Regional entropy lower bound applies directly to physical hash selection
Provedmme_regional_physical_entropy_selected_boundregional-extractiontensor-complexity
For arbitrary regional histograms with a target reference and nonnegative tolerance, the explicit entropy retention expression bounds the physical hash selected-count lower bound from below.
Preamble
import Definitions.Def_mme_regional_entropy_copy_bound import Theorems.Thm_mme_regional_physical_hash_load_entropy_bounds import Theorems.Thm_mme_common_hash_scale_real_upper_bound import Theorems.Thm_mme_regional_target_entropy_bounds import Theorems.Thm_mme_entropy_retention_lower_bound import Mathlib open BigOperators MME MME.RegionRate MME.RegionRealization MME.RecursiveYZ MME.CompleteSplit open scoped Classical set_option autoImplicit false universe u
Formal statement
theorem mme_regional_physical_entropy_selected_bound {half R ell : ℕ} {parent : Fin R → Fin 3 → ℕ}
(htotal : ∀ r, parent r 0 + parent r 1 + parent r 2 = 2 * half)
(n : Fin R → ℕ) (m : ∀ r, RecursiveThinSplit.Split half (parent r) → ℕ)
(mu : Fin 3 → Cell half R parent → CompleteWord ell → ℕ)
(d : ℕ) (eps : ℝ) (heps : 0 ≤ eps)
(ref : Address half R parent n) (href : ref ∈ RecursiveXHash.target m) :
let Q := commonScale half (loadNum htotal m d (fun i ↦ mu (yzMode i))
(fun i _ ↦ parentTypical htotal n m (mu (yzMode i)) eps)) (loadDen m)
let factor := scaleFactor (half := half) (parent := parent) n d ell
let theta := scaleExponent htotal n m mu eps
Real.exp (regionalRate htotal n m mu - ((∑ r, n r : ℕ) : ℝ) *
entropyModulus (Fin 2 → CompleteWord ell) eps -
4 * Real.sqrt (Real.log factor + theta)) /
(32 * polynomialFactor n (Fintype.card (Cell half R parent)) * factor) ≤
((RecursiveXHash.target (n := n) m).card : ℝ) *
Real.exp (-4 * Real.sqrt (Real.log Q)) / (32 * Q) := by sorrySource
Generic physical hash bounds for regional extraction, derived from target entropy and physical hash load estimates.