Documentation

Atlas.HighDimensionalStatistics.code.Chapter5.Prop_5_6

Proposition 5.6: Properties of KL Divergence #

This file provides the bundled statement of Proposition 5.6:

  1. Nonnegativity (Gibbs' inequality): KL(P, Q) ≥ 0
    • ENNReal form: trivial by type
    • Real-valued form (substantive, via Jensen): 0 ≤ ∫ llr dP + Q(Ω) − P(Ω)
  2. Tensorization (additivity for products):
    • Binary: KL(P₁⊗P₂, Q₁⊗Q₂) = KL(P₁,Q₁) + KL(P₂,Q₂)
    • General n-fold: KL(⊗ᵢ Pᵢ, ⊗ᵢ Qᵢ) = Σᵢ KL(Pᵢ,Qᵢ)

The individual parts and their proofs are in Def_5_5.lean.

Helper lemmas for the n-fold tensorization #

klDiv is invariant under pushforward by a measurable equivalence. Uses the f-divergence formula: klDiv μ ν = ∫⁻ x, ofReal(klFun(rnDeriv μ ν x)) ∂ν

Binary tensorization of KL divergence for probability measures (unconditional). Uses the Mathlib chain rule for compProd.

theorem Rigollet.Chapter5.prop_5_6_bundle :

Proposition 5.6 (Properties of KL divergence, bundled).