Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Definitions for Sending quantum entanglement through noisy channels

Definition
QAlgorithms_Papers_Independent_Support50_P41

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

prove2quantum

The objects the formalized statements of Sending quantum entanglement through noisy channels (arXiv:quant-ph/9604023) are stated in. Each was checked against the paper by a blind read-back and a moderator.

  • OperatorSum: PDF pp.5--6, (2)--(3): a specified normalized operator-sum representation.
  • CountableOperatorSum: PDF pp.9--10,34,36: countably indexed operator sums with actual convergence.
  • outcomeEntropy: PDF p.21, (15): Shannon entropy of the computed exchange-matrix diagonal.
  • indexState: PDF p.31: conjugate coefficients in the paired reference basis.
  • omegaTilde: PDF p.31: unnormalized maximally entangled vector, squared norm d.
  • relativeVector: PDF pp.7,31: partial inner product with the reference vector.
  • relativeOperator: PDF pp.32--33: contraction of both reference legs with z.
  • normalizedOperatorVector: PDF pp.34,36: subnormalized vector of D/d, with coefficient A[j,i]/sqrt(d). The p.36 parenthetical tilde is read using the consistent p.34 scaling.
  • dilationChoiVector: PDF pp.35--36: reorder the actual dilation output as (RQ),E.
  • padKraus: PDF p.37: extend a finite Kraus list by zero operators.
  • mixedInteraction: PDF pp.23--24: fixed unitary interaction with an independent mixed environment.
  • jointInformation: PDF pp.24--25: countable mutual information in bits, as a nonnegative relative-entropy sum, so infinite marginal entropies cause no subtraction.
Definition code
import Mathlib
import Definitions.Def_QAlgorithms_Defs_MatrixMapPositivity
import Definitions.Def_QAlgorithms_Defs_FiniteRegisters
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P40
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P45
import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P38
import Definitions.Def_QAlgorithms_Papers_Independent_Gap20_P06
import Definitions.Def_QAlgorithms_Papers_Independent_Information100_P50
import Definitions.Def_QAlgorithms_Papers_Independent_Renew46_P43

/-! Schumacher, arXiv:quant-ph/9604023v1. PDF page numbers; reference-first
coordinates; entropy in bits; squared fidelity. Compilation is pending.
The unquantified recovery claim on p.30 is deferred in HARD.md. -/

noncomputable section

namespace QAlgorithms.Papers.Independent.Support50.P41

open QAlgorithms
open QAlgorithms.Entanglement (IsCompletelyPositive)
open QAlgorithms.Papers.Independent.Dependency50.P40 (IsPurification purificationFidelity)
open QAlgorithms.Papers.Independent.Dependency50.P45 (entropyExchange)
open QAlgorithms.Papers.Independent.Information100.P08 (canonicalPurification)
open QAlgorithms.Papers.Independent.Information100.P34 (choi dephasing)
open QAlgorithms.Papers.Independent.Dependency50.P38 (vectorizeInputFirst)
open QAlgorithms.Papers.Independent.Gap20.P06 (dilationChannel environmentOutput dilatedVector)
open QAlgorithms.Papers.Independent.Information100.P50 (complementaryMap)
open QAlgorithms.Papers.Independent.Renew46.P43 (IsCountablePOVM classicalKLTerm countableEntropy)
open scoped Matrix ComplexConjugate ComplexOrder ENNReal

attribute [local instance] Classical.propDecidable Classical.decEq

-- The accepted generic dilation functions fix classical equality on each factor.
-- Match that instance for Fin before forming their product unitary groups.
local instance (priority := 2000) (n : ℕ) : DecidableEq (Fin n) := Classical.decEq _

/-- PDF pp.5--6, (2)--(3): a specified normalized operator-sum representation. -/
def OperatorSum {d m : ℕ} (Phi : Superop (Fin d) (Fin d))
    (A : Fin m → Matrix (Fin d) (Fin d) ℂ) : Prop :=
  IsKrausFamily A ∧ ∀ X, Phi X = krausApply A X

