Theorem 3.1 — full Tanner-to-quantum transformation
OpenQuantumCupGates644.ConstantWeight.Supplement.theorem_3_1Let Γ be a finite directed graph of total incident degree at most Δ≥2, let F be a field, and let r≥2. Assume Γ is oriented from one bipartition class to the other, or F has characteristic two. Let be the chosen local systems. Let M⊆E(Γ) be extendable for the first r−1 Tanner codes, and assume their local Schur product lies in the rth local image. The codes Q^(i) and protected spaces L^(i) are the actual cyclic tensor construction and Equation (12).
The selected subsystems support a transversal C^(r−1)Z gate with
This algebraic conclusion holds over the full field generality of §3.1. At finite fields it also requires the actual trace-phase action on CSS coset states. If all classical Tanner codes have distance bounds d^(j) and radius-ρ^(j) decoders, the same constructed subsystems are valid and satisfy
The quantum clock is c_r(T+Δ+|V(Γ)|+1)^{k_r r}, with witnesses preceding the field, graph and systems. This retains the unconditional gate assertion and every conditional parameter and decoder clause. Locality is the source's separate row and column convention.
Differences from the source:
- encoding: one-based finite coordinate sets and quotient logical spaces → Fin labels and full-preimage logical subspaces; reason: the relabellings are bijective and the published CSS convention represents the same quotients
- encoding: radius-min{∏ρ^(j)/Δ^(r−1),ρ^(r)} time-poly(Δ,|V|)^r decoding → nonnegative radii, strict error inequalities, and c(T+Δ+|V|+1)^(kr) elementary-operation bounds; reason: c and k depend only on fixed r and precede every graph, field, system and constituent clock T; polynomial constituent clocks specialize to the printed bound
- generalization: finite directed graphs → finite edge-labelled loopless directed graphs, also allowing parallel edges; reason: the proof uses incidence and total incident degree only; simple source graphs are included without changing any bound
- generalization: integer minimum-distance parameters → nonnegative real lower bounds for constituent-code distances; reason: substituting the source minimum distances recovers both source bounds; the displayed nonnegative distance expressions are monotone in these lower bounds
import Definitions.Def_QuantumCupGates644_ConstantWeight_Supplement_cup import Definitions.Def_QuantumCupGates644_ConstantWeight_Supplement_decoder
namespace QuantumCupGates644.ConstantWeight.Supplement
open scoped BigOperators
open Classical
theorem theorem_3_1 :
∀ r : ℕ, 2 ≤ r → ∃ c k : ℕ, 0 < c ∧ 0 < k ∧
∀ (F : Type) [Field F] (G : Graph) (Δ : ℕ), 2 ≤ Δ →
∀ (S : Fin r → LocalSystem F G), G.DegreeAtMost Δ →
(G.LeftToRight ∨ (2 : F) = 0) →
∀ (M : Finset (Fin G.ne)), Extendable S M → Multiplication S →
AlgebraicGate (productSubsystem S) ((4*Δ)^(r^2)) (M.card^r) ∧
(∀ [Fintype F] (p : ℕ) [Fact p.Prime] [Algebra (ZMod p) F],
SupportsGate p (productSubsystem S) ((4*Δ)^(r^2)) (M.card^r)) ∧
∀ (d ρ : Fin r → NNReal) (T : ℕ),
(∀ j, DistanceAtLeast (S j).tanner (d j)) →
(∀ j, DecodesClassical (S j).tanner (ρ j) T) →
∀ i, (productSubsystem S i).Valid ∧
SourceLocality S i (r * (Δ : ℝ)^2) ∧
(productSubsystem S i).XDistanceAtLeast (xDistanceBound d) ∧
(productSubsystem S i).ZDistanceAtLeast (zDistanceBound d Δ) ∧
DecodesQuantum (productSubsystem S i) (radius (fun j => ρ j) Δ)
(quantumClock c k r Δ G.nv T) := by sorry
end QuantumCupGates644.ConstantWeight.Supplement