Lemma 3: valence-bond projection gives the cluster
ProvedQAlgorithms.Papers.Independent.Intakenext100.P52.valence_projection_gives_clusterLet be an rectangular grid in which every vertex is incident to an edge; nontrivial one-dimensional rows are included. Let , let be the number of grid edges, and prepare one normalized -bond per edge, with one virtual qubit at each endpoint. At every vertex apply , and let be their tensor product. Then
Thus the projected vector is nonzero and division by gives precisely the normalized cluster state. The bond-only construction excludes an isolated physical site, for which the source supplies no degree-zero local-resource convention. The empty grid uses the scalar vacuum convention.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P52
noncomputable section
open scoped Matrix ComplexConjugate
namespace QAlgorithms.Papers.Independent.Intakenext100.P52
open QAlgorithms.Papers.Independent.Information100.P21 (Vec)
open QAlgorithms.Papers.Independent.Gap20.P01 (applyRect)
/-- Lemma 3, PDF p.17: rectangular grids, including nontrivial 1D rows.
Every site is incident to a bond, since the source's bond-only resource does
not specify a degree-zero local state. Nonzero norm is a conclusion.
The exact coefficient is (sqrt(2))^V / 2^E, so normalization is well-defined. -/
theorem valence_projection_gives_cluster (r c : ℕ)
(hincident : ∀ v : RectangularSites r c, ∃ e ∈ gridEdges r c, e.1 = v ∨ e.2 = v) :
let E := gridEdges r c
let v := applyRect (valenceProjection E) (bondGrid E)
let factor : ℝ := (Real.sqrt 2) ^ (r*c) / (2 : ℝ) ^ E.card
v = (factor : ℂ) • graphClusterState E ∧ 0 < factor ∧ ‖v‖ = factor ∧
v ≠ 0 ∧ ((factor : ℂ)⁻¹) • v = graphClusterState E := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P52
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For all natural numbers , let and let consist of the ordered pairs of sites satisfying either and , or and . Assume every site is an endpoint of some pair in . Define , and let be when and for every , and otherwise. Set
Then , , , , and , with real scalars regarded as complex scalars in vector equalities. The quantification includes or : there are then no sites or edges, the incidence assumption is vacuous, , and both vectors have their sole component equal to ; for , the incidence assumption cannot hold.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.