/-- PDF pp.9--10,34,36: countably indexed operator sums with actual convergence. -/
def CountableOperatorSum {K : Type*} {d : ℕ} (Phi : Superop (Fin d) (Fin d))
    (A : K → Matrix (Fin d) (Fin d) ℂ) : Prop :=
  HasSum (fun k => (A k)ᴴ * A k) 1 ∧
    ∀ X, HasSum (fun k => A k * X * (A k)ᴴ) (Phi X)

/-- PDF p.21, (15): Shannon entropy of the computed exchange-matrix diagonal. -/
def outcomeEntropy {d m : ℕ} (A : Fin m → Matrix (Fin d) (Fin d) ℂ)
    (rho : Matrix (Fin d) (Fin d) ℂ) : ℝ :=
  -∑ k, (complementaryMap A rho k k).re * Real.logb 2 (complementaryMap A rho k k).re

/-- PDF p.31: conjugate coefficients in the paired reference basis. -/
def indexState {d : ℕ} (v : EuclideanSpace ℂ (Fin d)) : EuclideanSpace ℂ (Fin d) :=
  WithLp.toLp 2 fun i => conj (v i)

/-- PDF p.31: unnormalized maximally entangled vector, squared norm d. -/
def omegaTilde (d : ℕ) : EuclideanSpace ℂ (Fin d × Fin d) :=
  WithLp.toLp 2 fun p => if p.1 = p.2 then 1 else 0

/-- PDF pp.7,31: partial inner product with the reference vector. -/
def relativeVector {a d : ℕ} (z : EuclideanSpace ℂ (Fin a))
    (v : EuclideanSpace ℂ (Fin a × Fin d)) : EuclideanSpace ℂ (Fin d) :=
  WithLp.toLp 2 fun j => ∑ i, conj (z i) * v (i, j)

/-- PDF pp.32--33: contraction of both reference legs with z. -/
def relativeOperator {a d : ℕ} (z : EuclideanSpace ℂ (Fin a))
    (D : Matrix (Fin a × Fin d) (Fin a × Fin d) ℂ) : Matrix (Fin d) (Fin d) ℂ :=
  fun j k => ∑ i, ∑ l, conj (z i) * D (i, j) (l, k) * z l

/-- PDF pp.34,36: subnormalized vector of D/d, with coefficient A[j,i]/sqrt(d).
The p.36 parenthetical tilde is read using the consistent p.34 scaling. -/
def normalizedOperatorVector {d : ℕ} (A : Matrix (Fin d) (Fin d) ℂ) :
    EuclideanSpace ℂ (Fin d × Fin d) :=
  ((Real.sqrt d)⁻¹ : ℂ) • vectorizeInputFirst A

/-- PDF pp.35--36: reorder the actual dilation output as (RQ),E. -/
def dilationChoiVector {d e : ℕ} [NeZero d]
    (U : Matrix.unitaryGroup (Fin d × Fin e) ℂ) (eta : EuclideanSpace ℂ (Fin e)) :
    EuclideanSpace ℂ ((Fin d × Fin d) × Fin e) :=
  WithLp.toLp 2 fun p => dilatedVector U eta (finiteMaxEntangled (Fin d))
    (p.1.1, (p.1.2, p.2))

/-- PDF p.37: extend a finite Kraus list by zero operators. -/
def padKraus {d m : ℕ} (A : Fin m → Matrix (Fin d) (Fin d) ℂ) (n : ℕ) :
    Fin n → Matrix (Fin d) (Fin d) ℂ :=
  fun k => if h : k.val < m then A ⟨k.val, h⟩ else 0

/-- PDF pp.23--24: fixed unitary interaction with an independent mixed environment. -/
def mixedInteraction {d e : ℕ} (U : Matrix.unitaryGroup (Fin d × Fin e) ℂ)
    (sigma : Matrix (Fin e) (Fin e) ℂ) (X : Matrix (Fin d) (Fin d) ℂ) :
    Matrix (Fin d × Fin e) (Fin d × Fin e) ℂ :=
  conjSuperop (U : Matrix (Fin d × Fin e) (Fin d × Fin e) ℂ)
    (Matrix.kronecker X sigma)

/-- PDF pp.24--25: countable mutual information in bits, as a nonnegative
relative-entropy sum, so infinite marginal entropies cause no subtraction. -/
def jointInformation {X Y : Type*} (P : X × Y → ℝ) : ℝ≥0∞ :=
  ∑' z : X × Y, classicalKLTerm (P z)
    ((∑' y, P (z.1, y)) * (∑' x, P (x, z.2)))

