Inherited matrix positivity on a subalgebra of bounded operators
DefinitionRandomDimReduction_Channel_Supplement_SubalgebraCPp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-informationrepresentation-theory
Let be any complex Hilbert space and a complex subalgebra of the bounded operators on . A finite operator block matrix is positive when and
A complex-linear map is completely positive for the inherited matrix order when entrywise application of preserves this cone for every finite size . Left multiplication is the actual complex-linear map ; its definition assumes neither positivity of nor commutation with .
This is the full amplification condition needed in Lemma 3.2. It allows infinite-dimensional and does not require to be commutative, norm closed or star closed, or require to belong to .
Definition code
import Mathlib
set_option autoImplicit false
noncomputable section
open scoped InnerProductSpace
namespace RandomDimReduction.Channel.Supplement
variable {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H] [CompleteSpace H]
/-- The inherited positive cone of finite operator block matrices on the Hilbert direct sum H^r.
Self-adjointness and nonnegative quadratic form are imposed in the ambient bounded operators. -/
def PositiveOperatorBlock {r : ℕ} (X : Matrix (Fin r) (Fin r) (H →L[ℂ] H)) : Prop :=
(∀ i j, (X i j).adjoint = X j i) ∧
∀ v : Fin r → H, 0 ≤ (∑ i, ∑ j, inner ℂ (v i) (X i j (v j))).re
/-- Complete positivity on an arbitrary (not necessarily star-closed) complex subalgebra,
with the inherited matrix order, at every finite amplification. -/
def CompletelyPositiveOn (S : NonUnitalSubalgebra ℂ (H →L[ℂ] H))
(f : S →ₗ[ℂ] (H →L[ℂ] H)) : Prop :=
∀ (r : ℕ) (X : Matrix (Fin r) (Fin r) S),
PositiveOperatorBlock (fun i j => (X i j : H →L[ℂ] H)) →
PositiveOperatorBlock (fun i j => f (X i j))
/-- Literal left multiplication; A need not belong to S. -/
def leftMultiplication (S : NonUnitalSubalgebra ℂ (H →L[ℂ] H)) (A : H →L[ℂ] H) :
S →ₗ[ℂ] (H →L[ℂ] H) where
toFun X := A * (X : H →L[ℂ] H)
map_add' X Y := by simp [mul_add]
map_smul' c X := by simp
end RandomDimReduction.Channel.Supplement
Source
Angus Lowe, Xinyu Tan, Random dimension reduction and learning symmetric properties of quantum states, STOC 2027 submission #994 (October 2026), https://github.com/sigact-stoc/stoc2027/releases/download/papers-2/994.pdf, Lemma 3.2, p.9