Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Denominator exponent pattern with alpha in the enlarged field

Open
QAlgorithms.Papers.Independent.Intakenext100.P03.denominator_pattern_alpha

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let nnn be a positive even integer. Put ζm=e2πi/m\zeta_m=e^{2\pi i/m}ζm​=e2πi/m, Rn=Z[ζ2n,1/2]R_n=\mathbb Z[\zeta_{2n},1/2]Rn​=Z[ζ2n​,1/2], Uz(θ)=diag⁡(1,eiθ)U_z(\theta)=\operatorname{diag}(1,e^{i\theta})Uz​(θ)=diag(1,eiθ), and Gn=⟨C,Uz(π/n)⟩G_n=\langle C,U_z(\pi/n)\rangleGn​=⟨C,Uz​(π/n)⟩, where C=⟨H0,S⟩C=\langle H_0,S\rangleC=⟨H0​,S⟩, H0=1+i2(111−1)H_0=\frac{1+i}{2}\begin{pmatrix}1&1\\1&-1\end{pmatrix}H0​=21+i​(11​1−1​) and S=diag⁡(1,i)S=\operatorname{diag}(1,i)S=diag(1,i).

For an axis p∈{x,y,z}p\in\{x,y,z\}p∈{x,y,z} with Pauli matrix PPP, write Up(θ)=1+eiθ2I+1−eiθ2PU_p(\theta)=\frac{1+e^{i\theta}}2I+\frac{1-e^{i\theta}}2PUp​(θ)=21+eiθ​I+21−eiθ​P. A canonical pair consists of an ordered finite list (pi,ai)i=1m(p_i,a_i)_{i=1}^m(pi​,ai​)i=1m​ and an actual matrix DDD, with consecutive axes distinct, 1≤ai<n/21\le a_i<n/21≤ai​<n/2, DDD proportional to an element of CCC, and U=(∏i=1mUpi(aiπ/n))DU=(\prod_{i=1}^m U_{p_i}(a_i\pi/n))DU=(∏i=1m​Upi​​(ai​π/n))D. Products have the printed left-to-right order, and the empty list is permitted.

Write n=2ksn=2^ksn=2ks with k≥1k\ge1k≥1 and positive odd sss. Define qa=2k−1−2k−jq_a=2^{k-1}-2^{k-j}qa​=2k−1−2k−j if n/gcd⁡(a,n)=2jn/\gcd(a,n)=2^jn/gcd(a,n)=2j, and qa=2k−1q_a=2^{k-1}qa​=2k−1 otherwise. Set α=21/2k−1\alpha=2^{1/2^{k-1}}α=21/2k−1 and β=2\beta=2β=2 for k=1k=1k=1, β=2cos⁡(π/2k)\beta=2\cos(\pi/2^k)β=2cos(π/2k) for k≥2k\ge2k≥2.

Use ordered Pauli axes X,Y,ZX,Y,ZX,Y,Z and the homomorphic Bloch matrix B(U)ij=12Re⁡tr⁡(PiUPjU†)B(U)_{ij}=\frac12\operatorname{Re}\operatorname{tr}(P_iUP_jU^\dagger)B(U)ij​=21​Retr(Pi​UPj​U†). This transposes printed equation (9), as required by the source's left-cancellation proof and row patterns.

Let U∈GnU\in G_nU∈Gn​ with B(U)∉B(C)B(U)\notin B(C)B(U)∈/B(C), take any canonical pair, and set N=∑iqaiN=\sum_iq_{a_i}N=∑i​qai​​. Work in the ring of integers OF\mathcal O_FOF​ of F=K=Q({ca,sa:1≤a<n/2},α), ca=αqacos⁡(πa/n), sa=αqasin⁡(πa/n)F=K=\mathbb Q(\{c_a,s_a:1\le a<n/2\},\alpha),\ c_a=\alpha^{q_a}\cos(\pi a/n),\ s_a=\alpha^{q_a}\sin(\pi a/n)F=K=Q({ca​,sa​:1≤a<n/2},α), ca​=αqa​cos(πa/n), sa​=αqa​sin(πa/n). A reduced exponent of a nonzero entry zzz means z=w/αrz=w/\alpha^rz=w/αr, with r∈Nr\in\mathbb Nr∈N, w∈OFw\in\mathcal O_Fw∈OF​, and www not divisible by α\alphaα. Coprimality means that the principal ideals sum to OF\mathcal O_FOF​.

  1. Every nonzero entry has a reduced exponent.
  2. The global maximum is attained and satisfies
max⁡i,j:B(U)ij≠0rij=N.\max_{i,j:B(U)_{ij}\ne0}r_{ij}=N.i,j:B(U)ij​=0max​rij​=N.

