Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5.1 (lattices, 3SAT) — fixed lattices with gap M1/s−εM^{1/s-\varepsilon}M1/s−ε in every ℓp\ell_pℓp​ norm

Proved
LatticePrepro.ThreeSAT.exists_lattice_gap

by WizardOfMenlo · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

coding-theorycvp-ncp-preprocessinghardness-of-approximationlatticespreprocessing

Let 1≤p≤∞1 \le p \le \infty1≤p≤∞ (a real number p≥1p \ge 1p≥1, or p=∞p = \inftyp=∞), let s=min⁡{p,2}s = \min\{p, 2\}s=min{p,2} (so s=2s = 2s=2 for p=∞p = \inftyp=∞), and let 0<ε<1/s0 < \varepsilon < 1/s0<ε<1/s. Then there is a constant C∈NC \in \mathbb{N}C∈N such that for every number of variables v≥1v \ge 1v≥1, with n=(2v)3n = (2v)^3n=(2v)3 the instance length of 3SAT on vvv variables, there exist

  • a finite coordinate set of size MMM with 1≤M≤nC1 \le M \le n^C1≤M≤nC, a lattice Λ⊆ZM\Lambda \subseteq \mathbb{Z}^MΛ⊆ZM (a subgroup) containing N⋅ZMN \cdot \mathbb{Z}^MN⋅ZM for some integer N>0N > 0N>0, so that Λ\LambdaΛ has full rank, and an integer target tx∈ZMt_x \in \mathbb{Z}^Mtx​∈ZM for every 3SAT instance xxx on vvv variables;
  • an integer radius R≥1R \ge 1R≥1,

such that, writing ∥⋅∥p\|\cdot\|_p∥⋅∥p​ for the ℓp\ell_pℓp​ norm and dist⁡p(t,Λ)=inf⁡u∈Λ∥t−u∥p\operatorname{dist}_p(t, \Lambda) = \inf_{u \in \Lambda} \|t - u\|_pdistp​(t,Λ)=infu∈Λ​∥t−u∥p​:

  1. YES instances. If xxx is satisfiable, then dist⁡p(tx,Λ)≤R\operatorname{dist}_p(t_x, \Lambda) \le Rdistp​(tx​,Λ)≤R.
  2. NO instances. If xxx is unsatisfiable, then
dist⁡p(tx,Λ)>M1/s−ε⋅R.\operatorname{dist}_p(t_x, \Lambda) > M^{1/s-\varepsilon} \cdot R.distp​(tx​,Λ)>M1/s−ε⋅R.

This is the lattice half of Theorem 5.1, instantiated with a verifier for 3SAT: one fixed lattice per number of variables separates satisfiable from unsatisfiable formulas with gap M1/s−εM^{1/s-\varepsilon}M1/s−ε. It is the combinatorial core of the CVP part of Theorem 1.1; the remaining ingredients there (polynomial-time computability and the complexity-theoretic argument) are not formalized.

Formalization Note: The conclusion is Nonempty (LatticeGap pn ε (SAT v) (len v) C), a structure bundling the lattice family (LatticeFamily: a finite coordinate type, a Submodule ℤ (M → ℤ) and the targets), RRR, NNN and the properties above. The norm exponent is a PNorm: any real p≥1p \ge 1p≥1 (the paper states rational ppp, which only matters for computability) or p=∞p = \inftyp=∞; ∥v∥p=(∑i∣vi∣p)1/p\|v\|_p = (\sum_i |v_i|^p)^{1/p}∥v∥p​=(∑i​∣vi​∣p)1/p uses real powers and ∥v∥∞=max⁡i∣vi∣\|v\|_\infty = \max_i |v_i|∥v∥∞​=maxi​∣vi​∣. dist⁡p\operatorname{dist}_pdistp​ is the real infimum latticeDist, and full rank is expressed by N⋅ZM⊆ΛN\cdot\mathbb{Z}^M \subseteq \LambdaN⋅ZM⊆Λ. Instances are encoded by one bit per possible clause (nnn is len v =(2v)3= (2v)^3=(2v)3) and indexed by v≥1v \ge 1v≥1; CCC may depend on ppp and ε\varepsilonε but not on vvv. Polynomial-time computability, the description-length bounds and the search test of Lemma 5.2 (stated separately for the concrete test) are not part of the statement.

Note (existence form). This statement asserts only that some lattice with targets has the gap. The Lean statement does not require the targets to be computable, so this form is also satisfied by a construction unrelated to the paper, in which the target records whether the instance is in the language. The statement for the object that the paper constructs is LatticePrepro.ThreeSAT.explicit_lattice_gap.

Preamble
import Mathlib
import Definitions.Def_LatticePrepro_Basic_Blocks
import Definitions.Def_LatticePrepro_Reduction_LinearSystem
import Definitions.Def_LatticePrepro_Conversions_Codes
import Definitions.Def_LatticePrepro_Conversions_ConstructionA
import Definitions.Def_LatticePrepro_Conversions_Hadamard
import Definitions.Def_LatticePrepro_Conversions_Lattices
import Definitions.Def_LatticePrepro_Lattices_Distance
import Definitions.Def_LatticePrepro_Lattices_Mathlib
import Definitions.Def_LatticePrepro_Moments_Basic
import Definitions.Def_LatticePrepro_Moments_Representation
import Definitions.Def_LatticePrepro_Verifier_Quadratic
import Definitions.Def_LatticePrepro_Reduction_FixedSystem
import Definitions.Def_LatticePrepro_Tables_Digits
import Definitions.Def_LatticePrepro_Tables_FieldSetup
import Definitions.Def_LatticePrepro_Tables_Hankel
import Definitions.Def_LatticePrepro_Tables_Construction
import Definitions.Def_LatticePrepro_Tables_ElementTable
import Definitions.Def_LatticePrepro_Tables_TableSoundness
import Definitions.Def_LatticePrepro_Tables_Explicit
import Definitions.Def_LatticePrepro_Main
import Definitions.Def_LatticePrepro_Verifier_Gates
import Definitions.Def_LatticePrepro_Verifier_Circuit
import Definitions.Def_LatticePrepro_Verifier_Expr
import Definitions.Def_LatticePrepro_Verifier_ThreeSAT
import Definitions.Def_LatticePrepro_ThreeSAT

open LatticePrepro
open LatticePrepro.ThreeSAT
Formal statement
theorem LatticePrepro.ThreeSAT.exists_lattice_gap (pn : PNorm) (ε : ℝ) (hε : 0 < ε) (hε1 : ε < 1 / pn.s) :
    ∃ C : ℕ, ∀ v : ℕ, 1 ≤ v → Nonempty (LatticeGap pn ε (SAT v) (len v) C) := by sorry
Source
G. Fenzi, G. Weissenberg, Nearly Optimal Hardness for CVP and NCP with Preprocessing (STOC 2027 submission #847; IACR ePrint 2026/2634, https://eprint.iacr.org/2026/2634), Theorem 5.1 (\label{thm:approximation-hardness}), lattices, for the 3SAT verifier; Lean: LatticePrepro/ThreeSAT.lean, LatticePrepro.ThreeSAT.exists_lattice_gap
Human review
  • Endorsed by Shuze Chen · Nov 15, 2026

    Confirmed by the moderator at approval.

  • Endorsed by WizardOfMenlo · Nov 15, 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