Lemma 1: , , is an isomorphism with inverse , and (10)
ProvedMosonyiPetz.lemma_1Throughout, is a linear, trace-preserving, 2-positive map between the matrix algebras of finite-dimensional Hilbert spaces, and are invertible density matrices on with , invertible (the standing assumption of the paper, p.3). is the adjoint of for the Hilbert–Schmidt inner product .
Assume that is sufficient for . Let , be the fixed-point sets of and , and
Then
- ;
- ;
- restricted to is an algebra isomorphism onto with inverse : both sets are closed under multiplication, maps onto , for , on and on ;
- for all and ,
The paper does not prove the lemma; it refers to 9.1 of Strătilă, Modular theory in operator algebras (its reference [17]); (10) is Choi's multiplicative-domain theorem for the unital 2-positive map .
Formalization Note is Matrix H H ℂ for a finite index type H; the densities are D 0 and D 1 ; an invertible density is a positive definite matrix of trace one; is the published Hilbert–Schmidt adjoint channelAdjoint. Square roots are the continuous functional calculus of and .
Differences from the source:
- encoding: the two densities D_1, D_2 (s = 1, 2) → D : Fin 2 → Matrix H H ℂ with D 0 = D₁, D 1 = D₂; reason: indexing convention; same instances
- encoding: invertible density matrices D_s and invertible T(D_s) → (D s).PosDef ∧ (D s).trace = 1 and (T (D s)).PosDef; reason: an invertible positive semidefinite matrix is positive definite; same instances
- encoding: T*, the adjoint of T for ⟨A, B⟩ = Tr A*B → channelAdjoint T (published Hilbert–Schmidt adjoint); reason: the same map
- reading: T* restricted to D_K is an algebraic isomorphism onto D_H with inverse α → D_K, D_H closed under products; T* maps D_K onto D_H, is multiplicative on D_K; α ∘ T* = id on D_K and T* ∘ α = id on D_H; reason: spells out 'algebraic isomorphism' between the fixed point algebras (p.8); T* is linear by construction
import Mathlib import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P84 import Definitions.Def_QAlgorithms_Papers_Independent_Dependency50_P21 import Definitions.Def_MosonyiPetz_IsSufficient import Definitions.Def_MosonyiPetz_alpha import Definitions.Def_MosonyiPetz_fixedPointsH import Definitions.Def_MosonyiPetz_fixedPointsK import Definitions.Def_MosonyiPetz_schwarzDomain open scoped Matrix ComplexOrder open QAlgorithms.Papers.Independent.Intakethird100.P84 (IsTwoPositive) open QAlgorithms.Papers.Independent.Dependency50.P21 (channelAdjoint)
namespace MosonyiPetz
/-- **Lemma 1: the fixed-point algebras lie in the multiplicative domains, and T* : D_K → D_H is
an algebra isomorphism with inverse α.**
Mosonyi–Petz, *Structure of sufficient quantum coarse-grainings*, arXiv:quant-ph/0312221v1, §4, p.8, Lemma 1: "Let A_{T*} := {X ∈ B(K) : T*(XX*) = T*(X)T*(X*) and T*(X*X) = T*(X*)T*(X)}
and A_α := {A ∈ B(H) : α(AA*) = α(A)α(A*) and α(A*A) = α(A*)α(A)}. Then DK ⊂ A_{T*} and
DH ⊂ A_α. Moreover, T* restricted to DK is an algebraic isomorphism onto DH with inverse α and
α(AB) = α(A)α(B) (10) for all A ∈ A_α and B ∈ B(H)."
Under the standing assumption and sufficiency (§4), with `D_K`, `D_H` the fixed-point sets of
`γ_K = α ∘ T*` and `γ_H = T* ∘ α`:
1. `D_K ⊆ A_{T*}`;
2. `D_H ⊆ A_α`;
3. `D_K` and `D_H` are closed under multiplication, `T*` maps `D_K` onto `D_H`, `T*` is
multiplicative on `D_K`, `α (T* X) = X` for `X ∈ D_K` and `T* (α A) = A` for `A ∈ D_H`;
4. (10): `α (A B) = α A * α B` for every `A ∈ A_α` and every `B`.
The standing assumption of the paper (p.3: "T always denotes a trace-preserving 2-positive
mapping T : B(H) → B(K) and we assume that the density matrices D1, D2, T(D1) and T(D2) are all
invertible") is in the binders: `T` is `ℂ`-linear, 2-positive (`IsTwoPositive`) and trace
preserving, and `D 0 = D₁`, `D 1 = D₂` are positive definite of trace one with `T (D s)` positive
definite.
The paper does not prove the lemma ("concerning the proof see 9.1 in [17]", Strătilă, *Modular theory in operator algebras*). (10) is Choi's
multiplicative-domain theorem for the unital 2-positive map α.
**Formalization Note.** `B(H)` is `Matrix H H ℂ` for a finite index type `H`. The two densities
are indexed by `Fin 2` (`D 0 = D₁`, `D 1 = D₂`); an invertible density is a positive definite matrix
of trace one. `T*` is the Hilbert–Schmidt adjoint `channelAdjoint T`. The maps α, T#, γ_H, γ_K and the
sets D_H, D_K are built from `D₁ = D 0`, as in the paper (p.4). "Algebraic isomorphism onto D_H" is read as: `D_K`, `D_H` closed under products (they are
the "fixed point algebras" of p.8), `T*` (linear by construction) multiplicative on `D_K`,
mapping `D_K` onto `D_H`, with two-sided inverse α. -/
theorem lemma_1
{H K : Type*} [Fintype H] [DecidableEq H] [Fintype K] [DecidableEq K]
(T : Matrix H H ℂ →ₗ[ℂ] Matrix K K ℂ) (hT : IsTwoPositive T)
(htr : ∀ A, (T A).trace = A.trace)
(D : Fin 2 → Matrix H H ℂ) (hD : ∀ s, (D s).PosDef ∧ (D s).trace = 1)
(hTD : ∀ s, (T (D s)).PosDef)
(hsuff : IsSufficient T (D 0) (D 1)) :
fixedPointsK T (D 0) ⊆ schwarzDomain (channelAdjoint T) ∧
fixedPointsH T (D 0) ⊆ schwarzDomain (alpha T (D 0)) ∧
((∀ X ∈ fixedPointsK T (D 0), ∀ Y ∈ fixedPointsK T (D 0), X * Y ∈ fixedPointsK T (D 0)) ∧
(∀ A ∈ fixedPointsH T (D 0), ∀ B ∈ fixedPointsH T (D 0), A * B ∈ fixedPointsH T (D 0)) ∧
(∀ X ∈ fixedPointsK T (D 0), channelAdjoint T X ∈ fixedPointsH T (D 0)) ∧
(∀ A ∈ fixedPointsH T (D 0), ∃ X ∈ fixedPointsK T (D 0), channelAdjoint T X = A) ∧
(∀ X ∈ fixedPointsK T (D 0), ∀ Y ∈ fixedPointsK T (D 0),
channelAdjoint T (X * Y) = channelAdjoint T X * channelAdjoint T Y) ∧
(∀ X ∈ fixedPointsK T (D 0), alpha T (D 0) (channelAdjoint T X) = X) ∧
(∀ A ∈ fixedPointsH T (D 0), channelAdjoint T (alpha T (D 0) A) = A)) ∧
∀ A ∈ schwarzDomain (alpha T (D 0)), ∀ B : Matrix H H ℂ,
alpha T (D 0) (A * B) = alpha T (D 0) A * alpha T (D 0) B := by sorry
end MosonyiPetz
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.