Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3.12 — formal-to-feasible expectation error

Proved
QuantumBipartiteTesting.Bipartiteness.lemma_3_12

by mikedeng1 · Nov 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

complexity-theorygraph-property-testingp2o-batch-stoc27ap2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-information

There are positive absolute constants C1,c1C_1,c_1C1​,c1​, independent of the circuit and grid point. For each fixed positive query-band constant BBB, the following holds eventually for every NNN, every actual qqq-query circuit with q≤BN1/3/log⁡Nq\le BN^{1/3}/\log Nq≤BN1/3/logN, every (M,ℓ)∈GN(M,\ell)\in\mathcal G_N(M,ℓ)∈GN​ and both signing laws σ\sigmaσ:

∣pσfeas−pσform∣≤exp⁡(C1qlog⁡N−c1N1/3).|p^{\rm feas}_\sigma-p^{\rm form}_\sigma|\le\exp(C_1q\log N-c_1N^{1/3}).∣pσfeas​−pσform​∣≤exp(C1​qlogN−c1​N1/3).

With the defined difference η=Δ−Δform\eta=\Delta-\Delta^{\rm form}η=Δ−Δform,

Δ=Δform+η,∣η∣≤2exp⁡(C1qlog⁡N−c1N1/3).\Delta=\Delta^{\rm form}+\eta,\qquad |\eta|\le2\exp(C_1q\log N-c_1N^{1/3}).Δ=Δform+η,∣η∣≤2exp(C1​qlogN−c1​N1/3).

There is a sufficiently small positive absolute c0c_0c0​ for which q≤c0N1/3/log⁡Nq\le c_0N^{1/3}/\log Nq≤c0​N1/3/logN implies uniform η=o(1)\eta=o(1)η=o(1) over all circuits and grid points. The latter means that every positive tolerance has one eventual size threshold working for all these data. This compares two separately constructed functionals; their equality is not assumed.

Differences from the source:

  • encoding: finite quantum query circuits with binary encoded outcomes and arbitrary workspace → published QueryAlg, arbitrary finite complex registers, injective null-zero bit encoding, unitary XOR oracle, input-independent output event; reason: every finite workspace is represented up to a basis bijection; unused answer bit strings and dummy vertex queries are retained
  • reading: formal expectation described both by (A.2)–(A.4) and by arbitrary output on overflow → the algebraic partition functional (A.2)–(A.4), followed by signing probabilities and linear extension; reason: these descriptions disagree at N=M=4,l=2: a one-edge moment is 1/2 algebraically versus 1/8 with empty output on overflow; the feasible law stays conditional and its comparison remains a theorem
  • reading: Definition 3.6 gives structural matching/response consistency and calls it equivalent to positive sampling probability → consistency is the structural condition, without a parameter-dependent positivity certificate; reason: a three-vertex two-color path is structurally consistent but cannot fit any block when N=M=4,l=2; retaining the structural definition preserves the full algebraic pattern domain
  • encoding: q=O(N^(1/3)/log N) and uniform o(1) → quantified positive band B with an eventual threshold, and an arbitrary positive tolerance uniformly over circuits and grid pairs; reason: retains the source asymptotic quantifier order and both error clauses
Preamble
import Definitions.Def_QuantumBipartiteTesting_Bipartiteness_Reduction
Formal statement
namespace QuantumBipartiteTesting.Bipartiteness
open Finset

theorem lemma_3_12 :
    (∃ C₁ c₁ : ℝ, 0 < C₁ ∧ 0 < c₁ ∧ ∀ B : ℝ, 0 < B → ∃ N₀ : ℕ,
      ∀ N ≥ N₀, ∀ k w q : ℕ, (q : ℝ) ≤ B * LowerBoundScale N →
      ∀ E : Encoding (Fin N × Bool) k, ∀ A : Algorithm (Address N 3) k w q,
      ∀ out : (Address N 3 × Bits k) × Fin w → Bool, ∀ M l : ℕ, Grid N M l →
        (∀ s : Signing, |FeasibleAcceptance E A out s M l - FormalAcceptance E A out s M l| ≤
          Real.exp (C₁*q*Real.log N - c₁*(N : ℝ)^(1/3 : ℝ))) ∧
        FeasibleGap E A out M l = FormalGap E A out M l + GapError E A out M l ∧
        |GapError E A out M l| ≤ 2*Real.exp (C₁*q*Real.log N - c₁*(N : ℝ)^(1/3 : ℝ))) ∧
    (∃ c₀ : ℝ, 0 < c₀ ∧ UniformGapVanishing c₀) := by sorry

end QuantumBipartiteTesting.Bipartiteness
Source
Chandrima Kayal, Sayantan Sen, Dániel Szabó, STOC 2027 submission #878 (October 2026), Lemma 3.12, PDF = printed page 17. https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/878.pdf
Human review
  • Endorsed by Shuze Chen · Nov 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Nov 9, 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