Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 4.13 — from the single-cycle to the product distribution

Proved
WeldedPath.lemma_4_13

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

compressed-oraclep2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1quantum-query-complexitywelded-trees

There is an absolute constant CCC such that the following holds. Let nnn be even, l∈[2,n−4]l\in[2,n-4]l∈[2,n−4], GnG_nGn​ a valid welded trees instance, and AAA a quantum algorithm making q≥1q\ge 1q≥1 queries to Oσ(wt)O^{(wt)}_\sigmaOσ(wt)​ that accepts when its final measurement outcome lies in a fixed set. Writing pD(A)p_D(A)pD​(A) for its acceptance probability when σ∼D\sigma\sim Dσ∼D,

pDsing(A) ≤ C(pDprod(A)+n q2N),N=2n.p_{D_{\mathrm{sing}}}(A)\ \le\ C\Big(p_{D_{\mathrm{prod}}}(A)+\frac{n\,q^2}{N}\Big),\qquad N=2^n .pDsing​​(A) ≤ C(pDprod​​(A)+Nnq2​),N=2n.

Conditioning the weld on being one cycle changes acceptance probabilities only by the additive O(nq2/N)O(nq^2/N)O(nq2/N), so the analysis can work with the product distribution.

Differences from the source:

  • reading: acceptance probability of a quantum algorithm → probability that the final computational-basis outcome lies in a fixed set acc not depending on σ; reason: an algorithm accepts on a fixed set of measurement outcomes
  • reading: O(p_{D_prod}(A) + n·q²/N) → one absolute constant C chosen before n, l, the instance, the workspace, q, A and acc; reason: the bound is uniform in all parameters
  • generalization: the standing assumption (49) of §4 → not assumed; reason: the proof (pp.41–42) uses only Lemma 3.2 and the hybrid argument; adding (49) gives the source's statement
  • generalization: one fixed instance and coloring (pp.14–15) → every valid weld and proper 3-coloring; reason: the fixed instance is one of them
Preamble
import Mathlib
import Definitions.Def_QAlgorithms_Defs_Query
import Definitions.Def_WeldedPath_successProb
Formal statement
namespace WeldedPath

open QAlgorithms

/-- **Lemma 4.13** (Carolan, Childs, Coudron, Gilani, Miloschewsky, Podder, arXiv:2611.05912
(sha256 b31400a0…), Lemma 4.13, p.41). Let `p_D(A)` be the acceptance probability of a quantum
algorithm `A` making `q ≥ 1` queries to an oracle sampled from `D`. Then
`p_{D_sing}(A) = O(p_{D_prod}(A) + n·q²/N)`, `N = 2^n`.

**Formalization Note.** `A` accepts when the final computational-basis measurement gives an
outcome in the fixed set `acc` (which does not depend on `σ`). `O(·)` is one absolute constant
`C`, chosen before `n`, `l`, the instance, the workspace, `q`, `A` and `acc`. The standing
assumptions of §3.1 (`n` even, `l ∈ [2, n − 4]`) are hypotheses; (49) is not. -/
theorem lemma_4_13 :
    ∃ C : ℝ, 0 < C ∧
      ∀ n l : ℕ, Even n → 2 ≤ l → l ≤ n - 4 →
      ∀ G : WeldedInstance n, G.IsValid →
      ∀ (W : Type) [Fintype W] [DecidableEq W] (q : ℕ), 1 ≤ q →
      ∀ (A : QueryAlg (Qubits (2 * n + 1) × Fin 3) (Qubits (2 * n + 1)) W q)
        (acc : ((Qubits (2 * n + 1) × Fin 3) × Qubits (2 * n + 1)) × W → Prop),
      G.probOver l (G.singSupport l) A (fun _ => acc) ≤
        C * (G.probOver l (G.prodSupport l) A (fun _ => acc) +
          (n : ℝ) * (q : ℝ) ^ 2 / 2 ^ n) := by sorry

end WeldedPath
Source
Carolan, Childs, Coudron, Gilani, Miloschewsky, Podder, Algorithmic Wave-Particle Duality: Pathfinding in Welded Trees is Hard, arXiv:2611.05912 (sha256 b31400a0…), Lemma 4.13, p.41
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