Theorem 3: two characterizations of equality in strong subadditivity
OpenQAlgorithms.Papers.Independent.Intakethird100.P84.equalityInStrongSubadditivityLet be an invertible density matrix on a tensor product of nonzero finite-dimensional complex Hilbert spaces. Let be its actual reduced densities, obtained by partial traces. Write , with natural logarithms. Each of the following conditions is separately equivalent to
This gives multiplicative and logarithmic characterizations of strong-subadditivity equality. The smaller densities are constructed from ; their invertibility follows from the hypothesis on the full state.
Formalization Note Registers have order . The entropy equality is written with the frozen entropy in bits; multiplication by converts it to the displayed equality in natural units. Powers and logarithms are taken before identity embedding.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakethird100_P84
noncomputable section
namespace QAlgorithms.Papers.Independent.Intakethird100.P84
open scoped Matrix ComplexConjugate ComplexOrder
attribute [local instance] Classical.propDecidable Classical.decEq
-- Pinned accepted definitions, inspected at their recorded source locations.
open QAlgorithms.Papers.Independent.Renew46.P35 (traceAdjoint)
open QAlgorithms.Papers.Independent.Dependency50.P21 (imaginaryPower marginalAB)
/-- Petz, arXiv:quant-ph/0209053v1, PDF p.11, Theorem 3.
Every marginal is computed from rho; the only invertibility assumption is on rho.
All smaller-system logarithms/powers are embedded with ordinary identities. -/
theorem equalityInStrongSubadditivity {I J K : Type*}
[Fintype I] [Fintype J] [Fintype K]
[Nonempty I] [Nonempty J] [Nonempty K]
(rho : Matrix (I × (J × K)) (I × (J × K)) ℂ)
(hrho : IsDensityMatrix rho) (hinv : IsUnit rho) :
let D12 := marginalAB rho
let D23 := traceLeft rho
let D2 := traceRight D23
let eqSSA := vnEntropy rho + vnEntropy D2 = vnEntropy D12 + vnEntropy D23
(eqSSA ↔ ∀ t : ℝ,
imaginaryPower rho t * embed12 (K := K) (imaginaryPower D12 (-t)) =
Matrix.kronecker (1 : Matrix I I ℂ) (imaginaryPower D23 t) *
Matrix.kronecker (1 : Matrix I I ℂ)
(Matrix.kronecker (imaginaryPower D2 (-t)) (1 : Matrix K K ℂ))) ∧
(eqSSA ↔
cfc Real.log rho - embed12 (K := K) (cfc Real.log D12) =
Matrix.kronecker (1 : Matrix I I ℂ) (cfc Real.log D23) -
Matrix.kronecker (1 : Matrix I I ℂ)
(Matrix.kronecker (cfc Real.log D2) (1 : Matrix K K ℂ))) := by sorry
end QAlgorithms.Papers.Independent.Intakethird100.P84
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be nonempty finite types and let , indexed by , be an invertible positive semidefinite complex matrix of trace one. Define its marginals by
Write , using spectral functional calculus, and write for the spectral function . For a matrix indexed by , let . The declaration asserts both equivalences
and
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.