Exactly two rows contain an entry at exponent NNN with numerator coprime to α\alphaα. 3. Exactly one row has no entry at exponent NNN. It is the row indexed by the first axis p1p_1p1​, and its attained maximum is

max⁡j:B(U)p1j≠0rp1j=N−qa1.\max_{j:B(U)_{p_1j}\ne0}r_{p_1j}=N-q_{a_1}.j:B(U)p1​j​=0max​rp1​j​=N−qa1​​.

That row contains an entry at this exponent with numerator coprime to α\alphaα. The non-Clifford premise guarantees a nonempty list.

The pattern exposes the leading canonical axis while retaining the stronger coprimality conclusion.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P03
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P03

open scoped BigOperators Classical

-- Reuse the Mathlib matrix type without declaring a second Mat.

local notation "Mat" => Matrix Bool Bool ℂ

/-- Canonical data: the axis/exponent list and the actual residual matrix D,
not a nonunique phase/Clifford witness pair, PDF pp.6-7. -/
local notation "CanonicalData" => List (Fin 3 × ℕ) × Matrix Bool Bool ℂ

/-- Theorem 6.3(a)-(c), PDF pp.16-18. Uses the enlarged field K and alpha,
not the smaller denominatorBeta field. Same disclosed Bloch transpose correction. -/
theorem denominator_pattern_alpha (n k s : ℕ) (hk : 1 ≤ k) (hs : 0 < s)
    (hodd : Odd s) (hfactor : n = 2 ^ k * s)
    (U : Mat) (hU : U ∈ cyclotomicGroup n) (hnot : homomorphicBloch U ∉ blochCliffords)
    (d : CanonicalData) (hd : IsCyclotomicCanonical n U d) :
    ∃ p a rest, d.1 = (p, a) :: rest ∧
      DenominatorPattern (numeratorField n k) (denominatorAlpha k) (homomorphicBloch U)
        (canonicalTotalWeight n k d) (denominatorWeight n k a) p := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P03

end
Source
Exact synthesis of single-qubit unitaries over Clifford-cyclotomic gate sets, arXiv:1501.04944v2, PDF p. 16, Theorem 6.3
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

For all natural numbers n,k,sn,k,sn,k,s satisfying k≥1k\ge1k≥1, s>0s>0s>0, sss odd, and n=2ksn=2^k sn=2ks, define Z(θ)=diag⁡(1,eiθ)Z(\theta)=\operatorname{diag}(1,e^{i\theta})Z(θ)=diag(1,eiθ), let C\mathcal CC be the multiplicative submonoid generated by H0=1+i2(111−1)H_0=\frac{1+i}{2}\left(\begin{smallmatrix}1&1\\1&-1\end{smallmatrix}\right)H0​=21+i​(11​1−1​), S=diag⁡(1,i)S=\operatorname{diag}(1,i)S=diag(1,i), and their conjugate transposes, and let Gn\mathcal G_nGn​ be the multiplicative submonoid generated by C\mathcal CC, Z(π/n)Z(\pi/n)Z(π/n), and Z(π/n)†Z(\pi/n)^\daggerZ(π/n)†. Index the Pauli matrices by 0,1,20,1,20,1,2, with σ0=(0110)\sigma_0=\left(\begin{smallmatrix}0&1\\1&0\end{smallmatrix}\right)σ0​=(01​10​), σ1=(0−ii0)\sigma_1=\left(\begin{smallmatrix}0&-i\\i&0\end{smallmatrix}\right)σ1​=(0i​−i0​), and σ2=diag⁡(1,−1)\sigma_2=\operatorname{diag}(1,-1)σ2​=diag(1,−1), and define the real matrix

B(U)ij=12Re⁡tr⁡(σiUσjU†).B(U)_{ij}=\frac12\operatorname{Re}\operatorname{tr}(\sigma_i U\sigma_j U^\dagger).B(U)ij​=21​Retr(σi​Uσj​U†).

