Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5.8 — O(n1/4/ϵ)O(n^{1/4}/\sqrt\epsilon)O(n1/4/ϵ​) quantum-example query bound

Open
QuantumLTF.Example.theorem_5_8_query

by stockfish-soup · Nov 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

gaussianhermite-transformlinear-threshold-functionsquantum-examplesquantum-learningquantum-query-complexity

Let n≥2n\ge2n≥2, let w∈Sn−1w\in\mathbb S^{n-1}w∈Sn−1, and let 0<ϵ<1/20<\epsilon<1/20<ϵ<1/2. For a sufficiently fine power-of-two discretization MMM, coherent access to the discretized Gaussian quantum-example oracle QEX⁡M(fw,γn)\operatorname{QEX}_M(f_w,\gamma_n)QEXM​(fw​,γn​) and its inverse suffices to output a unit vector w~\widetilde ww satisfying

∥w~−w∥≤ϵ\|\widetilde w-w\|\le\epsilon∥w−w∥≤ϵ

with probability at least 2/32/32/3, using

O ⁣(n1/4ϵ)O\!\left(\frac{n^{1/4}}{\sqrt\epsilon}\right)O(ϵ​n1/4​)

oracle calls. The circuit is fixed before the unknown direction www, and the guarantee holds for every unitary extension of the prescribed example state.

This formalizes the query-complexity clause of Theorem 5.8, the paper’s quartic quantum speedup in dimension. The additional-gate and classical-postprocessing bounds are intentionally separated into later formalization targets rather than hidden in an unconstrained decoder.

Formalization Note The output coordinates are measured fixed-point numbers on designated wires, and the output is required to have unit norm exactly.

Preamble
import Mathlib
import Definitions.Def_QAlgorithms_Defs_Circuit
import Definitions.Def_QuantumLTF_Membership_gates12
import Definitions.Def_QuantumLTF_Membership_readFp
import Definitions.Def_QuantumLTF_Example_qex
Formal statement
namespace QuantumLTF.Example

open QAlgorithms QuantumLTF.Membership

/-- **Theorem 5.8, query-complexity clause.** A homogeneous LTF can be learned from the paper’s finite discretized Gaussian quantum-example oracle using `O(n^(1/4)/sqrt ε)` oracle calls. -/
theorem theorem_5_8_query :
    ∃ C : ℝ, 0 < C ∧ ∀ n : ℕ, 2 ≤ n → ∀ ε : ℝ, 0 < ε → ε < 1 / 2 →
      ∃ (q m ℓo fo : ℕ)
        (c : OracleCircuit gates12 (fun _ : Fin 1 => n * q + 1) m)
        (out : Fin n → (Fin ℓo ↪ Fin m)),
        (c.queryCount 0 : ℝ) ≤ C * Real.rpow n (1 / 4) / Real.sqrt ε ∧
        ∀ w : EuclideanSpace ℝ (Fin n), ‖w‖ = 1 →
          ∀ U : Matrix (Fin (n * q + 1) → Bool) (Fin (n * q + 1) → Bool) ℂ, IsQEX w U →
            probEvent (act (c.unitary fun _ => U) (zeroKet m))
              (fun z => let wt := WithLp.toLp 2 fun i => readFp (out i) fo z
                ‖wt‖ = 1 ∧ ‖wt - w‖ ≤ ε) ≥ 2 / 3 := by sorry

end QuantumLTF.Example
Source
Krivcenko, Nguyen, de Wolf, Fast Quantum Algorithms for Learning Linear Threshold Functions, arXiv:2609.40331v3, Theorem 5.8, pp. 72–74 (PDF pp. 73–75), query-complexity clause

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