Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Universal probabilistic-PBT success bound

Open
QAlgorithms.Papers.Independent.Intakenext100.P53.probabilistic_success_bound

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

prove2quantum

Let N≥1N\ge1N≥1 and d≥2d\ge2d≥2 be integers. Alice and Bob have NNN ordered ddd-dimensional ports, and the input register CCC has dimension ddd. Write s=(d−1)/2s=(d-1)/2s=(d−1)/2 and ∣S⟩=d−1/2∑m=−ss(−1)s−m∣s,m⟩∣s,−m⟩|S\rangle=d^{-1/2}\sum_{m=-s}^s(-1)^{s-m}|s,m\rangle|s,-m\rangle∣S⟩=d−1/2∑m=−ss​(−1)s−m∣s,m⟩∣s,−m⟩. Put Pi=∣S⟩⟨S∣AiC⊗IAˉiP_i=|S\rangle\langle S|_{A_iC}\otimes I_{\bar A_i}Pi​=∣S⟩⟨S∣Ai​C​⊗IAˉi​​, σi=d−(N−1)Pi\sigma_i=d^{-(N-1)}P_iσi​=d−(N−1)Pi​, and ρ=∑iσi\rho=\sum_i\sigma_iρ=∑i​σi​. A resource filter OOO gives ∣ΨO⟩=(OA⊗IB)∣S⟩⊗N|\Psi_O\rangle=(O_A\otimes I_B)|S\rangle^{\otimes N}∣ΨO​⟩=(OA​⊗IB​)∣S⟩⊗N and obeys tr⁡(O†O)=dN\operatorname{tr}(O^\dagger O)=d^Ntr(O†O)=dN. Alice measures ACACAC; Bob selects the reported port and discards the others. Entanglement fidelities are tested against a normalized reference-input singlet, with no correction on the selected port.

Let {Π0,Π1,…,ΠN}\{\Pi_0,\Pi_1,\ldots,\Pi_N\}{Π0​,Π1​,…,ΠN​} be an arbitrary POVM, with 000 indicating failure, which teleports the input exactly in every successful branch. Let pip_ipi​ be the probability of branch iii in the reference-singlet experiment, and p=∑i=1Npip=\sum_{i=1}^Np_ip=∑i=1N​pi​. Then

p≤NN+d2−1.p\le\frac{N}{N+d^2-1}.p≤N+d2−1N​.

In the associated coding experiment Bob sends BkB_kBk​ to Alice and she uses this same measurement. Equality in the bound requires q0∣k=0q_{0|k}=0q0∣k​=0 for every kkk, where q0∣kq_{0|k}q0∣k​ is the actual failure probability in that experiment. There is no symmetry or maximally entangled-resource restriction, and no general attainability claim. This is the paper's alternative derivation of the known numerical bound.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P53
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P53

open scoped Classical ComplexConjugate ComplexOrder Matrix Matrix.Norms.L2Operator

-- Accepted generic matrix and positive-operator interfaces. The support is the
-- orthogonal range projection; all uses below are on positive physical rho.

open QAlgorithms.Papers.Independent.Information100.P21 (Op)

open QAlgorithms.Papers.Independent.Information100.P07 (support)

open QAlgorithms.Papers.Independent.Next46.P39 (inverseSqrt)

/-- PDF p.4, Eqs.(19)-(22): universal bound and necessary coding saturation. -/
theorem probabilistic_success_bound (d N : ℕ) (hd : 2 ≤ d) (hN : 1 ≤ N)
    (O : Op (Ports d N)) (E : Option (Fin N) → Op (AC d N))
    (hexact : ExactProbabilistic O E) :
    pbtSuccessProbability O E ≤ (N : ℝ) / ((N : ℝ) + (d : ℝ) ^ 2 - 1) ∧
    (pbtSuccessProbability O E = (N : ℝ) / ((N : ℝ) + (d : ℝ) ^ 2 - 1) →
      ∀ k, codingProbability O (E none) k = 0) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P53

end
Source
Some remarks on port-based teleportation, arXiv:1506.01555v2, PDF p. 4, V Eqs.(19)-(22): rederived general probabilistic-PBT upper bound
Read-back

