Finite-codimension tensor avoidance
OpenQAlgorithms.Papers.Independent.Intakenext100.P05.finiteCodimension_tensor_avoidanceLet be an infinite-dimensional separable complex Hilbert space, let be an integer, and let be a finite-dimensional complex linear subspace of . Then there is a closed complex linear subspace such that
Here the tensor subspace is viewed inside through the map induced by . Since the first factor is finite-dimensional, its algebraic tensor product gives the same vector space as the Hilbert tensor product. The statement supplies the linear avoidance step used in the torsion index construction.
Formalization Note Finite codimension is expressed by finite-dimensionality of the quotient , without a numerical finrank test.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P05
namespace QAlgorithms.Papers.Independent.Intakenext100.P05
open scoped TensorProduct
/-- Lemma 6.2 (PDF p. 10): a finite-dimensional subspace of C^n tensor H
is disjoint from C^n tensor V for some closed finite-codimension subspace V of H.
The standing Hilbert-space assumptions are from PDF pp. 5 and 21.
The tensor subspace is the range of the actual map induced by V's inclusion;
finite codimension means that H/V is finite-dimensional. The source proof's
interchanged V/W names are read using its coordinate-span orthogonal complement.
No positivity condition on n is imposed. -/
theorem finiteCodimension_tensor_avoidance
{H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H]
[CompleteSpace H] [TopologicalSpace.SeparableSpace H]
(hInfinite : ¬ FiniteDimensional ℂ H)
(n : ℕ) (W : Submodule ℂ ((Fin n → ℂ) ⊗[ℂ] H))
[FiniteDimensional ℂ W] :
∃ V : Submodule ℂ H,
IsClosed (V : Set H) ∧
FiniteDimensional ℂ (H ⧸ V) ∧
(TensorProduct.map
(LinearMap.id : (Fin n → ℂ) →ₗ[ℂ] (Fin n → ℂ)) V.subtype).range ⊓ W = ⊥ := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P05
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every complete, separable complex inner product space with its compatible norm, assuming that is not finite-dimensional over , every natural number , and every finite-dimensional complex linear subspace of the algebraic tensor product , there exists a complex linear subspace that is closed in , whose quotient is finite-dimensional over , and such that
where is the space of complex-valued functions on an -element index set and is inclusion. The zero subspace is allowed for , and is included; in that case the ambient tensor product is zero and the intersection condition is automatic.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.