(7), p. 4 — relative entropy through the relative modular operator: S(ρ‖σ) = ⟨ρ^{1/2}, −ln(Δ)(ρ^{1/2})⟩
ProvedSimpleSSA.Monotone.eq_7Let and be invertible density matrices on a finite-dimensional system . Let be the relative modular operator, a strictly positive linear map on for the Hilbert–Schmidt inner product , and let be its logarithm, obtained by diagonalizing and taking the logarithm of its eigenvalues. Then the relative entropy satisfies
The identity combines the two logarithms of the relative entropy, which need not commute, into the logarithm of a single positive map. It is the form in which monotonicity (4) is proved in §3, through (8)–(12).
Formalization Note Invertibility of and is the paper's own standing assumption of §3 (p. 3: "In defining this operator we will assume that and are invertible"). Linear maps on matrices are represented by matrices on in the row-major vectorization ; then is the Kronecker matrix and is its matrix logarithm by functional calculus, exactly the p. 4 definition. The equality is stated in , so it also asserts that the inner product is real. Logarithms are natural.
import Mathlib import Definitions.Def_SimpleSSA_Monotone_Basic open scoped Matrix.Norms.L2Operator ComplexOrder Kronecker Matrix
namespace SimpleSSA.Monotone
open TroppMatrixConcentration QAlgorithms
/-- (7), p. 4: for invertible density matrices `ρ`, `σ` (the §3 assumption, p. 3),
`S(ρ‖σ) = ⟨ρ^{1/2}, −ln(Δ)(ρ^{1/2})⟩`, where `ln(Δ)` is the logarithm of the relative modular
operator `Δ(X) = σ X ρ⁻¹` as a strictly positive linear map on matrices. -/
theorem eq_7 {A : Type} [Fintype A] [DecidableEq A] [Nonempty A]
(ρ σ : Matrix A A ℂ) (hρ : IsDensityMatrix ρ) (hρ_inv : ρ.PosDef)
(hσ : IsDensityMatrix σ) (hσ_inv : σ.PosDef) :
(relEntropy ρ σ : ℂ) =
hsInner (sqrtM ρ) (applyHS (-(matrixLog (modularOp ρ σ))) (sqrtM ρ)) := by sorry
end SimpleSSA.Monotone
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.