Cancellation of the log-product term in the hardcore derivative
ProvedZeroFreeRRG.Hardcore.hardcore_gradient_log_product_cancellationLet be independent from a mean-zero law supported on , and let . Writing and , the derivative numerator satisfies
The formal statement writes as the sum of the logarithms of its positive factors.
Preamble
import Mathlib set_option autoImplicit false open MeasureTheory open scoped Classical BigOperators
Formal statement
theorem ZeroFreeRRG.Hardcore.hardcore_gradient_log_product_cancellation (d : ℕ) (μ : Measure ℝ) [IsProbabilityMeasure μ] (θ : ℝ)
(hθ : 0 ≤ θ) (hθ1 : θ < 1)
(hb : ∀ᵐ x ∂μ, -1 ≤ x ∧ x ≤ 1)
(hint : Integrable (fun x : ℝ => x) μ) (hmean : (∫ x : ℝ, x ∂μ) = 0) :
(∫ xs : Fin d → ℝ, ((1+θ)*(∑ j, (∏ k ∈ Finset.univ.erase j, (1+θ*xs k))*xs j) +
(∏ k, (1-xs k)) - ∏ k, (1+θ*xs k)) * ∑ i, Real.log (1+θ*xs i)
∂(Measure.pi fun _ : Fin d => μ)) = 0 := 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, pp. 29–31.