Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1.2 — Narrow-DNF approximation

Proved
NarrowDNF.Approx.theorem_1_2

by mikedeng1 · Nov 10, 2026 · Mathlib 0df444a (Lean v4.33.1)

combinatoricsp2o-batch-stoc27ap2o-gran-per-chapterp2o-plan-paperp2o-v1probability

There is an absolute real constant C>0C>0C>0 such that, for every integer n≥1n\ge1n≥1, every K>0K>0K>0, every 0<p≤1/20<p\le1/20<p≤1/2, every 0<ε<10<\varepsilon<10<ε<1, and every coordinatewise increasing Boolean function f:{0,1}n→{0,1}f:\{0,1\}^n\to\{0,1\}f:{0,1}n→{0,1} satisfying Ip(f)≤KI_p(f)\le KIp​(f)≤K, there is a finite monotone DNF ggg such that

width⁡(g)≤exp⁡ ⁣(C(K+1)2ε2),Pp[f(X)≠g(X)]≤ε.\operatorname{width}(g)\le\exp\!\left(\frac{C(K+1)^2}{\varepsilon^2}\right),\qquad \mathbb P_p[f(X)\ne g(X)]\le\varepsilon.width(g)≤exp(ε2C(K+1)2​),Pp​[f(X)=g(X)]≤ε.

The constant is chosen before the dimension, influence bound, bias, tolerance and function. Influence is independent-coordinate resampling influence, and disagreement is the two-sided error under the actual product law. There is no mean, symmetry, Fourier-degree or term-count restriction. Both constant functions have width-zero witnesses.

This is the dimension- and low-bias-uniform approximation result; it asserts existence without an efficient-construction claim.

Differences from the source:

  • encoding: coordinates [n] and bits {0,1} → Fin n and Bool with pointwise false < true; reason: coordinate relabeling preserves all source instances
  • encoding: product probabilities and expectations → published Bernoulli event sums and exact real finite weighted sums; reason: the finite product law is identical and every cost is integrable
  • encoding: a monotone DNF with bounded width → a finite set of positive coordinate terms, its actual evaluation, and a real bound on its natural syntactic width; reason: every finite source DNF has the same representation after deleting duplicate terms, and every such witness is a source DNF
Preamble
import Definitions.Def_NarrowDNF_Approx_DNF
Formal statement
namespace NarrowDNF.Approx

theorem theorem_1_2 :
    ∃ C : ℝ, 0 < C ∧ ∀ n : ℕ, 1 ≤ n → ∀ K : ℝ, 0 < K →
      ∀ p : ℝ, 0 < p → p ≤ 1 / 2 → ∀ ε : ℝ, 0 < ε → ε < 1 →
      ∀ f : Cube n → Bool, Monotone f → FriedgutCoalition.totalPInfl p f ≤ K →
      ∃ g : DNF n, (width g : ℝ) ≤ Real.exp (C * (K + 1) ^ 2 / ε ^ 2) ∧
        error p f (evalDNF g) ≤ ε := by sorry

end NarrowDNF.Approx
Source
Chenghua Liu, Boning Meng, STOC 2027 submission #1509 (October 2026), Theorem 1.2, p. 2 (PDF p. 3); proof p. 9 (PDF p. 10)
Human review
  • Endorsed by Shuze Chen · Nov 10, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Nov 10, 2026

    Confirmed by the mission captain (proposal self-audit).

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