Local conjugation forces a tensor product of unitaries
OpenQAlgorithms.Papers.Independent.Intakenext100.P81.local_unitary_factorizationLet and let be nonzero finite-dimensional complex Hilbert spaces. Let be unitary on . Suppose that is supported on factor whenever is supported on that factor. Then there are factor unitaries such that
For every subset of the sites, there is a unitary whose conjugation agrees with that of on every operator supported in and fixes every operator supported in its complement. This gives the lattice splittability consequence. The factors have arbitrary positive dimensions, and equality includes the scalar phase.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P81
noncomputable section
namespace QAlgorithms.Papers.Independent.Intakenext100.P81
open scoped BigOperators ComplexConjugate Manifold ContDiff
/-- Theorem 2.1, PDF p.23, proof pp.146--148, lattice consequence p.24.
The finite family is nonempty and its factors are nonzero. Scalar phase is absorbed
in a factor; no qubit restriction and no equality merely up to phase is imposed. -/
theorem local_unitary_factorization {n : ℕ} (hn : 0 < n) (dim : Fin n → ℕ)
(hdim : ∀ i, 0 < dim i) (U : Matrix.unitaryGroup (FactorBasisIndex dim) ℂ)
(hlocal : ∀ i (O : Matrix (Fin (dim i)) (Fin (dim i)) ℂ),
∃ P : Matrix (Fin (dim i)) (Fin (dim i)) ℂ,
(star (U : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ)) *
embedSite dim i O * (U : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) =
embedSite dim i P) :
(∃ V : (i : Fin n) → Matrix.unitaryGroup (Fin (dim i)) ℂ,
(U : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) =
factorTensorMatrix dim (fun i => (V i : Matrix (Fin (dim i)) (Fin (dim i)) ℂ))) ∧
(∀ s : Finset (Fin n), ∃ V : Matrix.unitaryGroup (FactorBasisIndex dim) ℂ,
(∀ O : Matrix (RegionIndex dim s) (RegionIndex dim s) ℂ,
star (V : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) * embedRegion dim s O *
(V : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) =
star (U : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) * embedRegion dim s O *
(U : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ)) ∧
(∀ O : Matrix (RegionIndex dim sᶜ) (RegionIndex dim sᶜ) ℂ,
star (V : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) * embedRegion dim sᶜ O *
(V : Matrix (FactorBasisIndex dim) (FactorBasisIndex dim) ℂ) = embedRegion dim sᶜ O)) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P81
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every natural number , every family of positive natural dimensions for , and every unitary complex matrix indexed by the assignment set , assume the following: for each and each complex matrix , there exists a complex matrix such that , where when for every , and otherwise; here denotes conjugate transpose. Then both of these conclusions hold: there exists a family of unitary complex matrices such that
and, for every subset , there exists a unitary complex matrix indexed by such that
Here is the complement within , consists of assignments for , and equals if for every , and equals otherwise. Dimensions equal to are included, as are empty and full subsets ; is a singleton, and is the sole entry of times the identity.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.