Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 6.3 — Cofactor argument removes the bottom fan-in bound

Open
ParityBarrier563.Trilinear.Supplement.lemma_6_3

by mikedeng1 · Nov 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

boolean-circuitscomplexity-theoryp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let N≥0N\ge0N≥0, integers d,s≥1d,s\ge1d,s≥1, 0≤R≤N0\le R\le N0≤R≤N, a test ψ:{0,1}N→[−1,1]\psi:\{0,1\}^N\to[-1,1]ψ:{0,1}N→[−1,1], real k≥2log⁡2(4s)k\ge2\log_2(4s)k≥2log2​(4s) and γ≥0\gamma\ge0γ≥0 be given. Suppose that for every restriction τ fixing at most R coordinates and every layered depth-d circuit C of size at most s whose restricted bottom fan-in is at most k, ∣E[ψτCτ]∣≤γ|\mathbb E[\psi_\tau C_\tau]|\le\gamma∣E[ψτ​Cτ​]∣≤γ. Then every layered depth-d circuit C of size at most s satisfies

∣E[ψC]∣≤2γ+2−R/2.|\mathbb E[\psi C]|\le2\gamma+2^{-R/2}.∣E[ψC]∣≤2γ+2−R/2.

Restricted fan-in counts only live literals. This removes the bottom fan-in bound using hypotheses on every allowed cofactor.

Differences from the source:

  • encoding: layered circuits and their restricted bottom fan-in → a layering of the published DAG with a live-literal wire count; reason: constants and fixed literals are omitted from fan-in while every actual gate is counted
  • encoding: alternating layered circuits, with actual gate count and the stated fan-in bounds → Layering data over the published Circuit DAG, retaining constants and unary gates and counting literal or gate wires; reason: this is the source layer convention and keeps both output types; duplicate wires can be deleted without increasing any bound
Preamble
import Definitions.Def_ParityBarrier563_Trilinear_Supplement_Layering

set_option autoImplicit false
open scoped Classical
Formal statement
namespace ParityBarrier563.Trilinear.Supplement

/-- Lemma 6.3: every cofactor of at most R fixed bits, and real fan-in bounds k. -/
theorem lemma_6_3 (N d s R : ℕ) (hd : 1 ≤ d) (hs : 1 ≤ s) (hR : R ≤ N)
    (ψ : RealFn (Fin N)) (hψ : BoundedTest ψ) (k γ : ℝ)
    (hk : 2*Real.logb 2 (4*s) ≤ k) (hγ : 0 ≤ γ)
    (hcofactor : ∀ (τ : Restriction (Fin N)), fixedCount τ ≤ R →
      ∀ (C : Circuit (Fin N)) (L : Layering C d), C.size ≤ s → BottomBound L τ k →
        |avg (fun z => restrict τ ψ z * ind (restrict τ C.eval) z)| ≤ γ) :
  ∀ (C : Circuit (Fin N)) (L : Layering C d), C.size ≤ s →
    |avg (fun x => ψ x * ind C.eval x)| ≤ 2*γ + (2 : ℝ)^(-(R : ℝ)/2) := by sorry

end ParityBarrier563.Trilinear.Supplement
Source
Nikolai Chukhin, Xin Li, Ivan Mihajlin, Yan Zhong, Breaking the Parity Barrier for Constant-Depth Circuits, STOC 2027 submission #563 (October 2026), Lemma 6.3, printed p.27 / PDF p.29

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me