Proposition 8.1.3 — Matrix relative entropy is nonnegative
ProvedTroppMatrixConcentration.ch8_entropy_nonnegativeFor any positive-definite complex matrices of the same positive dimension,
The matrices need not commute or have trace one. The corrected entropy used here provides the nonnegativity statement underlying the variational trace formula.
import Definitions.Def_TroppMatrixConcentration_ch8_entropy open scoped Matrix.Norms.L2Operator ComplexOrder
namespace TroppMatrixConcentration
theorem ch8_entropy_nonnegative {d : ℕ} [NeZero d]
(A H : Matrix (Fin d) (Fin d) ℂ) (hA : A.PosDef) (hH : H.PosDef) :
0 ≤ ch8_relativeEntropy A H := by sorry
end TroppMatrixConcentration
Read-back
What the Lean code literally says, in plain math · gpt-6
For every nonzero natural number and every pair of complex matrices and , with rows and columns indexed by , if and are positive definite (each is Hermitian and has strictly positive quadratic form on every nonzero complex vector), then . Here each matrix logarithm is obtained by applying the real natural logarithm to that matrix through continuous functional calculus; for these positive definite matrices, this applies the scalar logarithm to their positive eigenvalues. The trace is the sum of the diagonal entries, and the asserted inequality is an inequality of real numbers after taking its real part. There is no assumption that and commute or have a prescribed or equal trace. The dimension-zero case is excluded by the hypothesis .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.