Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-generatedPipeline-generatedjudge pass (gpt-6-luna)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Purifications exist and are unitarily equivalent after padding

Statement

Every density operator ρ on a nonzero finite-dimensional complex-inner-product space H has a purification on some finite-dimensional auxiliary space. More precisely, if r=rank⁡(ρ), it has one on an auxiliary space of dimension r. If Ψi∈H⊗Ki (i=1,2) are any two purifications of the same ρ, then for every integer m≥max⁡(dim⁡K1,dim⁡K2) and every pair of linear isometric embeddings ιi:Ki→Cm, where Cm has its standard complex inner product, there is a unitary U on Cm such that

(IH⊗U)(IH⊗ι2)Ψ2=(IH⊗ι1)Ψ1.

Facts & Assumptions

Given: A density operator ρ on nonzero finite-dimensional H, and when proving equivalence, two finite-dimensional purifications Ψi∈H⊗Ki of ρ, an integer m≥max⁡(dim⁡K1,dim⁡K2), and linear isometric embeddings ιi:Ki→Cm with the standard inner product on Cm.

[F1]

Density operators have an orthogonal spectral ensemble: ρ=∑j=1rλj∣ej⟩⟨ej∣ with λj>0, orthonormal ej, and ∑jλj=1.

[F2]

Ensemble and purification: a purification is a unit vector whose auxiliary partial trace is ρ.

[F3]

Schmidt decomposition: a finite bipartite vector has a Schmidt decomposition with positive coefficients and orthonormal factor families.

[F4]

Ordered tensor registers, their inner product, and wire permutations, Tensor products of finitely many linear maps and Dirac Notation Conversion: tensor factors are ordered with product pairings, tensor maps apply each linear map to its factor, and ∣x⟩⟨y∣ acts by z↦⟨z,y⟩x.

[F5]

Partial trace over a finite tensor factor: in ordered orthonormal product bases, partial trace sums entries with equal auxiliary indices.

[F7]

The partial trace is characterized by the trace pairing and is basis-independent: the partial-trace formula is independent of the chosen ordered orthonormal basis of the auxiliary space.

[F12]

Rank and nullity of a linear map with finite-dimensional domain: the rank of ρ is the dimension of its image.

[F13]

The product basis is orthonormal and tensor dimensions multiply: elementary tensors of ordered factor bases form a basis of the tensor product, and if the factor bases are orthonormal then so is this product basis.

[F11]

Linear isometries, and orthogonal or unitary operators on finite-dimensional inner product spaces: a norm-preserving linear map is a linear isometry, and a bijective linear isometry on a finite-dimensional complex inner-product space is unitary.

Proof

1.1F1F2F4F5F9F12constructalgebra

Use [F1] and let K0=Cr with the standard ordered basis relabelled f1,…,fr, orthonormal by [F9]. Define Ψ0=∑j=1rλj ej⊗fj. The product pairing [F4] gives ∥Ψ0∥2=∑jλj=1, and [F5] gives Tr⁡K0(∣Ψ0⟩⟨Ψ0∣)=∑jλj∣ej⟩⟨ej∣=ρ. Thus Ψ0 is a purification on an auxiliary space of dimension r=rank⁡(ρ) by [F12].

1.2F1F2F3F4F5F6F7F9F11F13choosealgebra

For each i, choose an orthonormal basis (gj(i))j<di of Ki by [F6], and put hj(i)=ιigj(i). Since ιi is linear and norm-preserving by [F11], each hj(i) has norm one. For j≠ℓ, the four vectors gj(i)±gℓ(i) and gj(i)±igℓ(i) have squared norm 2, so their images do too. With the root inner product convention, for z=⟨hj(i),hℓ(i)⟩ the differences ∥hj(i)+hℓ(i)∥2−∥hj(i)−hℓ(i)∥2=2(z+z‾) and ∥hj(i)+ihℓ(i)∥2−∥hj(i)−ihℓ(i)∥2=2i(z‾−z) both vanish. Thus z+z‾=0 and z=z‾, so z=0; the image vectors are orthonormal. Let Li=span⁡{hj(i):j<di}. By [F6], Cm=Li⊕Li⊥ and Li⊥ is finite-dimensional; choose an orthonormal basis of Li⊥ by [F6] and append it to (hj(i)) to obtain an ordered orthonormal basis of Cm. Choose an ordered orthonormal basis of H by [F6]. The product basis result [F13] shows that IH⊗ιi sends the orthonormal product basis of H⊗Ki to an orthonormal family, so the padded vector Ψ~i=(IH⊗ιi)Ψi is still a unit vector. Write Ψi=∑α,jcαjeα⊗gj(i) in this product basis. In the extended basis of Cm, Ψ~i=∑α,jcαjeα⊗hj(i) has zero coefficients on every complement basis vector, so the partial-trace coordinate sum [F5] over the extended basis equals the original partial trace over (gj(i)), namely ρ. By [F3] write each Ψ~i in Schmidt form ∑jsj(i)uj(i)⊗fj(i); [F5]–[F7] give its marginal ∑j(sj(i))2∣uj(i)⟩⟨uj(i)∣=ρ. Therefore the vectors with (sj(i))2=λ form an orthonormal basis of each positive λ-eigenspace of ρ.

2.1step 1.2F4F10F11constructalgebra

Fix an eigenvalue λ>0. By step 1.2, its two lists of Schmidt vectors are orthonormal bases of the same eigenspace. Write uλ,j(2)=∑k(Vλ)kjuλ,k(1). Orthonormality gives ∑k(Vλ)kj(Vλ)kℓ‾=δjℓ; conjugating this identity gives Vλ∗Vλ=I, so the square matrix Vλ is unitary. Prescribe fλ,j(2)↦∑k(Vλ)kj‾fλ,k(1). For indices j,ℓ, the inner product of the two images is ∑k(Vλ)kj‾(Vλ)kℓ=δjℓ; distinct eigenvalue blocks remain orthogonal. Thus this basis map extends linearly by [F10] to an onto isometry between the Schmidt-support subspaces.

3.1F1F6F8F10F11step 1.2step 2.1algebrachoose∎

By [F6], each padded auxiliary space is the orthogonal direct sum of its Schmidt-support subspace and its orthogonal complement. Step 1.2 shows that each support has dimension r, so [F8] gives dimension m−r for both complements. Choose orthonormal bases for them by [F6] and prescribe a bijection between those bases. The complement map and the support isometry from step 2.1 together send an orthonormal basis of Cm to an orthonormal basis of Cm; [F10] extends this map uniquely and linearly, and the basis map is surjective and preserves norms. Hence [F11] makes it a unitary U. In each eigenvalue block, ∑j(Vλ)kj(Vλ)ℓj‾=δkℓ, and therefore ∑jλ uλ,j(2)⊗Ufλ,j(2)=∑k=1mλλ uλ,k(1)⊗fλ,k(1). Summing these equalities over the positive eigenvalues gives the displayed equivalence after padding.

Remarks

The proof uses finite-dimensional spectral families, orthonormal complements and finite basis extensions; it makes no arbitrary-index choice.

Depends on

Used by

Cited to discharge well-definedness by Ensemble and purification.

Dependency tree · two levels

91 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources