Exponential integrability of an independent Hermitian sum
ProvedTroppMatrixConcentration.ch3_master_sum_exponential_integrablematrix-concentrationrandom-matrices
On any probability space, let be a finite independent family of measurable, almost surely Hermitian complex matrices, with . Fix . If every is Bochner integrable, then
The family may be empty and the parameter may be zero. No integrability of the summands themselves is assumed. This auxiliary lemma supplies the exponential integrability needed to apply Laplace bounds to independent sums.
Preamble
import Definitions.Def_TroppMatrixConcentration_probability open MeasureTheory ProbabilityTheory open scoped Matrix.Norms.L2Operator set_option autoImplicit false
Formal statement
namespace TroppMatrixConcentration
/-- Exponential integrability passes from independent Hermitian summands to their sum. -/
theorem ch3_master_sum_exponential_integrable {Ω : Type*} [MeasurableSpace Ω]
(μ : Measure Ω) [IsProbabilityMeasure μ] {d N : ℕ} [NeZero d]
(X : Fin N → Ω → Matrix (Fin d) (Fin d) ℂ) (θ : ℝ)
(hMeas : ∀ k, Measurable (X k))
(hHerm : ∀ k, ∀ᵐ ω ∂μ, (X k ω).IsHermitian)
(hIndep : iIndepFun X μ)
(hExp : ∀ k, Integrable (fun ω => matrixExp (θ • X k ω)) μ) :
Integrable (fun ω => matrixExp (θ • ∑ k, X k ω)) μ := by sorry
end TroppMatrixConcentration
Source
Joel A. Tropp, An Introduction to Matrix Concentration Inequalities, arXiv:1501.01571v1 (7 January 2015), https://arxiv.org/abs/1501.01571v1; Auxiliary regularity lemma for the proof of Theorem 3.6.1, printed p. 36; Section 2.2.1, printed p. 25, states the standing regularity convention. This is explicit analytic groundwork, not a separately numbered theorem in the book.