Quantum Sibson identity for the Petz divergence
OpenQAlgorithms.Papers.Independent.Intakenext100.P15.quantum_sibson_identityAll Hilbert spaces are finite-dimensional and nonzero. A density operator is positive semidefinite with trace one; singular states are included. All logarithms are base two. For real , denotes the sandwiched Rényi divergence and the traditional Petz divergence, with value on support failure and negative spectral powers taken on the support.
For a bipartite density , let and put , , and . Partial-trace cyclicity gives .
For every density and every , the normalization satisfies , the constructed is a density, and
The equality holds in the extended reals for singular comparisons as well. This decomposition supplies the explicit minimizing comparison state for Petz mutual information.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P15
noncomputable section
open scoped Matrix ComplexOrder ComplexConjugate Topology
open Filter
namespace QAlgorithms.Papers.Independent.Intakenext100.P15
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
-- These are pinned accepted definitions, not other chunks from the current run.
open QAlgorithms.Papers.Independent.Bridge50.P43 (NormalizedDensity payoff)
open QAlgorithms.Papers.Independent.Dependency50.P09 (Divergence logExt sandwichedD petzD)
open QAlgorithms.Papers.Independent.Information100.P34 (choi)
open QAlgorithms.Papers.Independent.Information100.P63 (schatten)
open QAlgorithms.Papers.Independent.Information100.P71
(supportPower SupportIncluded stateRelative channelPower diamond)
/-- PDF pp.14--15, Lemma 7, (6.7)--(6.8). Z>0 and normalization are conclusions;
the optimizer is constructed from the actual marginal and retains singular comparisons. -/
theorem quantum_sibson_identity
{A B : Type} [Fintype A] [Fintype B] [DecidableEq A] [DecidableEq B]
[Nonempty A] [Nonempty B] (rho : NormalizedDensity (A × B)) (sigma : NormalizedDensity B)
(alpha : ℝ) (ha : 1 < alpha) :
0 < sibsonZ alpha rho.1 ∧ IsDensityMatrix (sibsonState alpha rho.1) ∧
sibsonK alpha rho.1 =
traceLeft (Matrix.kronecker (supportPower (traceRight rho.1) (1 - alpha))
(1 : Matrix B B ℂ) * supportPower rho.1 alpha) ∧
petzD alpha rho.1 (Matrix.kronecker (traceRight rho.1) sigma.1) =
petzD alpha (sibsonState alpha rho.1) sigma.1 +
petzD alpha rho.1 (Matrix.kronecker (traceRight rho.1) (sibsonState alpha rho.1)) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P15
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be finite nonempty index sets with decidable equality, let be a positive semidefinite matrix of trace on , let be a positive semidefinite matrix of trace on , and let . Write , where , and write . For a positive semidefinite matrix , define by spectral functional calculus using the scalar function that is at and at every nonzero eigenvalue; in particular, zero eigenvalues remain zero even for negative exponents. Define
For matrices , the divergence used here is if ; otherwise it is
with arithmetic in the extended real numbers, where for and for . The assertion is that , that is positive semidefinite with trace , and that
No invertibility of or is assumed; the equality uses the stated extended-real conventions.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.