/-- The one-operator identity representation satisfies both normalization and action. -/
example {d : ℕ} : OperatorSum (fun X : Matrix (Fin d) (Fin d) ℂ => X)
    (fun _ : Fin 1 => (1 : Matrix (Fin d) (Fin d) ℂ)) := by
  constructor
  · simp [IsKrausFamily]
  · intro X
    simp [krausApply]

/-- The countable interface agrees with the one-label identity representation. -/
example {d : ℕ} : CountableOperatorSum (fun X : Matrix (Fin d) (Fin d) ℂ => X)
    (fun _ : Fin 1 => (1 : Matrix (Fin d) (Fin d) ℂ)) := by
  constructor
  · simpa using (hasSum_fintype (fun _ : Fin 1 =>
      (1 : Matrix (Fin d) (Fin d) ℂ)ᴴ * 1))
  · intro X
    simpa using (hasSum_fintype (fun _ : Fin 1 =>
      (1 : Matrix (Fin d) (Fin d) ℂ) * X * (1 : Matrix (Fin d) (Fin d) ℂ)ᴴ))

example {d m : ℕ} (Phi : Superop (Fin d) (Fin d))
    (A : Fin m → Matrix (Fin d) (Fin d) ℂ) (h : OperatorSum Phi A) :
    IsKrausFamily A := h.1

example {K : Type*} {d : ℕ} (Phi : Superop (Fin d) (Fin d))
    (A : K → Matrix (Fin d) (Fin d) ℂ) (h : CountableOperatorSum Phi A) :
    HasSum (fun k => (A k)ᴴ * A k) 1 := h.1

example {d : ℕ} (rho : Matrix (Fin d) (Fin d) ℂ) :
    outcomeEntropy (fun k : Fin 0 => Fin.elim0 k) rho = 0 := by simp [outcomeEntropy]

example {d : ℕ} (v : EuclideanSpace ℂ (Fin d)) (i : Fin d) :
    indexState (indexState v) i = v i := by simp [indexState]

example : omegaTilde 1 (0, 0) = 1 := by simp [omegaTilde]

example {a d : ℕ} (v : EuclideanSpace ℂ (Fin a × Fin d)) :
    relativeVector 0 v = 0 := by ext j; simp [relativeVector]

example {a d : ℕ} (D : Matrix (Fin a × Fin d) (Fin a × Fin d) ℂ) :
    relativeOperator 0 D = 0 := by ext j k; simp [relativeOperator]

example (A : Matrix (Fin 1) (Fin 1) ℂ) : normalizedOperatorVector A (0, 0) = A 0 0 := by
  simp [normalizedOperatorVector, vectorizeInputFirst, Matrix.vec]

example {d e : ℕ} [NeZero d] (U : Matrix.unitaryGroup (Fin d × Fin e) ℂ)
    (eta : EuclideanSpace ℂ (Fin e)) (i j : Fin d) (k : Fin e) :
    dilationChoiVector U eta ((i, j), k) =
      dilatedVector U eta (finiteMaxEntangled (Fin d)) (i, (j, k)) := rfl

example {d m : ℕ} (A : Fin m → Matrix (Fin d) (Fin d) ℂ) (n : ℕ)
    (k : Fin n) (h : m ≤ k.val) : padKraus A n k = 0 := by
  simp [padKraus, Nat.not_lt.mpr h]

example {d e : ℕ} (U : Matrix.unitaryGroup (Fin d × Fin e) ℂ)
    (sigma : Matrix (Fin e) (Fin e) ℂ) (X : Matrix (Fin d) (Fin d) ℂ) :
    mixedInteraction U sigma X = conjSuperop (U : Matrix (Fin d × Fin e) (Fin d × Fin e) ℂ)
      (Matrix.kronecker X sigma) := rfl

example {X Y : Type*} : jointInformation (fun _ : X × Y => 0) = 0 := by
  simp [jointInformation, classicalKLTerm]

end QAlgorithms.Papers.Independent.Support50.P41
Source
Sending quantum entanglement through noisy channels, arXiv:quant-ph/9604023v1, https://arxiv.org/abs/quant-ph/9604023v1
Read-back