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

For every d,N∈Nd,N\in\mathbb Nd,N∈N with d≥2d\ge2d≥2 and N≥1N\ge1N≥1, let P={0,…,d−1}{0,…,N−1}P=\{0,\ldots,d-1\}^{\{0,\ldots,N-1\}}P={0,…,d−1}{0,…,N−1}, H=P×{0,…,d−1}H=P\times\{0,\ldots,d-1\}H=P×{0,…,d−1}, and let OOO be a complex matrix on PPP. Let E∅E_\varnothingE∅​ and EiE_iEi​, i<Ni<Ni<N, be complex matrices on HHH, with ∅\varnothing∅ denoting the “none” outcome. Define

s(a,c)=(−1)ad1{c=d−1−a},Σ=ss†,ΨO(x,y)=∑uOx,u∏js(uj,yj),s(a,c)=\frac{(-1)^a}{\sqrt d}\mathbf1_{\{c=d-1-a\}}, \quad \Sigma=ss^\dagger, \quad \Psi_O(x,y)=\sum_uO_{x,u}\prod_js(u_j,y_j),s(a,c)=d​(−1)a​1{c=d−1−a}​,Σ=ss†,ΨO​(x,y)=u∑​Ox,u​j∏​s(uj​,yj​), (BA)(a,x),(b,y)=∑u,v,c,eA(v,e),(u,c)ΨO(u,x)ΨO(v,y)‾s(a,c)s(b,e)‾.(B_A)_{(a,x),(b,y)} =\sum_{u,v,c,e}A_{(v,e),(u,c)} \Psi_O(u,x)\overline{\Psi_O(v,y)}s(a,c)\overline{s(b,e)}.(BA​)(a,x),(b,y)​=u,v,c,e∑​A(v,e),(u,c)​ΨO​(u,x)ΨO​(v,y)​s(a,c)s(b,e)​.

Let ins⁡k(c,r)\operatorname{ins}_k(c,r)insk​(c,r) insert ccc at port kkk, and define

(QkBA)(a,c),(b,e)=∑r(BA)(a,ins⁡k(c,r)),(b,ins⁡k(e,r)),(Ck)(x,c),(y,e)=∑rΨO(x,ins⁡k(c,r))ΨO(y,ins⁡k(e,r))‾,(Q_kB_A)_{(a,c),(b,e)} =\sum_r(B_A)_{(a,\operatorname{ins}_k(c,r)),(b,\operatorname{ins}_k(e,r))}, \qquad (C_k)_{(x,c),(y,e)} =\sum_r\Psi_O(x,\operatorname{ins}_k(c,r)) \overline{\Psi_O(y,\operatorname{ins}_k(e,r))},(Qk​BA​)(a,c),(b,e)​=r∑​(BA​)(a,insk​(c,r)),(b,insk​(e,r))​,(Ck​)(x,c),(y,e)​=r∑​ΨO​(x,insk​(c,r))ΨO​(y,insk​(e,r))​,

where rrr ranges over assignments to the ports other than kkk. Suppose that tr⁡(O†O)=dN\operatorname{tr}(O^\dagger O)=d^Ntr(O†O)=dN, all E∅,EiE_\varnothing,E_iE∅​,Ei​ are positive semidefinite, E∅+∑iEi=IHE_\varnothing+\sum_iE_i=I_HE∅​+∑i​Ei​=IH​, and, writing bi=Re⁡tr⁡(BEi)b_i=\operatorname{Re}\operatorname{tr}(B_{E_i})bi​=Retr(BEi​​), one has QiBEi=biΣQ_iB_{E_i}=b_i\SigmaQi​BEi​​=bi​Σ for every iii. Then, with p=∑ibip=\sum_i b_ip=∑i​bi​, the theorem asserts

p≤NN+d2−1,p=NN+d2−1 ⟹ ∀k<N, Re⁡tr⁡(E∅Ck)=0.p\le\frac{N}{N+d^2-1}, \qquad p=\frac{N}{N+d^2-1} \ \Longrightarrow\ \forall k<N,\ \operatorname{Re}\operatorname{tr}(E_\varnothing C_k)=0.p≤N+d2−1N​,p=N+d2−1N​ ⟹ ∀k<N, Retr(E∅​Ck​)=0.
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