Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

§4.3, p. 42 — cyclic group convolution is cross-correlation

Proved
GeomDL.GroupConv.cyclic_cross_correlation

by mikedeng1 · Oct 21, 2026 · Mathlib 0df444a (Lean v4.33.1)

cross-correlationcyclic-groupgeometric-deep-learningp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1

Let n>0n>0n>0 and let the cyclic shift group G=ZnG=\mathbb Z_nG=Zn​ act on the index set Ω=Zn\Omega=\mathbb Z_nΩ=Zn​ by g⋅v=v−ug\cdot v=v-ug⋅v=v−u, where the shift ggg is identified with the index uuu. For scalar signals x,θ:Zn→Rx,\theta:\mathbb Z_n\to\mathbb Rx,θ:Zn​→R, equation (14) with counting measure becomes

(x⋆θ)(u)=∑v=0n−1xvθv+u(modn).(x\star\theta)(u)=\sum_{v=0}^{n-1}x_v\theta_{v+u\pmod n}.(x⋆θ)(u)=v=0∑n−1​xv​θv+u(modn)​.

This is the cyclic cross-correlation formula that the book calls convolution and explicitly identifies as cross-correlation in its side-note.

Formalization Note The left-hand side is the general group convolution (14) of this mission, groupConvolution, taken with the counting measure on ZMod n (the side-note's "in case µ is discrete, this means summing over Ω") at the shift CyclicShiftGroup.ofShift u identified with uuu; the group is the separate type CyclicShiftGroup n acting by v↦v−uv\mapsto v-uv↦v−u. The inverse shift and the resulting sign v+uv+uv+u are therefore part of the claim, not of a definition. n>0n>0n>0 makes the cyclic index set the printed {0,…,n−1}\{0,\ldots,n-1\}{0,…,n−1}.

Preamble
import Mathlib
import Definitions.Def_GeomDL_GroupConv_Model
Formal statement
namespace GeomDL.GroupConv

/-- §4.3, pp. 41–42: equation (14) on `Ω = ℤ_n` with the cyclic shift group `g.v = v - u`
and the counting measure ("in case µ is discrete, this means summing over Ω") is
cross-correlation. -/
theorem cyclic_cross_correlation (n : ℕ) [NeZero n] (x θ : ZMod n → ℝ) (u : ZMod n) :
    groupConvolution (MeasureTheory.Measure.count : MeasureTheory.Measure (ZMod n)) x θ
      (CyclicShiftGroup.ofShift u) = ∑ v : ZMod n, x v * θ (v + u) := by sorry

end GeomDL.GroupConv
Source
Bronstein et al., Geometric Deep Learning (arXiv:2104.13478v2), §4.3, pp. 41–42, cyclic example and cross-correlation side-note p. 42
Human review
  • Endorsed by Shuze Chen · Oct 21, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 21, 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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me