For every complex 2×22\times22×2 matrix U∈GnU\in\mathcal G_nU∈Gn​ such that B(U)≠B(C′)B(U)\ne B(C')B(U)=B(C′) for every C′∈CC'\in\mathcal CC′∈C, and every pair d=(L,C)d=(L,C)d=(L,C), where LLL is a finite list of pairs (p,a)∈{0,1,2}×N(p,a)\in\{0,1,2\}\times\mathbb N(p,a)∈{0,1,2}×N and CCC is a complex 2×22\times22×2 matrix, assume that every (p,a)(p,a)(p,a) in LLL satisfies 1≤a<⌊n/2⌋1\le a<\lfloor n/2\rfloor1≤a<⌊n/2⌋, consecutive pairs have different first coordinates, some real ϕ\phiϕ satisfies eiϕC∈Ce^{i\phi}C\in\mathcal CeiϕC∈C, and

U=(∏(p,a) in L[1+eiaπ/n2I+1−eiaπ/n2σp])C,U=\left(\prod_{(p,a)\text{ in }L} \left[\frac{1+e^{i a\pi/n}}2 I+ \frac{1-e^{i a\pi/n}}2\sigma_p\right]\right)C,U=​(p,a) in L∏​[21+eiaπ/n​I+21−eiaπ/n​σp​]​C,

with the product taken in list order. Set α=21/2k−1\alpha=2^{1/2^{k-1}}α=21/2k−1, and, for c∈Nc\in\mathbb Nc∈N, set

W(c)={2k−1−2 k−log⁡2(n/gcd⁡(c,n)),n/gcd⁡(c,n)=2j for some j∈N,2k−1,otherwise,W(c)= \begin{cases} 2^{k-1}-2^{\,k-\log_2(n/\gcd(c,n))}, & n/\gcd(c,n)=2^j\text{ for some }j\in\mathbb N,\\ 2^{k-1},&\text{otherwise}, \end{cases}W(c)={2k−1−2k−log2​(n/gcd(c,n)),2k−1,​n/gcd(c,n)=2j for some j∈N,otherwise,​

using the natural-number logarithm and subtraction truncated at zero. Let FFF be the smallest subfield of C\mathbb CC containing α\alphaα and all αW(c)cos⁡(πc/n)\alpha^{W(c)}\cos(\pi c/n)αW(c)cos(πc/n), αW(c)sin⁡(πc/n)\alpha^{W(c)}\sin(\pi c/n)αW(c)sin(πc/n) for 1≤c<⌊n/2⌋1\le c<\lfloor n/2\rfloor1≤c<⌊n/2⌋, and put OF={z∈F:z is integral over Z}\mathcal O_F=\{z\in F:z\text{ is integral over }\mathbb Z\}OF​={z∈F:z is integral over Z}. For a real number zzz and r∈Nr\in\mathbb Nr∈N, define R(z,r)R(z,r)R(z,r) to mean that z≠0z\ne0z=0 and there exists w∈OFw\in\mathcal O_Fw∈OF​ such that z=w/αrz=w/\alpha^rz=w/αr and it is not the case that α,w∈OF\alpha,w\in\mathcal O_Fα,w∈OF​, α≠0\alpha\ne0α=0, and w=αvw=\alpha vw=αv for some v∈OFv\in\mathcal O_Fv∈OF​. Define E(i,r)E(i,r)E(i,r) to mean that R(B(U)ij,r)R(B(U)_{ij},r)R(B(U)ij​,r) holds for some column jjj, and define K(i,r)K(i,r)K(i,r) to mean that some column jjj has B(U)ij≠0B(U)_{ij}\ne0B(U)ij​=0 and

B(U)ij=w/αr,w,α,u,v∈OF,uw+vα=1B(U)_{ij}=w/\alpha^r,\qquad w,\alpha,u,v\in\mathcal O_F,\qquad uw+v\alpha=1B(U)ij​=w/αr,w,α,u,v∈OF​,uw+vα=1

for some w,u,vw,u,vw,u,v. Then there exist p,ap,ap,a and a remaining list L′L'L′ such that L=(p,a)::L′L=(p,a)::L'L=(p,a)::L′. With N=∑(p′,c) in LW(c)N=\sum_{(p',c)\text{ in }L}W(c)N=∑(p′,c) in L​W(c) and q=W(a)q=W(a)q=W(a), every nonzero entry of B(U)B(U)B(U) satisfies R(z,r)R(z,r)R(z,r) for some natural rrr; every rrr satisfying R(B(U)ij,r)R(B(U)_{ij},r)R(B(U)ij​,r) is at most NNN; some row satisfies E(i,N)E(i,N)E(i,N); exactly two rows satisfy K(i,N)K(i,N)K(i,N); there is exactly one row for which E(i,N)E(i,N)E(i,N) fails; and every such row is ppp, has all its reduced exponents at most N−qN-qN−q, and satisfies both E(i,N−q)E(i,N-q)E(i,N−q) and K(i,N−q)K(i,N-q)K(i,N−q), where N−qN-qN−q is natural subtraction truncated at zero.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 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