Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Claim 4.14 — Wt=ker⁡(CLt∣Wt−1)W_t=\ker(C_{L_t}|_{W_{t-1}})Wt​=ker(CLt​​∣Wt−1​​) and Xt=rank⁡(CLt∣Wt−1)X_t=\operatorname{rank}(C_{L_t}|_{W_{t-1}})Xt​=rank(CLt​​∣Wt−1​​)

Proved
F3SubsetSum.claim_4_14

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

binary-error-lwecommutative-algebraf3-subset-sump2o-gran-per-chapterp2o-plan-paperp2o-v1polynomial-systemsquantum-algorithms

Let q≥3q\ge3q≥3 be prime and D≥2D\ge2D≥2. For a history a1,…,at−1a_1,\dots,a_{t-1}a1​,…,at−1​ and a new sample ata_tat​, let Vt−1⊆Vt⊆RDV_{t-1}\subseteq V_t\subseteq R_DVt−1​⊆Vt​⊆RD​ be the sample images (39), Ws=Vs⊥⊆RD∗W_s=V_s^\perp\subseteq R_D^*Ws​=Vs⊥​⊆RD∗​ their annihilators, and CLt:RD∗→RD−2∗C_{L_t}:R_D^*\to R_{D-2}^*CLt​​:RD∗​→RD−2∗​ the contraction by Lt=⟨at,x⟩L_t=\langle a_t,x\rangleLt​=⟨at​,x⟩. Then

Wt=ker⁡(CLt∣Wt−1)andXt=rank⁡(CLt∣Wt−1).W_t=\ker\bigl(C_{L_t}|_{W_{t-1}}\bigr)\qquad\text{and}\qquad X_t=\operatorname{rank}\bigl(C_{L_t}|_{W_{t-1}}\bigr).Wt​=ker(CLt​​∣Wt−1​​)andXt​=rank(CLt​​∣Wt−1​​).

This moves the martingale step into the dual space, where it becomes a rank question.

Differences from the source:

  • encoding: the standing integer D ≥ 2 of §4.1 (eq. (29), p.18), left implicit in the statement → hypothesis D ≥ 2; reason: R_{D−2} is undefined for D < 2 (natural-number subtraction would make it R_0)
  • encoding: ker(C_{L_t}|{W{t−1}}) as a subspace of RD → the kernel, a subspace of W{t−1}, mapped into R_D by the inclusion; reason: the paper identifies the two
Preamble
import Mathlib
import Definitions.Def_F3SubsetSum_TruncatedRing
import Definitions.Def_F3SubsetSum_SampleImage
import Definitions.Def_F3SubsetSum_SymbolicRank
Formal statement
namespace F3SubsetSum

/-- **Claim 4.14** (Kothari, Metger, O'Donnell, Shutty, Wu, arXiv:2609.40321v1, Claim 4.14, p.25):
"Let C_{L_t}|_{W_{t−1}} : W_{t−1} → R*_{D−2} be the restriction of C_{L_t} to W_{t−1}. Then
W_t = ker(C_{L_t}|_{W_{t−1}}) and X_t = rank(C_{L_t}|_{W_{t−1}})."

Over `F_q`, `q ≥ 3` prime, `D ≥ 2`: for a history `a_1, …, a_{t−1}` and a new sample `a_t`, with
`V_{t−1}`, `V_t` the sample images (39) and `W = V^⊥ ⊆ R*_D` their annihilators (p.24): `W_t` is
the kernel of the contraction `C_{L_t}` (p.25) restricted to `W_{t−1}`, viewed inside `R*_D`, and
the increment `X_t` (40) is the rank of that restriction.

**Formalization Note** The kernel, a subspace of `W_{t−1}`, is mapped into `R*_D` by the
inclusion. -/
theorem claim_4_14 (q : ℕ) [Fact q.Prime] (hq : 3 ≤ q) (n D : ℕ) (hD : 2 ≤ D) {t : ℕ}
    (a : Fin t → Fin n → ZMod q) (a' : Fin n → ZMod q) :
    (sampleImage q (Fin.snoc a a') D).dualAnnihilator =
        (LinearMap.ker ((contraction (ZMod q) n q D a').domRestrict
            (sampleImage q a D).dualAnnihilator)).map
          (sampleImage q a D).dualAnnihilator.subtype ∧
      increment q a a' D =
        Module.finrank (ZMod q) (LinearMap.range ((contraction (ZMod q) n q D a').domRestrict
          (sampleImage q a D).dualAnnihilator)) := by sorry

end F3SubsetSum
Source
Kothari, Metger, O'Donnell, Shutty, Wu, Exponential quantum speedup for F_3^n-Subset-Sum? Or, rigorous classical algorithms for Binary-Error LWE, arXiv:2609.40321v1, Claim 4.14, p.25
Human review
  • Endorsed by Shuze Chen · Nov 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 6, 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