Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§IV.B: vectorization and transfer identities (172)

Proved
QAlgorithms.Papers.Independent.Advance50.P38.bipartiteVectorizationTransfer

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

prove2quantum

Let N1,N2>0N_1,N_2>0N1​,N2​>0, let X,Y∈CN1×N2X,Y\in\mathbb C^{N_1\times N_2}X,Y∈CN1​×N2​, and let V∈U(N1),W∈U(N2)V\in\mathrm U(N_1),W\in\mathrm U(N_2)V∈U(N1​),W∈U(N2​). For column vectorization and the usual conjugate-linear-first inner product,

vec⁡(VYW)=(W⊤⊗V)vec⁡Y,⟨vec⁡X,vec⁡Y⟩=tr⁡(X†Y).\operatorname{vec}(VYW)=(W^\top\otimes V)\operatorname{vec}Y,\qquad\langle\operatorname{vec}X,\operatorname{vec}Y\rangle=\operatorname{tr}(X^\dagger Y).vec(VYW)=(W⊤⊗V)vecY,⟨vecX,vecY⟩=tr(X†Y).

Putting x=vec⁡X,y=vec⁡Yx=\operatorname{vec}X,y=\operatorname{vec}Yx=vecX,y=vecY and U=W⊗VU=W\otimes VU=W⊗V gives tr⁡(xx†Uyy†U†)=∣⟨x,vec⁡(VYW⊤)⟩∣2=∣tr⁡(X†VYW⊤)∣2\operatorname{tr}(xx^\dagger Uyy^\dagger U^\dagger)=|\langle x,\operatorname{vec}(VYW^\top)\rangle|^2=|\operatorname{tr}(X^\dagger VYW^\top)|^2tr(xx†Uyy†U†)=∣⟨x,vec(VYW⊤)⟩∣2=∣tr(X†VYW⊤)∣2. In particular this trace is real and equals the transfer objective.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Advance50_P38
Formal statement
namespace QAlgorithms.Papers.Independent.Advance50.P38

open scoped BigOperators Topology Manifold ContDiff

noncomputable section

/-- Vectorization and pureStateTransfer (172), PDF p.27: column order, ordinary
transpose on W, and a modulus-square trace. -/
theorem bipartiteVectorizationTransfer {a b : ℕ} (ha : 0 < a) (hb : 0 < b)
    (X Y : Matrix (Fin a) (Fin b) ℂ)
    (V : Matrix.unitaryGroup (Fin a) ℂ) (W : Matrix.unitaryGroup (Fin b) ℂ) :
    colVec ((V : (Matrix (Fin a) (Fin a) ℂ)) * Y * (W : (Matrix (Fin b) (Fin b) ℂ))) =
      (Matrix.kronecker (W : (Matrix (Fin b) (Fin b) ℂ)).transpose (V : (Matrix (Fin a) (Fin a) ℂ))).mulVec (colVec Y) ∧
    vInner (colVec X) (colVec Y) = Matrix.trace (X.conjTranspose * Y) ∧
    Matrix.trace (rankOneMatrix (colVec X) * Matrix.kronecker (W : (Matrix (Fin b) (Fin b) ℂ)) (V : (Matrix (Fin a) (Fin a) ℂ)) *
      rankOneMatrix (colVec Y) * (Matrix.kronecker (W : (Matrix (Fin b) (Fin b) ℂ)) (V : (Matrix (Fin a) (Fin a) ℂ))).conjTranspose) =
      (Complex.normSq (Matrix.trace (X.conjTranspose * (V : (Matrix (Fin a) (Fin a) ℂ)) * Y * (W : (Matrix (Fin b) (Fin b) ℂ)).transpose)) : ℂ) ∧
    pureStateTransfer (colVec X) (colVec Y) (Matrix.kronecker (W : (Matrix (Fin b) (Fin b) ℂ)) (V : (Matrix (Fin a) (Fin a) ℂ))) =
      Complex.normSq (vInner (colVec X) (colVec ((V : (Matrix (Fin a) (Fin a) ℂ)) * Y * (W : (Matrix (Fin b) (Fin b) ℂ)).transpose))) := by sorry

end

end QAlgorithms.Papers.Independent.Advance50.P38
Source
Gradient Flows for Optimisation and Quantum Control: Foundations and Applications, arXiv:0802.4195v2, PDF p. 27, §IV.B: vectorization and transfer identities (172)
Read-back

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

For all positive natural numbers a,ba,ba,b, arbitrary complex a×ba\times ba×b matrices X,YX,YX,Y, and unitary matrices V∈U(a)V\in U(a)V∈U(a), W∈U(b)W\in U(b)W∈U(b), define cB(j,i)=Bijc_B(j,i)=B_{ij}cB​(j,i)=Bij​ for any a×ba\times ba×b matrix BBB, define ⟨u,v⟩=∑j=0b−1∑i=0a−1u(j,i)‾v(j,i)\langle u,v\rangle=\sum_{j=0}^{b-1}\sum_{i=0}^{a-1}\overline{u(j,i)}v(j,i)⟨u,v⟩=∑j=0b−1​∑i=0a−1​u(j,i)​v(j,i), and let RuR_uRu​ have entries (Ru)pq=upuq‾(R_u)_{pq}=u_p\overline{u_q}(Ru​)pq​=up​uq​​. With M=W⊗VM=W\otimes VM=W⊗V, all four identities hold:

cVYW=(WT⊗V)cY,⟨cX,cY⟩=tr⁡(X∗Y),tr⁡(RcXMRcYM∗)=∣tr⁡(X∗VYWT)∣2,Re⁡tr⁡(RcXMRcYM∗)=∣⟨cX,cVYWT⟩∣2.\begin{aligned} c_{VYW}&=(W^{\mathsf T}\otimes V)c_Y,\\ \langle c_X,c_Y\rangle&=\operatorname{tr}(X^*Y),\\ \operatorname{tr}(R_{c_X}MR_{c_Y}M^*)&= \left|\operatorname{tr}(X^*VYW^{\mathsf T})\right|^2,\\ \operatorname{Re}\operatorname{tr}(R_{c_X}MR_{c_Y}M^*)&= \left|\langle c_X,c_{VYW^{\mathsf T}}\rangle\right|^2. \end{aligned}cVYW​⟨cX​,cY​⟩tr(RcX​​MRcY​​M∗)Retr(RcX​​MRcY​​M∗)​=(WT⊗V)cY​,=tr(X∗Y),=​tr(X∗VYWT)​2,=∣⟨cX​,cVYWT​⟩∣2.​

Here T{}^{\mathsf T}T is ordinary transpose, ∗{}^*∗ is conjugate transpose, matrix action on a vector uses the usual sum over column indices, and the real number on the right of the third identity is regarded as a complex number. The vectors cX,cYc_X,c_YcX​,cY​ need not be normalized and may be zero.

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