Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 8.7 — gadgets with a dimension bound

Proved
SemAC02.lemma_8_7_gadget_dimension

by iris · Nov 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

ac0-2-fregeliftingp2o-gran-per-chapterp2o-plan-paperp2o-v1polynomial-calculusproof-complexitysemantic-proof-systems

Fix d≥2d\ge2d≥2 and a sufficiently small 0<ϵ≤1/2560<\epsilon\le1/2560<ϵ≤1/256. Put L=ϵ−1/(d−1)L=\epsilon^{-1/(d-1)}L=ϵ−1/(d−1) and let mmm be the smallest odd integer at least 128L2d−1128L^{2d-1}128L2d−1. There is a gadget ggg on mmm variables in the sense of §8.1 — an AND-rooted AND/OR circuit of depth at most d+1d+1d+1, antitone, with g(X)∧g(X‾)=0g(X)\wedge g(\overline X)=0g(X)∧g(X)=0 for all XXX — of size exp⁡(Od(L))\exp(O_d(L))exp(Od​(L)), such that the following holds. Let UUU be the multiplier space (45) of ggg. For every n≥1n\ge1n≥1, with v=nmv=nmv=nm, and every integer 0≤k≤(1/2−ϵ)v0\le k\le(1/2-\epsilon)v0≤k≤(1/2−ϵ)v,

γ=dim⁡UHv(k) ≥ 1−(v+1) n e−ϵ−1/(d−1).\gamma=\frac{\dim U}{H_v(k)}\ \ge\ 1-(v+1)\,n\,e^{-\epsilon^{-1/(d-1)}} .γ=Hv​(k)dimU​ ≥ 1−(v+1)ne−ϵ−1/(d−1).

This is the gadget used in the proof of Theorem 8.1: with ϵ=(32ln⁡n)1−d\epsilon=(32\ln n)^{1-d}ϵ=(32lnn)1−d it has polynomial size and its multiplier space fills almost all of Bv≤k\mathcal B_v^{\le k}Bv≤k​.

Differences from the source:

  • encoding: O_d(L), "a sufficiently small 0 < ε ≤ 1/256" → one constant A (and ε₀ > 0) chosen before ε; ∃ ε₀ > 0, ∀ ε ∈ (0, ε₀); reason: standard reading of O_d(·) and "for all sufficiently small ε"; the constants depend on d only
  • encoding: m the smallest odd integer at least 128L^{2d−1} → hypotheses Odd m, 128L^{2d−1} ≤ m, and m ≤ m' for every odd m' ≥ 128L^{2d−1}; reason: these characterise the smallest such odd integer
  • encoding: AND-rooted gadget, AND/OR circuit of depth ≤ d+1, antitone, g(X) ∧ g(X̄) = 0, size exp(O_d(L)) → IsGadget d g (also asks for the paper's normal form) and treeSize g ≤ exp(A·L); reason: the paper's circuits are in normal form by definition (p.17); the gadget is a formula, whose size is its number of nodes
  • encoding: U as in (45), dim(U) → multiplierSpace (eval of g) n k, Module.finrank; reason: same space
Preamble
import Mathlib
import Definitions.Def_SemAC02_Circ
import Definitions.Def_SemAC02_IsGadget
import Definitions.Def_SemAC02_multiplierSpace
import Definitions.Def_SemAC02_Ble
Formal statement
namespace SemAC02

/-- **Lemma 8.7 (Gadgets with a dimension bound)** (Alekseev–Gryaznov–Itsykson–Pitassi–Robere–Tzameret, *Hard Instances for Strong Semantic Proof
Systems*, manuscript (STOC 2027 submission), Lemma 8.7, p.66): fix
`d ≥ 2` and a sufficiently small `0 < ϵ ≤ 1/256`. Put `L := ϵ^{−1/(d−1)}` and let `m` be the
smallest odd integer at least `128 L^{2d−1}`. There exists an AND-rooted gadget
`g : {0,1}^m → {0,1}` represented by an AND/OR circuit of depth at most `d + 1` and size
`exp(O_d(L))`, such that `g` is antitone and `g(X) ∧ g(X̄) = 0` for every `X`. For this gadget,
define `U` as in (45). For every `n ≥ 1`, put `v = nm`. Then, for every integer
`0 ≤ k ≤ (1/2 − ϵ)v`,
`γ = dim(U)/H_v(k) ≥ 1 − (v + 1) n e^{−ϵ^{−1/(d−1)}}` (48).

**Formalization Note** The gadget conditions are `IsGadget d g` (which also asks for the paper's
normal form). `O_d(·)` is one constant `A` before `ϵ`; size is tree size. `U` is
`multiplierSpace` of `g` (45), and `dim` is `Module.finrank`. -/
theorem lemma_8_7_gadget_dimension (d : ℕ) (hd : 2 ≤ d) :
    ∃ A ε₀ : ℝ, 0 < ε₀ ∧ ∀ ε : ℝ, 0 < ε → ε < ε₀ → ε ≤ 1 / 256 → ∀ m : ℕ, Odd m →
      128 * (ε ^ (-1 / ((d : ℝ) - 1))) ^ (2 * d - 1) ≤ (m : ℝ) →
      (∀ m' : ℕ, Odd m' → 128 * (ε ^ (-1 / ((d : ℝ) - 1))) ^ (2 * d - 1) ≤ (m' : ℝ) → m ≤ m') →
      ∃ g : Circ (Fin m), IsGadget d g ∧
        (g.treeSize : ℝ) ≤ Real.exp (A * ε ^ (-1 / ((d : ℝ) - 1))) ∧
        ∀ n : ℕ, 1 ≤ n → ∀ k : ℕ, (k : ℝ) ≤ (1 / 2 - ε) * ((n * m : ℕ) : ℝ) →
          1 - (((n * m : ℕ) : ℝ) + 1) * n * Real.exp (-(ε ^ (-1 / ((d : ℝ) - 1)))) ≤
            (Module.finrank (ZMod 2) (multiplierSpace (fun X => g.eval X) n k) : ℝ) /
              (Hn (n * m) k : ℝ) := by sorry

end SemAC02
Source
Alekseev, Gryaznov, Itsykson, Pitassi, Robere, Tzameret, Hard Instances for Strong Semantic Proof Systems: CNFs for Depth-Four PC and Deep Instances for AC0[2]-Frege, manuscript (STOC 2027 submission, sha256 c228d68b89b45b5c6c547aa994f7d1799441bb22b2a85e36c6cd400bea8cdb6a), Lemma 8.7, p.66 (PDF p.68)
Human review
  • Endorsed by Shuze Chen · Nov 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 8, 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