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

QAlgorithms.Papers.Independent.Support50.P41.CountableOperatorSum

For any index type KKK, any natural number ddd, any function Φ:Cd×d→Cd×d\Phi:\mathbb C^{d\times d}\to\mathbb C^{d\times d}Φ:Cd×d→Cd×d, and any family A:K→Cd×dA:K\to\mathbb C^{d\times d}A:K→Cd×d, this predicate means that the unordered sums converge and satisfy

∑k∈KAk†Ak=Idand∀X∈Cd×d,∑k∈KAkXAk†=Φ(X).\sum_{k\in K}A_k^\dagger A_k=I_d \quad\text{and}\quad \forall X\in\mathbb C^{d\times d},\qquad \sum_{k\in K}A_kXA_k^\dagger=\Phi(X).k∈K∑​Ak†​Ak​=Id​and∀X∈Cd×d,k∈K∑​Ak​XAk†​=Φ(X).

Here convergence means convergence of the net of sums over finite subsets of KKK. No countability assumption on KKK is imposed, and XXX ranges over all matrices. The case d=0d=0d=0 is included; if KKK is empty, the first equality is impossible for d>0d>0d>0.

QAlgorithms.Papers.Independent.Support50.P41.OperatorSum

For any natural numbers d,md,md,m, any function Φ\PhiΦ from complex d×dd\times dd×d matrices to complex d×dd\times dd×d matrices, and any family A0,…,Am−1A_0,\ldots,A_{m-1}A0​,…,Am−1​ of such matrices, this predicate means

∑k=0m−1Ak†Ak=Idand∀X∈Cd×d,Φ(X)=∑k=0m−1AkXAk†,\sum_{k=0}^{m-1}A_k^\dagger A_k=I_d \quad\text{and}\quad \forall X\in\mathbb C^{d\times d},\qquad \Phi(X)=\sum_{k=0}^{m-1}A_kXA_k^\dagger,k=0∑m−1​Ak†​Ak​=Id​and∀X∈Cd×d,Φ(X)=k=0∑m−1​Ak​XAk†​,

where †\dagger† denotes conjugate transpose. The equality is required for every matrix XXX, without positivity or normalization restrictions. Both dimensions may be zero; if m=0m=0m=0, the sums are zero, so the predicate is impossible when d>0d>0d>0.

QAlgorithms.Papers.Independent.Support50.P41.dilationChoiVector

For any natural numbers d,ed,ed,e with d≠0d\ne0d=0, any unitary matrix UUU indexed by {0,…,d−1}×{0,…,e−1}\{0,\ldots,d-1\}\times\{0,\ldots,e-1\}{0,…,d−1}×{0,…,e−1}, and any vector η∈Ce\eta\in\mathbb C^eη∈Ce, this defines a vector www indexed by ((i,j),t)((i,j),t)((i,j),t) through

w((i,j),t)=1d∑b=0e−1U(j,t),(i,b) ηb.w_{((i,j),t)} =\frac{1}{\sqrt d}\sum_{b=0}^{e-1}U_{(j,t),(i,b)}\,\eta_b.w((i,j),t)​=d​1​b=0∑e−1​U(j,t),(i,b)​ηb​.

Equivalently, it applies UUU to the second ddd-dimensional factor and the environment of the vector with amplitudes δijηt/d\delta_{ij}\eta_t/\sqrt dδij​ηt​/d​, then regroups the indices as ((i,j),t)((i,j),t)((i,j),t). No normalization of η\etaη is assumed. The environment dimension eee may be zero, in which case the result is the unique zero dimensional vector.

QAlgorithms.Papers.Independent.Support50.P41.indexState

For every natural number ddd and every vector v∈Cdv\in\mathbb C^dv∈Cd, this defines the vector whose iii-th coordinate is vi‾\overline{v_i}vi​​ for each 0≤i<d0\le i<d0≤i<d. No unit norm assumption is imposed. When d=0d=0d=0, the result is the unique vector in the zero dimensional space.

