Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Abstract and §2, pp. 1–4 — every single-qubit unitary within ε by O(log(1/ε)) H, T, CNOT gates and two |0⟩ ancillas

Proved
CliffordTApprox.Ancilla.main_result

by mikedeng1 · Oct 22, 2026 · Mathlib 0df444a (Lean v4.33.1)

approximationclifford-tgate-synthesisp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-circuits

There is an absolute constant A>0A > 0A>0 such that for every 2×22\times22×2 unitary matrix UUU and every 0<ε≤10 < \varepsilon \le 10<ε≤1 there is a circuit CCC of gates from {H,T,CNOT}\{H, T, \mathrm{CNOT}\}{H,T,CNOT} on three wires, the data qubit followed by two ancillas initialized to ∣0⟩|0\rangle∣0⟩, such that

∥ ψ↦C(ψ⊗∣00⟩) − ψ↦(Uψ)⊗∣00⟩ ∥F≤εand#C≤A(log⁡21ε+1).\big\|\,\psi\mapsto C(\psi\otimes|00\rangle)\ -\ \psi\mapsto(U\psi)\otimes|00\rangle\,\big\|_F \le \varepsilon \qquad\text{and}\qquad \#C \le A\left(\log_2\frac1\varepsilon + 1\right).​ψ↦C(ψ⊗∣00⟩) − ψ↦(Uψ)⊗∣00⟩​F​≤εand#C≤A(log2​ε1​+1).

This is the paper's main result: single-qubit unitaries are approximated to precision ε\varepsilonε with O(log⁡(1/ε))O(\log(1/\varepsilon))O(log(1/ε)) Clifford and T gates using two ancillary qubits, which matches the Ω(log⁡(1/ε))\Omega(\log(1/\varepsilon))Ω(log(1/ε)) lower bound and improves on the O(log⁡3+δ(1/ε))O(\log^{3+\delta}(1/\varepsilon))O(log3+δ(1/ε)) of the Solovay–Kitaev algorithm.

Formalization Note. The error is the Frobenius norm of the difference of two linear maps from one qubit to three qubits: it is sensitive to the global phase of UUU and charges any amplitude left on the ancillas, and it dominates the per-state distance used on the page. "Up to two ancillary qubits" is formalized as exactly two ancilla wires (an unused one is idle). The gate set is {H,T,CNOT}\{H, T, \mathrm{CNOT}\}{H,T,CNOT} (S=T2S=T^2S=T2, X=HT4HX=HT^4HX=HT4H). The page's O(log⁡(1/ε))O(\log(1/\varepsilon))O(log(1/ε)) is pinned as A(log⁡2(1/ε)+1)A(\log_2(1/\varepsilon)+1)A(log2​(1/ε)+1) with AAA quantified before UUU and ε\varepsilonε; the +1+1+1 makes the bound meaningful for ε\varepsilonε near 111.

Preamble
import Mathlib
import Definitions.Def_CliffordTApprox_Ancilla_Setting
Formal statement
namespace CliffordTApprox.Ancilla

open QAlgorithms
open QAlgorithms.Papers.Independent.Renew46.P12

/-- Abstract, p. 1; §2, pp. 2–4: there is an absolute constant `A` such that every single-qubit
unitary `U` is approximated to precision `ε ∈ (0, 1]` (Frobenius norm of the initialized map,
phase-sensitive, ancilla leakage charged) by an `{H, T, CNOT}` circuit on the data wire and two
ancillas initialized to `|0⟩`, with at most `A (log₂(1/ε) + 1)` gates. -/
theorem main_result :
    ∃ A : ℝ, 0 < A ∧ ∀ U : Matrix (Qubits 1) (Qubits 1) ℂ,
      U ∈ Matrix.unitaryGroup (Qubits 1) ℂ →
      ∀ ε : ℝ, 0 < ε → ε ≤ 1 →
      ∃ c : Circuit htcxGateSet (1 + 2),
        frobNorm (initializedMap c.unitary - idealMap U 2) ≤ ε ∧
        (c.size : ℝ) ≤ A * (Real.log (1 / ε) / Real.log 2 + 1) := by sorry

end CliffordTApprox.Ancilla
Source
Kliuchnikov, Maslov, Mosca, Asymptotically optimal approximation of single qubit unitaries by Clifford and T circuits using a constant number of ancillary qubits, arXiv:1212.0822v2, p. 1, Abstract; §2, pp. 2–4
Human review
  • Endorsed by Shuze Chen · Oct 22, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 22, 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