Differentiation under the hardcore vertex-entropy expectation
ProvedZeroFreeRRG.Hardcore.hardcore_entropy_integral_derivativeLet be any probability measure supported on , and let independent messages have law . For , the expected hardcore vertex entropy
is differentiable, and its derivative is the expectation of the pointwise derivative displayed in the formal statement. Here .
Preamble
import Mathlib set_option autoImplicit false open MeasureTheory Set open scoped Classical BigOperators Topology
Formal statement
theorem ZeroFreeRRG.Hardcore.hardcore_entropy_integral_derivative (d : ℕ) (η : Measure ℝ) [IsProbabilityMeasure η]
(hb : ∀ᵐ h ∂η, -1 ≤ h ∧ h ≤ 1) (θ : ℝ) (hθ : 0 < θ) (hθ1 : θ < 1) :
let μ := Measure.pi (fun _ : Fin d => η)
let V := fun (t : ℝ) (h : Fin d → ℝ) =>
((∏ i, (1+t*h i)) + t*(∏ i, (1-h i)))/(1+t)
let D := fun (t : ℝ) (h : Fin d → ℝ) =>
((1+t)*(∑ i, (∏ j ∈ Finset.univ.erase i, (1+t*h j))*h i) +
(∏ i, (1-h i)) - ∏ i, (1+t*h i)) / (1+t)^2
HasDerivAt (fun t => ∫ h, V t h * Real.log (V t h) ∂μ)
(∫ h, D θ h * (Real.log (V θ h) + 1) ∂μ) θ := by sorrySource
Chen–Wang–Zhang, Zero-Freeness for Random Regular Graphs up to the Replica Symmetry Breaking Threshold, https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/1461.pdf, proof of Lemma 4.17, equation (25), pp. 29–31.