QAlgorithms.Papers.Independent.Support50.P41.jointInformation

For arbitrary types X,YX,YX,Y and any real valued function P:X×Y→RP:X\times Y\to\mathbb RP:X×Y→R, put pxy=P(x,y)p_{xy}=P(x,y)pxy​=P(x,y) and qxy=(∑y′P(x,y′))(∑x′P(x′,y))q_{xy}=(\sum_{y'}P(x,y'))(\sum_{x'}P(x',y))qxy​=(∑y′​P(x,y′))(∑x′​P(x′,y)). This defines the extended nonnegative real value

∑(x,y)∈X×Y{max⁡ ⁣(qxy/ln⁡2,0),pxy=0,+∞,pxy≠0 and qxy=0,max⁡ ⁣(pxylog⁡2(pxy/qxy)+(qxy−pxy)/ln⁡2,0),pxy≠0 and qxy≠0.\sum_{(x,y)\in X\times Y} \begin{cases} \max\!\left(q_{xy}/\ln2,0\right),&p_{xy}=0,\\ +\infty,&p_{xy}\ne0\ \text{and}\ q_{xy}=0,\\ \max\!\left(p_{xy}\log_2(p_{xy}/q_{xy})+(q_{xy}-p_{xy})/\ln2,0\right), & p_{xy}\ne0\ \text{and}\ q_{xy}\ne0. \end{cases}(x,y)∈X×Y∑​⎩⎨⎧​max(qxy​/ln2,0),+∞,max(pxy​log2​(pxy​/qxy​)+(qxy​−pxy​)/ln2,0),​pxy​=0,pxy​=0 and qxy​=0,pxy​=0 and qxy​=0.​

There are no finiteness, countability, nonnegativity, normalization, or summability assumptions on PPP. Each marginal sum is a total real unordered sum, assigned zero when its family is not summable; the outer sum is an extended nonnegative unordered sum and may be infinite. Real division and logarithms are total, with division by zero and log⁡20\log_2 0log2​0 equal to zero. If either type is empty, the outer sum is zero.

QAlgorithms.Papers.Independent.Support50.P41.mixedInteraction

For any natural numbers d,ed,ed,e, any unitary matrix UUU indexed by {0,…,d−1}×{0,…,e−1}\{0,\ldots,d-1\}\times\{0,\ldots,e-1\}{0,…,d−1}×{0,…,e−1}, and arbitrary matrices σ∈Ce×e\sigma\in\mathbb C^{e\times e}σ∈Ce×e and X∈Cd×dX\in\mathbb C^{d\times d}X∈Cd×d, this defines

U(X⊗σ)U†,(X⊗σ)(i,a),(j,b)=Xijσab.U(X\otimes\sigma)U^\dagger, \qquad (X\otimes\sigma)_{(i,a),(j,b)}=X_{ij}\sigma_{ab}.U(X⊗σ)U†,(X⊗σ)(i,a),(j,b)​=Xij​σab​.

Neither σ\sigmaσ nor XXX is required to be positive, Hermitian, or normalized. Both dimensions may be zero, yielding a matrix on an empty index set if either is zero.

QAlgorithms.Papers.Independent.Support50.P41.normalizedOperatorVector

For any natural number ddd and any matrix A∈Cd×dA\in\mathbb C^{d\times d}A∈Cd×d, this defines a vector www indexed by (i,j)∈{0,…,d−1}2(i,j)\in\{0,\ldots,d-1\}^2(i,j)∈{0,…,d−1}2 with coordinates

w(i,j)=(d)−1Aji.w_{(i,j)}=(\sqrt d)^{-1}A_{ji}.w(i,j)​=(d​)−1Aji​.

Thus the first vector index is the input, or column, index of AAA. No condition on AAA ensures that the resulting vector has norm one. The case d=0d=0d=0 is included: the reciprocal of zero is defined as zero, and the resulting vector has no coordinates.

QAlgorithms.Papers.Independent.Support50.P41.omegaTilde

For every natural number ddd, this defines a vector ω~d∈C{0,…,d−1}2\widetilde\omega_d\in\mathbb C^{\{0,\ldots,d-1\}^2}ωd​∈C{0,…,d−1}2 by

(ω~d)ij={1,i=j,0,i≠j.(\widetilde\omega_d)_{ij}= \begin{cases} 1,&i=j,\\ 0,&i\ne j. \end{cases}(ωd​)ij​={1,0,​i=j,i=j.​

There is no normalization factor; its norm is d\sqrt dd​. For d=0d=0d=0, it is the unique zero dimensional vector.

