Deterministic pushforward contracts finite L1 distance
ProvedAFT.FiniteDist.l1_map_lefinite-probabilityinfo-loss-lar-native-5c3becbstoc-2027
Let be probability laws on a finite set , and let map into a finite set. Their pushforwards satisfy
Thus deterministic postprocessing cannot increase this distance.
Preamble
import Definitions.Def_InfoLossLAR_Native_AFT_FiniteConcentration
import Mathlib
/-! # Total variation of actual independent finite products -/
namespace AFT.FiniteDist
end AFT.FiniteDist
section AFT
section FiniteDist
open AFT
open AFT.FiniteDist
open scoped BigOperators Classical
noncomputable section
set_option backward.isDefEq.respectTransparency false
variable {A B : Type*} [Fintype A] [Fintype B]
Formal statement
theorem AFT.FiniteDist.l1_map_le (first second : FiniteDist A) (map : A → B) :
(∑ b,|(first.map map).weight b-(second.map map).weight b|) ≤ ∑ a,|first.weight a-second.weight a| := by sorrySource
Local manuscript repository commit 5c3becb254daaa8738368b474349766c1769c2a0; lean/AFT/FiniteProductVariation.lean; lines 10–24; retained native proof and type checked against the native theorem.