Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-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.

Robust one-step OAA including ancilla leakage

Statement

Let S be a nonzero finite-dimensional system space, let A be the complete clean-ancilla register, let Jx=∣0A⟩⊗x, and put Π=JJ†, R0=I−2Π. Suppose W is unitary, J†WJ=M/2, V:S→S is unitary, and ∥M−V∥≤η≤1/100. For C:=−WR0W†R0W one has the clean-input isometry bound ∥CJ−JV∥≤10η. The estimate includes residual ancilla leakage; it does not assert exact cleanup for nonzero η.

Facts & Assumptions

Given: The finite system and ancilla spaces, the stated block condition, unitaries W,V, and 0≤η≤1/100.

[L1]

One-step oblivious amplitude amplification fixes the chronological circuit −WR0W†R0W.

[F1]

Oblivious-amplification block identity: J†WR0W†R0WJ=−3M/2+MM†M/2 when J†WJ=M/2.

[F2]

Oblivious ancilla reflection: R0=I−2Π for the projector Π onto the entire clean coefficient/work ancilla, tensored with identity on the system.

[F3]
[F5]

The operator norm is zero on the zero domain and otherwise is max_{||v||=1} ||Tv||: the operator norm is the maximum norm on unit vectors.

[F7]

The operator norm is submultiplicative and satisfies ||T^*T|| = ||T||^2: ∥AB∥≤∥A∥∥B∥ and ∥T†T∥=∥T∥2.

[F8]

Ordered tensor registers, their inner product, and wire permutations and Tensor products of finitely many linear maps: product registers have the product inner product and T⊗I acts factorwise.

[F9]

Every finite-dimensional real or complex inner product space has an orthonormal basis: every finite-dimensional inner-product space has an orthonormal basis.

[F10]

Finite registers and their ordered computational bases and The computational basis is orthonormal: the distinguished clean register vector is a unit vector. Together with [F8], the clean injection Jx=∣0A⟩⊗x is an isometry; Finite-dimensional linear isometries satisfy the adjoint identity gives J†J=I, and An isometric quantum encoding gives an orthogonal code projector makes Π=JJ† the orthogonal projection onto its image.

Proof

Given: As in Facts & Assumptions.

1.1givenL1F1F2F3F4F5F6F10algebra

By [F10], J†J=I and Π is an orthogonal projection. By the circuit order in [L1], the factors in C are unitary by [F2]–[F4], so CJ is an isometry. Put D:=J†CJ. By [F1] and the minus sign in C, D=3M/2−MM†M/2. Write E=M−V. For every unit vector x, ∥Mx∥≤∥Vx∥+∥Ex∥≤1+η, so ∥M∥≤1+η by [F5]. For each endomorphism T of the nonzero space S, Cauchy–Schwarz gives ∥Tx∥=max⁡∥y∥=1∣⟨Tx,y⟩∣ (with Tx=0 immediate), hence ∥T∥=max⁡∥x∥=∥y∥=1∣⟨Tx,y⟩∣. The adjoint identity interchanges the unit vectors, so ∥T†∥=∥T∥; in particular ∥M†∥≤1+η and ∥E†∥=∥E∥≤η.

2.1givenF4F5F6F7step 1.1algebra

Since VV†V=V, MM†M−V=(M−V)M†M+V(M†−V†)M+VV†(M−V). For every unit vector, apply the vector triangle inequality [F6] and [F7] to these three terms; using ∥E†∥=∥E∥≤η and ∥V∥=∥V†∥=1 from [F4] gives ∥MM†M−V∥≤η((1+η)2+(1+η)+1). Maximizing over unit vectors by [F5], ∥D−V∥≤32η+12η((1+η)2+(1+η)+1)≤4η for η≤1/100.

2.2givenF2F3F4F7F10step 1.1algebra

Let L=(I−Π)CJ. Since Π=JJ† and J†J=I by [F10], CJ=JD+L and J†L=J†(I−JJ†)CJ=J†CJ−(J†J)J†CJ=0. Hence both cross terms vanish in (CJ)†CJ=(JD+L)†(JD+L), and D†D+L†L=(CJ)†CJ=I because C and J are isometries. Thus L†L=I−D†D and ∥L∥2=∥I−D†D∥ by [F7].

3.1givenF4F5F6F7step 2.1step 2.2algebra

Set δ=∥D−V∥≤4η by step 2.1. For every unit vector x, ∥Dx∥≤∥Vx∥+∥(D−V)x∥≤1+δ, so ∥D∥≤1+δ. Since V†V=I, I−D†D=(V−D)†D+V†(V−D); the vector triangle inequality [F6] and [F7] therefore give ∥(I−D†D)x∥≤δ∥D∥+δ≤δ(2+δ) for unit x. Maximizing by [F5], ∥I−D†D∥≤δ(2+δ)≤4η(2+4η)≤9η. Thus ∥L∥≤3η.

4.1F4F5F6F10step 2.1step 3.1algebra

From CJ−JV=J(D−V)+L, for every unit x the vector triangle inequality [F6] and the isometry of J in [F10] give ∥(CJ−JV)x∥≤∥D−V∥+∥L∥≤4η+3η. Maximizing by [F5] and using η≤1/100 yields ∥CJ−JV∥≤10η.

5.1F5F8F9F10step 4.1algebra∎

For any finite reference space R, choose an orthonormal basis (ej) by [F9] and write ξ=∑jxj⊗ej. By [F8], ∥(CJ−JV)⊗IR ξ∥2=∑j∥(CJ−JV)xj∥2≤(10η)2∑j∥xj∥2=(10η)2∥ξ∥2. Thus the same clean-input isometry bound holds with any finite reference, including entangled inputs.

Depends on

Used by

Dependency tree · two levels

56 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