Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eqs.(40)–(42): covariance of graph-state encoding

Proved
QAlgorithms.Papers.Independent.Advance50.P50.graph_encoding_covariance

by Gohan · Oct 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let G,HG,HG,H be simple graphs on the same nnn vertices and ppp a vertex permutation satisfying {p(i),p(j)}∈E(H)\{p(i),p(j)\}\in E(H){p(i),p(j)}∈E(H) exactly when {i,j}∈E(G)\{i,j\}\in E(G){i,j}∈E(G). Let PPP be its physical site permutation, t>0t>0t>0, and ρin\rho_{\rm in}ρin​ a permutation-invariant density matrix. Put HG=∑{i,j}∈E(G)ZiZj+∑iXiH_G=\sum_{\{i,j\}\in E(G)}Z_iZ_j+\sum_iX_iHG​=∑{i,j}∈E(G)​Zi​Zj​+∑i​Xi​, with unordered edges counted once, WG=e−itHGW_G=e^{-itH_G}WG​=e−itHG​, and ρG=WGρinWG†\rho_G=W_G\rho_{\rm in}W_G^\daggerρG​=WG​ρin​WG†​. Then

HH=PHGP†,WH=PWGP†,ρH=PρGP†.H_H=PH_GP^\dagger,\qquad W_H=PW_GP^\dagger,\qquad \rho_H=P\rho_GP^\dagger.HH​=PHG​P†,WH​=PWG​P†,ρH​=PρG​P†.

The encoding therefore respects vertex relabelling and isomorphism-class labels. No injective graph encoding or distinguishing evolution time is claimed.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Advance50_P50
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Advance50.P50

open MeasureTheory ProbabilityTheory

open scoped ComplexConjugate ENNReal

-- Exact accepted primitives; no paper-local redeclarations.

open QAlgorithms.Papers.Independent.Information100.P63 (Mat tensorPower)

open QAlgorithms.Papers.Independent.Gap48.P20 (Copies)

open QAlgorithms.Papers.Independent.Gap48.P38 (subsystemReduction)

open QAlgorithms.Papers.Independent.Information100.P32 (purity)

open QAlgorithms.Papers.Independent.Information100.P08 (IsPure)

-- Matrices carry the associative commutator bracket X*Y-Y*X.

attribute [local instance 100] LieRing.ofAssociativeRing

/-- PDF pp.14–15, Eqs.(40)–(42): encoding is covariant under a vertex
bijection preserving adjacency. Printed P rho P is corrected to P rho P†;
general permutations are not self-adjoint. No injectivity is claimed.
Printed: PDF pp.14-15 defines graph isomorphism by an edge bijection and repeatedly writes P rho P and P H P for arbitrary permutations P.
Changed: Require a vertex permutation preserving adjacency and use P rho P-dagger, P H P-dagger and P exp(-itH) P-dagger throughout.
Why: An arbitrary edge bijection only fixes edge counts. The Ising-encoding proof relabels vertices and conjugates physical operators. A three-cycle has P-dagger different from P, so the adjoint is indispensable. No separating observable or injective encoding is added. -/
theorem graph_encoding_covariance {n : ℕ} (G H : SimpleGraph (Fin n))
    (p : Equiv.Perm (Fin n)) (hiso : ∀ i j, H.Adj (p i) (p j) ↔ G.Adj i j)
    (t : ℝ) (ht : 0 < t) (rho : Mat (Qubits n)) (hrho : IsDensityMatrix rho)
    (hfixed : ∀ q : Equiv.Perm (Fin n), conjSuperop (sitePermutation q) rho = rho) :
    graphHamiltonian H = conjSuperop (sitePermutation p) (graphHamiltonian G) ∧
    graphEvolution H t = conjSuperop (sitePermutation p) (graphEvolution G t) ∧
    graphEncoding H t rho = conjSuperop (sitePermutation p) (graphEncoding G t rho) := by sorry

end QAlgorithms.Papers.Independent.Advance50.P50

end
Source
Group-Invariant Quantum Machine Learning, arXiv:2205.02261v2, PDF p. 14, Eqs.(40)–(42): covariance of graph-state encoding
Read-back

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

For every natural number nnn, simple graphs G,HG,HG,H on nnn sites, and permutation ppp of those sites satisfying H(p(i),p(j))H(p(i),p(j))H(p(i),p(j)) is an edge exactly when G(i,j)G(i,j)G(i,j) is an edge, take any real t>0t>0t>0 and any positive semidefinite matrix ρ\rhoρ of trace 111 on nnn-bit configurations. Suppose PqρPq†=ρP_q\rho P_q^\dagger=\rhoPq​ρPq†​=ρ for every site permutation qqq, where Pq(y,x)=1{y=x∘q−1}P_q(y,x)=\mathbf 1_{\{y=x\circ q^{-1}\}}Pq​(y,x)=1{y=x∘q−1}​. Define HK=∑i<j, K(i,j) an edgeZiZj+∑iXiH_K=\sum_{i<j,\ K(i,j)\text{ an edge}}Z_iZ_j+\sum_iX_iHK​=∑i<j, K(i,j) an edge​Zi​Zj​+∑i​Xi​, UK(t)=e−itHKU_K(t)=e^{-itH_K}UK​(t)=e−itHK​, and EK,t(ρ)=UK(t)ρUK(t)†\mathcal E_{K,t}(\rho)=U_K(t)\rho U_K(t)^\daggerEK,t​(ρ)=UK​(t)ρUK​(t)†, where Z=diag⁡(1,−1)Z=\operatorname{diag}(1,-1)Z=diag(1,−1), X=(0110)X=\begin{pmatrix}0&1\\1&0\end{pmatrix}X=(01​10​), and each MiM_iMi​ acts by MMM at site iii and the identity elsewhere. Then

HH=PpHGPp†,UH(t)=PpUG(t)Pp†,EH,t(ρ)=PpEG,t(ρ)Pp†.H_H=P_pH_GP_p^\dagger,\qquad U_H(t)=P_pU_G(t)P_p^\dagger,\qquad \mathcal E_{H,t}(\rho)=P_p\mathcal E_{G,t}(\rho)P_p^\dagger.HH​=Pp​HG​Pp†​,UH​(t)=Pp​UG​(t)Pp†​,EH,t​(ρ)=Pp​EG,t​(ρ)Pp†​.

The quantification includes n=0n=0n=0, when the Hamiltonians vanish and the site permutation is unique.

Human review
  • Endorsed by Shuze Chen · Oct 29, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 29, 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