Aether Catalog definitions — Probability_WignerSecondMomentConcentration
DefinitionProbability_WignerSecondMomentConcentrationaether-catalogprobability
Definition bundle for the Aether Catalog module Probability.WignerSecondMomentConcentration, transplanted by skeleton subtraction; supplies the types and constants the catalog's theorems import.
Definition code
-- Def bundle generated from Probability/WignerSecondMomentConcentration.lean by skeleton subtraction
import Mathlib
import Definitions.Def_Probability_WignerUniversalFourthMoment
/-
Copyright (c) 2026 Harmonic. All rights reserved.
Released under Apache 2.0 license.
# Convergence in probability of the second spectral moment of a Wigner matrix
For an arbitrary centred, unit-variance entry law `ℒ` (finitely supported, fourth
moment `m₄`), the second moment of the empirical spectral distribution of `W/√N`
* has expectation exactly `1 - 1/N`, and
* has variance exactly `2 (m₄ - 1) (N-1) / N³`,
so by Chebyshev's inequality it converges **in probability** to `1 = C₁`, the
second moment of the Wigner semicircle law. This is the `k = 2` case of the
semicircle law, proved for a general Wigner ensemble rather than for a single
entry distribution.
-/
open Matrix BigOperators Finset Filter Topology
open RademacherWigner (edgeOf)
open scoped Classical
namespace WignerUniversal
variable {S : Type*} [Fintype S] {N : ℕ}
/-! ### Basic probabilistic infrastructure -/
/-- The weight (product law probability) of a configuration. -/
noncomputable def weight (L : EntryLaw S) (ω : Conf N S) : ℝ := ∏ e, L.w (ω e)
/-- The probability of an event, i.e. of a finite set of configurations. -/
noncomputable def gprob (L : EntryLaw S) (A : Finset (Conf N S)) : ℝ :=
∑ ω ∈ A, weight L ω
/-! ### The second trace moment and its square -/
/-! ### The second moment of the second trace moment -/
/-- `ne i j = 1` iff `i ≠ j`. -/
def neInd (i j : Fin N) : ℝ := if i = j then 0 else 1
/-- Indicator of the coinciding ordered edge pair `(k,l) = (i,j)`. -/
def pairInd1 (i j k l : Fin N) : ℝ :=
(if i = j then 0 else 1) * (if k = i then 1 else 0) * (if l = j then 1 else 0)
/-- Indicator of the reversed edge pair `(k,l) = (j,i)`. -/
def pairInd2 (i j k l : Fin N) : ℝ :=
(if i = j then 0 else 1) * (if k = j then 1 else 0) * (if l = i then 1 else 0)
/-! ### Mean, variance and convergence in probability -/
end WignerUniversal