QAlgorithms.Papers.Independent.Support50.P41.outcomeEntropy

For any natural numbers d,md,md,m, any family A0,…,Am−1∈Cd×dA_0,\ldots,A_{m-1}\in\mathbb C^{d\times d}A0​,…,Am−1​∈Cd×d, and any matrix ρ∈Cd×d\rho\in\mathbb C^{d\times d}ρ∈Cd×d, define the real numbers

qk=Re⁡ ⁣(∑y=0d−1∑a=0d−1∑b=0d−1(Ak)ya (Ak)yb‾ ρab).q_k=\operatorname{Re}\!\left( \sum_{y=0}^{d-1}\sum_{a=0}^{d-1}\sum_{b=0}^{d-1} (A_k)_{ya}\,\overline{(A_k)_{yb}}\,\rho_{ab} \right).qk​=Re(y=0∑d−1​a=0∑d−1​b=0∑d−1​(Ak​)ya​(Ak​)yb​​ρab​).

The defined value is −∑k=0m−1qklog⁡2qk-\sum_{k=0}^{m-1}q_k\log_2 q_k−∑k=0m−1​qk​log2​qk​. No Kraus completeness condition, density condition, nonnegativity, or normalization of the qkq_kqk​ is assumed. The logarithm is the total real logarithm, with log⁡20=0\log_2 0=0log2​0=0, so zero terms contribute zero. If m=0m=0m=0 or d=0d=0d=0, the value is zero.

QAlgorithms.Papers.Independent.Support50.P41.padKraus

For any natural numbers d,md,md,m, any family A0,…,Am−1∈Cd×dA_0,\ldots,A_{m-1}\in\mathbb C^{d\times d}A0​,…,Am−1​∈Cd×d, and any natural number nnn, this defines a family B0,…,Bn−1B_0,\ldots,B_{n-1}B0​,…,Bn−1​ by

Bk={Ak,k<m,0,m≤k.B_k= \begin{cases} A_k,&k<m,\\ 0,&m\le k. \end{cases}Bk​={Ak​,0,​k<m,m≤k.​

There is no assumption that m≤nm\le nm≤n: when n<mn<mn<m, the family is truncated; when n>mn>mn>m, zero matrices are appended. No Kraus completeness condition is imposed. All zero dimensions are allowed, and n=0n=0n=0 gives the empty family.

QAlgorithms.Papers.Independent.Support50.P41.relativeOperator

For any natural numbers a,da,da,d, any vector z∈Caz\in\mathbb C^az∈Ca, and any complex matrix DDD indexed by pairs (i,j)(i,j)(i,j) with 0≤i<a0\le i<a0≤i<a and 0≤j<d0\le j<d0≤j<d, this defines a d×dd\times dd×d matrix RRR by

Rjk=∑i=0a−1∑ℓ=0a−1zi‾ D(i,j),(ℓ,k) zℓ.R_{jk}=\sum_{i=0}^{a-1}\sum_{\ell=0}^{a-1} \overline{z_i}\,D_{(i,j),(\ell,k)}\,z_\ell.Rjk​=i=0∑a−1​ℓ=0∑a−1​zi​​D(i,j),(ℓ,k)​zℓ​.

No normalization of zzz, Hermiticity, positivity, or trace condition on DDD is assumed. If a=0a=0a=0, the result is zero; d=0d=0d=0 is also allowed.

QAlgorithms.Papers.Independent.Support50.P41.relativeVector

For any natural numbers a,da,da,d, any vector z∈Caz\in\mathbb C^az∈Ca, and any vector v∈C{0,…,a−1}×{0,…,d−1}v\in\mathbb C^{\{0,\ldots,a-1\}\times\{0,\ldots,d-1\}}v∈C{0,…,a−1}×{0,…,d−1}, this defines w∈Cdw\in\mathbb C^dw∈Cd by

wj=∑i=0a−1zi‾ v(i,j).w_j=\sum_{i=0}^{a-1}\overline{z_i}\,v_{(i,j)}.wj​=i=0∑a−1​zi​​v(i,j)​.

Neither vector is required to be normalized. If a=0a=0a=0, every output coordinate is zero; if d=0d=0d=0, the output has no coordinates.

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