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.

Hellinger entropy for compact-location normal mixtures in a scale band

Statement

For A≥1, τ>0, and 0<ϵ<1/4, let PA,τ:={pF,σ:F([−A,A])=1, τ≤σ≤2τ}. There is a universal constant C such that log⁡N(ϵ,PA,τ,h)≤C(A/τ∨1)[log⁡ ⁣(C(A/τ∨1)ϵ)]2. The same bound, with a changed universal constant, holds for PA,τ,ρ:={pF,σ:F([−A,A]c)≤ρ, τ≤σ≤2τ} whenever 0≤ρ≤cϵ2 for a sufficiently small universal c>0. Here N is the covering number in the stated metric.

Facts & Assumptions

Given: A≥1, τ>0, 0<ϵ<1/4, AC, and the mixture density of The Dirichlet normal-mixture density is jointly measurable and normalized.

[A1]

Assume AC for the normal-density and finite-support approximation interfaces. The Axiom of Choice

[F1]

Compactly supported Gaussian mixtures admit logarithmic-size finite approximations: a mixture with mixing law on [−A,A] is within L1 error η of a discrete mixture with at most C(A/σ∨1)log⁡(e/η) atoms.

[F2]

Hellinger distance is a metric and controls L1 and total variation: for probability densities, Hellinger distance is a metric and h2(p,q)≤∫∣p−q∣dλ.

[F3]

A positive-variance normal law has its location-scale density: ϕσ(x−z)=σ−1ϕ((x−z)/σ), with ϕ the standard normal density.

[F4]

Root-L2 Hellinger distance between density laws: the Hellinger metric used here is the root-L2 distance between square-root densities.

[F5]

For a standard normal random variable Z, E[Z2]=1. Location-scale normal laws have their stated mean and variance

[F6]
[F7]

A differentiable function with integrable derivative on a compact interval has endpoint difference equal to the integral of its derivative. The second fundamental theorem: if G is differentiable on [a,b] with G′=f and f is integrable, then ∫abf=G(b)−G(a)

Proof

Given: The hypotheses in the Statement.

1.1A1F1givenchoose

Put K:=⌈C0(A/τ∨1)log⁡(e/ϵ)⌉. By [F1], every pF,σ∈PA,τ is within L1 distance ϵ2/C1 of a mixture with at most K atoms, for a fixed sufficiently large C1. Pad shorter mixtures by zero-weight atoms.

2.1A1F2F3F4F5F6F7step 1.1givenconstruct

Differentiating the density in [F3] gives ∥∂zϕσ(⋅−z)∥1=σ−1E∣Z∣≤σ−1, since 2∣Z∣≤1+Z2 and EZ2=1 by [F5], and ∥∂σϕσ(⋅−z)∥1=σ−1E∣Z2−1∣≤2σ−1 for Z∼N(0,1). Integrating these derivatives along the location and scale segments by [F7] yields ∥ϕσ(⋅−z)−ϕσ(⋅−z′)∥1≤∣z−z′∣/σ and ∥ϕσ−ϕσ′∥1≤2∣σ−σ′∣/(σ∧σ′). For the first class use a scale grid in [τ,2τ]; for an actual band, instead grid [τ,min⁡{2τ,sn}], omitting 2τ when that endpoint is excluded. In either case choose grid mesh at most c2τϵ2, with every grid point admitted by the corresponding scale constraint. Use a location grid in [−A,A] with mesh at most c2τϵ2 and an L1 grid on the K-simplex of mesh at most c2ϵ2. Replacing the scale, locations and weights changes the mixture density in L1 by at most Cc2ϵ2: integrate the componentwise location and scale bounds against the finite mixing weights, and use Tonelli [F6] for the componentwise integrals. Choose c2 so this plus the approximation from step 1.1 is at most ϵ2/4. Thus the resulting finite set, with centers in the class being covered, is an L1 cover of radius ϵ2/4, hence an h-cover of radius ϵ/2 by [F2].

3.1step 2.1algebra

The scale grid has at most Cϵ−2 points and the location grid at most C(1+A/(τϵ2)) points. Rounding each of the first K−1 weights down to a mesh proportional to ϵ2/K leaves the last weight nonnegative after setting it to make the sum one, gives at most (CK/ϵ2)K weight vectors, and changes their L1 distance by at most 2c2ϵ2. Therefore the logarithm of the cover size is at most Clog⁡(e/ϵ)+K[log⁡ ⁣(C(A/τ∨1)ϵ2)+log⁡ ⁣(CKϵ2)]≤C(A/τ∨1)[log⁡ ⁣(C(A/τ∨1)ϵ)]2. The constructed finite set is an h-net of radius ϵ/2, hence also a net of radius ϵ; its cardinality therefore bounds the covering number and proves the first bound.

4.1A1F2F4F6step 2.1step 3.1givenalgebra∎

If r:=F([−A,A]c)=0, then F already belongs to the compact-location class. If 0<r≤ρ<1, let FA and FAc be the normalized restrictions to [−A,A] and its complement. Then pF,σ=(1−r)pFA,σ+rpFAc,σ, so ∥pF,σ−pFA,σ∥1≤2r by integrating the mixture densities (Tonelli [F6]). By [F2], h(pF,σ,pFA,σ)≤2r; the triangle inequality in [F4] then transfers the cover to the tail-constrained class. The scale grid is restricted as in step 2.1, so every center has F([−A,A]c)=0 and satisfies the band and cutoff constraints. For sufficiently small universal c, ρ≤cϵ2 makes the truncation error at most ϵ/2, so the radius-ϵ/2 cover from steps 2.1 and 3.1 becomes a radius-ϵ cover after the constant rescaling used in the entropy bound.

Depends on

Used by

Dependency tree · two levels

57 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