Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 4.15 — one-factor gain with coefficient-degree loss

Proved
AC02Frege.lemma_4_15_one_factor_gain

by iris · Nov 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

ac0-2-fregemajority-liftingp2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1polynomial-calculusproof-complexity

Let ℓ\ellℓ be odd, and consider a line with multilinear polynomials u1,…,usu_1,\dots,u_su1​,…,us​ of degree at most rrr and an intermediate approximator PPP. Suppose K≥rK\ge rK≥r and Q+deg⁡xP+3r≤DQ+\deg_xP+3r\le DQ+degx​P+3r≤D. Then some factor q=1+∑iaiuiq=1+\sum_ia_iu_iq=1+∑i​ai​ui​, ai∈F2a_i\in\mathbb F_2ai​∈F2​, satisfies

cPq(K) ≤ cP(K)−12cP(K−r)+EK+12EK−r.c_{Pq}(K)\ \le\ c_P(K)-\tfrac12c_P(K-r)+E_K+\tfrac12E_{K-r}.cPq​(K) ≤ cP​(K)−21​cP​(K−r)+EK​+21​EK−r​.

Appending a good Razborov–Smolensky factor removes half of the error conditions visible at degree K−rK-rK−r, up to the dimensions lost to the restriction β≤Q\beta\le Qβ≤Q.

Differences from the source:

  • encoding: ℓ odd, m = (ℓ − 1)/2 (standing assumption of §4) → hypothesis Odd ℓ, m = ℓ / 2; reason: same instances
  • encoding: polynomials over F2 identified with functions on the cube; products of functions → MvPolynomial over ZMod 2; products multilinearized by ML (membership in Ĉ_D and β depend only on ML p); reason: Definition 2.1 identifies a function with its multilinear polynomial; no instance changes
  • generalization: P an intermediate approximator → every multilinear P; reason: the statement and its proof use only deg_x P; specializes to the source
  • encoding: a line with polynomials u_1, …, u_s; c_P(j) → u : Fin s → multilinear polynomials of degree ≤ r; M_N(j) − finrank of the span of Z_P(j); reason: Definition 4.14; same values
Preamble
import Mathlib
import Definitions.Def_LiftSAParity_CNF
import Definitions.Def_AC02Frege_ML
import Definitions.Def_AC02Frege_PC
import Definitions.Def_AC02Frege_lift
import Definitions.Def_AC02Frege_majorityExpansion
import Definitions.Def_AC02Frege_rankCounts
Formal statement
namespace AC02Frege

open MvPolynomial LiftSAParity

/-- Alekseev–Gryaznov–Pitassi–Robere–Shekhovtsov–Tzameret, *Progress on AC0[2]-Frege Lower Bounds*, manuscript (sha256 88d149fa…), Lemma 4.15, (24), p.25 (one-factor gain with coefficient-degree loss): for a line with
polynomials `u_1, …, u_s` of degree at most `r` and an intermediate approximator `P`, if
`K ≥ r` and `Q + deg_x P + 3r ≤ D` then some factor `q = 1 + Σ_i a_i u_i`, `a_i ∈ 𝔽₂`, satisfies
`c_{Pq}(K) ≤ c_P(K) − ½ c_P(K − r) + E_K + ½ E_{K−r}`.

**Formalization Note** `ℓ` odd; `u_i` and `P` multilinear; `Pq` is the product of functions.
Stated for every multilinear `P` (the proof uses only `deg_x P`). -/
theorem lemma_4_15_one_factor_gain {n ℓ : ℕ} (hℓ : Odd ℓ) (F : CNF (Fin n)) (D Q K r s : ℕ)
    (u : Fin s → MvPolynomial (Fin n × Fin ℓ) (ZMod 2))
    (hu : ∀ i, IsMultilinear (u i) ∧ (u i).totalDegree ≤ r)
    (P : MvPolynomial (Fin n × Fin ℓ) (ZMod 2)) (hP : IsMultilinear P)
    (hrK : r ≤ K) (hD : Q + P.totalDegree + 3 * r ≤ D) :
    ∃ a : Fin s → ZMod 2,
      (cErr F ℓ D K (P * (1 + ∑ i, a i • u i)) u : ℝ) ≤
        (cErr F ℓ D K P u : ℝ) - (1 / 2) * (cErr F ℓ D (K - r) P u : ℝ)
          + (Eexcl n ℓ Q K : ℝ) + (1 / 2) * (Eexcl n ℓ Q (K - r) : ℝ) := by sorry

end AC02Frege
Source
Alekseev, Gryaznov, Pitassi, Robere, Shekhovtsov, Tzameret, Progress on AC0[2]-Frege Lower Bounds (manuscript dated November 3, 2026, STOC 2027 submission, sha256 88d149fa2bcae702e55a6b72efecda2b4137fd18dca3e23a056c3db5da0ece61), Lemma 4.15, (24), p.25; proof App. B.2, pp.44–45
Human review
  • Endorsed by Shuze Chen · Nov 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 7, 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