Alphabeta Math
ExampleConstruction: AI-generatedVerification: AI-generatedPipeline-generatedjudge pass (gpt-6-luna)verified 2026-10-22 (gpt-6.1-sol)
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.

A smooth Bessel-blur truth with an explicit TV source certificate

Example

Assume Countable Choice. Let 0<a<b, let ϕ∈Cc∞((a,b)) be nonnegative and not identically zero, define U(r)=∫rbϕ(s) ds for r≥0, and set u†(x)=U(∣x∣) on R2. Choose h∈Cc∞((a/2,2b)) with 0≤h≤1 and h=1 on supp⁡ϕ, put z(x)=h(∣x∣)x/∣x∣ extended by zero at the origin, and let q=div⁡z. On real L2(R2) define the Bessel blur A=F2−1MmF2, where F2 is the unitary Fourier transform with phase e−2πix⋅ξ and Mm multiplies by m(ξ)=(1+4π2∣ξ∣2)−1/2. Let f=Au† and define p=F2−1((1+4π2∣ξ∣2)1/2q^(ξ)). Then A is bounded, self-adjoint, and injective, p∈L2(R2;R), and A∗p∈∂TV2(u†). Thus u† is the unique minimum-TV exact solution satisfying the full-space source condition; for zero noise the parameter ratio ∥A∗∥∥wn∥2/αn is zero.

Facts & Assumptions

Given: The nonzero nonnegative radial profile ϕ, the radial truth u†, the Fourier multiplier m, and full-space TV2.

[A1]

Countable Choice is assumed by the Plancherel supplier. (The Axiom of Countable Choice (ACω))

[F1]

Plancherel theorem: F2 is a unitary map on complex L2(R2) and preserves inner products.

[F2]

Fourier transform on complex L1 classes: for integrable q, q^(ξ)=∫q(x)e−2πix⋅ξ dx; integration by parts gives the displayed derivative transform.

[F3]

Schwartz space is dense in L2: complex Schwartz functions are dense in complex L2; taking real parts gives density of real Schwartz functions in real L2.

[F4]

Distributional TV as an extended functional on L²: TV2(v) is the supremum of ∫vdiv⁡y over compactly supported smooth vector fields with ∥y∥∞≤1.

[F5]

Banach-space convex subgradients and dual pairings and Dual source certificates for TV regularization: q∈∂TV2(u) when TV2(v)≥TV2(u)+⟨q,v−u⟩ for all v∈L2, using the real Hilbert pairing.

[F6]

A smooth bump between concentric Euclidean balls: a smooth cutoff can equal one on the compact radial support of ϕ and vanish near the origin.

[F7]

Real and complex inner-product spaces and their induced length: on real-valued functions, the complex L2 inner product equals the real Hilbert pairing.

Verification

technique · Fourier multiplier verification and a calibration field

The real-valued subspace is closed in complex L2: if gj are real-valued classes and gj→g in complex L2, then ∥Im⁡g∥2=∥Im⁡(g−gj)∥2≤∥g−gj∥2→0, so g is real-valued as an L2 class.

1.1A1F1F2F3F7givenalgebra

The multiplier m is real, even, and satisfies 0<m≤1, so [F1] makes A a contraction on complex L2. For real Schwartz g, the integral formula [F2] gives g^(−ξ)=g^(ξ)‾, and multiplication by m preserves this symmetry. Approximate mg^ in L2 by Schwartz functions using [F3], then replace each approximant s by 12(s(ξ)+s(−ξ)‾). These symmetrized approximants remain Schwartz, converge to mg^, and have real inverse Fourier transforms by [F2]; [F1] and the closedness observation above pass this property to Ag. Density of real Schwartz functions and boundedness of A extend real-valuedness to real L2. By [F7], the complex Plancherel pairing restricts to the real Hilbert pairing. Multiplication by the real function m is self-adjoint, so ⟨Au,v⟩=⟨u,Av⟩ for real u,v, and A∗=A. If Au=0, then mu^=0 almost everywhere; strict positivity of m and injectivity of the unitary Fourier transform give u=0. Thus A is bounded, self-adjoint, and injective.

1.2F4F5F6givenalgebra

Because ϕ is supported away from zero and from b, u† is constant near the origin, smooth across radii a,b, and zero for ∣x∣≥b; hence it is a compactly supported smooth real function. Its gradient is ∇u†(x)=−ϕ(∣x∣)x/∣x∣ wherever ϕ(∣x∣)≠0 and is zero elsewhere. The cutoff in the Statement exists by [F6] and gives z∈Cc∞(R2;R2) with ∥z∥∞≤1. Integration by parts gives ∫R2u†div⁡z dx=−∫R2z⋅∇u† dx=∫R2ϕ(∣x∣) dx=∫R2∣∇u†∣ dx. For every admissible test field y, the same integration by parts gives ∫u†div⁡y≤∫∣∇u†∣, while the field z attains equality. Thus TV2(u†)=∫u†div⁡z. Setting q=div⁡z, [F4] gives TV2(v)≥∫vq=TV2(u†)+⟨q,v−u†⟩ for every v, so q∈∂TV2(u†).

2.1A1F1F2F5step 1.1step 1.2construct

The vector q is smooth and compactly supported, so q,∂1q,∂2q∈L2. Integration by parts using [F2] gives ∂jq^=2πiξjq^, and Plancherel [F1] yields ∫R2(1+4π2∣ξ∣2)∣q^(ξ)∣2 dξ=∥q∥22+∥∇q∥22<∞. Therefore p^(ξ):=(1+4π2∣ξ∣2)1/2q^(ξ) is in L2, defining p=F2−1p^∈L2. The integral formula [F2] gives Hermitian symmetry of q^, and the real even multiplier preserves it. As in step 1.1, approximate p^ by Schwartz functions, symmetrize them, apply inverse Fourier transformation, and use Plancherel continuity and closedness of the real subspace to conclude that p is real. By construction A∗p=Ap=q, so [F5] gives the required source certificate. Injectivity from step 1.1 makes u† the unique exact solution, hence the unique minimum-TV exact solution.

3.1step 1.1step 2.1givenalgebra∎

Step 2.1 provides a certificate at the unique minimum-TV exact truth. For every positive αn and zero noise wn=0, the parameter ratio is ∥A∗∥∥wn∥2/αn=0. Thus this example verifies the source condition and the zero-noise parameter side of the full-space deblurring setup without asserting existence of a regularized minimizer.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

53 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