Linear maps with equal norms differ by an isometry of the codomain
ProvedWildeQIT.Purification.exists_linearIsometry_apply_eq_of_norm_eqlinear-algebrapurificationquantum-informationwilde-qit
Let be a module over , let be a finite-dimensional inner product space over , and let be linear maps with for every . Then there is a linear isometry with for every . (The rule is a well-defined linear isometry , because , and every linear isometry from a subspace of a finite-dimensional inner product space into it extends to the whole space.)
Preamble
import Mathlib
Formal statement
namespace WildeQIT.Purification
open Matrix
/-- **Codomain isometry from equal norms.** Let `f g : E →ₗ[𝕜] F` be linear maps from a
`𝕜`-module `E` into a finite-dimensional inner product space `F` with `‖f v‖ = ‖g v‖` for every
`v`. Then there is a linear isometry `T : F →ₗᵢ[𝕜] F` with `T (g v) = f v` for every `v`.
(The map `g v ↦ f v` is a well-defined isometry `range g → F`, since `ker g = ker f`; it extends
to all of `F` because `F` is finite-dimensional.) -/
theorem exists_linearIsometry_apply_eq_of_norm_eq {𝕜 E F : Type*} [RCLike 𝕜] [AddCommGroup E] [Module 𝕜 E]
[NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [FiniteDimensional 𝕜 F]
(f g : E →ₗ[𝕜] F) (h : ∀ v, ‖f v‖ = ‖g v‖) :
∃ T : F →ₗᵢ[𝕜] F, ∀ v, T (g v) = f v := by sorry
end WildeQIT.Purification
Source
Nielsen and Chuang, Quantum Computation and Quantum Information, Theorem 2.6 (unitary freedom in the ensemble for density matrices) (key step of the proof); Wilde, From Classical to Quantum Shannon Theory, arXiv:1106.1445v8, proof of Theorem 5.1.1 and Corollary 5.1.1, pp. 185-186