Fe(E)=Re⁡tr⁡[Φd(id⁡⊗E)(Φd)]F_{\rm e}(\mathcal E)=\operatorname{Re}\operatorname{tr}[\Phi_d(\operatorname{id}\otimes\mathcal E)(\Phi_d)]Fe​(E)=Retr[Φd​(id⊗E)(Φd​)]. Then H(E)\mathcal H(\mathcal E)H(E) also admits a finite normalized Kraus representation, and Fe(H(E))=Fe(E)F_{\rm e}(\mathcal H(\mathcal E))=F_{\rm e}(\mathcal E)Fe​(H(E))=Fe​(E). If d=1d=1d=1, then H(E)=R1\mathcal H(\mathcal E)=\mathcal R_1H(E)=R1​. If d≥2d\ge2d≥2, then, for the real number

p=d2Fe(E)−1d2−1,p=\frac{d^2F_{\rm e}(\mathcal E)-1}{d^2-1},p=d2−1d2Fe​(E)−1​,

one has −1/(d2−1)≤p≤1-1/(d^2-1)\le p\le1−1/(d2−1)≤p≤1 and H(E)=Rp\mathcal H(\mathcal E)=\mathcal R_pH(E)=Rp​; this permitted interval includes negative values of ppp.

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