Alphabeta Math
TheoremStatement: 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.

The observed density is the latent marginal under Tonelli

Statement

Assume the hypotheses of Latent-variable joint and observed-data factorization. Then m(θ,x)=∫p(θ,x,z) μ(dz) is jointly measurable in (θ,x). For each fixed θ, it is an observed-data probability density: it is A-measurable and ∫Xm(θ,x) ν(dx)=1. Consequently m(θ,x)<∞ for ν-almost every x. On every valid fibre 0<m(θ,x)<∞, q(z∣θ,x)=p(θ,x,z)/m(θ,x) is a measurable μ-density of total mass one. Finally, if h:Θ×X×Z→C is product measurable and ∫∣h(θ,x,z)∣p(θ,x,z) d(ν⊗μ)<∞ for a fixed θ, Fubini permits exchanging the latent and observed integrals of hp. No exchange is claimed without this absolute integrability.

Facts & Assumptions

Given: The parameter and data spaces, sigma-finite measures ν,μ, jointly measurable normalized p, and, for the last clause, a product measurable complex-valued h with ∫∣h∣p<∞.

[F1]

Latent-variable joint and observed-data factorization: p is jointly measurable, each fixed-θ joint density integrates to one, m(θ,x)=∫p(θ,x,z) μ(dz), and q=p/m is defined only when 0<m<∞.

[F2]

Measure kernel and probability kernel: a uniformly sigma-finite kernel comes with a specified exhaustion (Zn) for which each kernel mass on Zn is finite.

[F3]

Measurability of integration against a kernel: integrating a nonnegative product-measurable function against a finite or uniformly sigma-finite kernel gives a measurable function of the parameter.

[F4]

Tonelli's theorem for nonnegative measurable functions on a sigma-finite product: for nonnegative product-measurable p, its section integrals are measurable and the iterated and product integrals agree.

[F5]

Fubini's theorem for L^1 functions on a sigma-finite product: for an L1(ν⊗μ) function, almost every section is integrable and the iterated integrals equal the product integral.

[F6]

Monotonicity and nonnegative homogeneity of the nonnegative integral: pointwise order and multiplication by a positive finite scalar pass to nonnegative integrals.

[F7]

Every complete ordered field is Archimedean: the positive integers are unbounded in R.

[F8]

The Borel sigma-algebra on the extended real line: the extended-real Borel structure contains [0,∞] and its Borel subsets.

[F9]

Threshold characterisations of real-valued and extended-real-valued measurability: an extended-real function is measurable if all strict upper threshold preimages are measurable.

[F10]

Composition with a Borel measurable outer map preserves measurability: composition of a measurable map with a Borel measurable outer map is measurable.

[F11]

Arithmetic and lattice operations preserve measurability whenever they are defined: products of extended-real measurable functions are measurable under its stated zero-times-infinity convention.

[F12]

The product sigma-algebra and its finite iterates: the finite product sigma-algebra is generated by measurable rectangles.

[F13]

Finite, sigma-finite, and semifinite measures: sigma-finiteness provides a countable measurable exhaustion by sets of finite measure.

Proof

1.1F10F12given

Fix θ and let Dθ be the collection of subsets E⊆Θ×X×Z whose section Eθ:={(x,z):(θ,x,z)∈E} belongs to A⊗C. Sections commute with complements and countable unions, so Dθ is a sigma-algebra. It contains each measurable rectangle T0×A0×C0, whose section is A0×C0 if θ∈T0 and empty otherwise. By [F12], every product-measurable set has a measurable section. Applying this to preimages of Borel sets shows that pθ(x,z):=p(θ,x,z) is product-measurable. For complex h, first compose with the continuous Borel coordinate maps Re⁡,Im⁡:C→R; the same section argument makes both Re⁡hθ and Im⁡hθ measurable, hence hθ is complex-measurable.

1.2F2F13givenchoose

By [F13], choose a measurable exhaustion Zn↑Z with μ(Zn)<∞. On S:=Θ×X, the constant assignment K(s,B):=μ(B) is a measure kernel: each section is μ and each evaluation map is constant. The same exhaustion makes it uniformly sigma-finite by [F2].

2.1F1F4F6F7givenstep 1.1

For fixed θ, step 1.1 makes pθ product-measurable, so Tonelli [F4] gives ∫Xm(θ,x) ν(dx)=∫X×Zpθ(x,z) d(ν⊗μ)=1 by [F1]. Thus m(θ,⋅) is a probability density. Also Eθ:={(x,z):pθ(x,z)=∞} is measurable. For every positive integer n, n1Eθ≤pθ, so [F6] gives n(ν⊗μ)(Eθ)≤1. Since this holds for every n, [F7] implies (ν⊗μ)(Eθ)=0.

2.2F1F3F12step 1.2

The function p is nonnegative and product-measurable by [F1]. Under the finite-product identification S×Z=(Θ×X)×Z from [F12], apply [F3] to the constant kernel K from step 1.2. It follows that s↦∫p(s,z) μ(dz)=m(s) is measurable on S.

3.1F5F11givenstep 1.1step 2.1

For fixed θ, write uθ=Re⁡hθ and vθ=Im⁡hθ. Define the finite-valued measurable function p~θ:=pθ1{pθ<∞} and then Hθ:=uθp~θ+ivθp~θ. By [F11] this is complex-valued product-measurable; it is defined to be zero on the null set where pθ=∞. It agrees with hθpθ off that null set, and the assumed absolute integrability gives ∫∣Hθ∣ d(ν⊗μ)<∞. Fubini [F5] applies to this L1 representative and gives ∫X×ZHθ d(ν⊗μ)=∫X(∫ZHθ(x,z) μ(dz))ν(dx)=∫Z(∫XHθ(x,z) ν(dx))μ(dz), with its zero assignment on exceptional null sections. Since Hθ=hp almost everywhere on the fixed-θ product space, these are the asserted exchanges for hp.

3.2F1F6F7F8F9F10F11F12step 2.2step 2.1algebra

Let E:={x:m(θ,x)=∞}, measurable by step 2.2. For each positive integer n, n1E≤m(θ,⋅), so [F6] and step 2.1 give nν(E)≤1. Since this holds for every n, [F7] forces ν(E)=0. If 0<m(θ,x)<∞, division by that scalar and the definition of m give ∫q(z∣θ,x) μ(dz)=m(θ,x)/m(θ,x)=1. The valid-fibre set V=m−1((0,∞)) is measurable by step 2.2 and [F8]. For measurability, define ρ:[0,∞]→[0,∞] by ρ(t)=1/t for finite t>0 and ρ(0)=ρ(+∞)=0. For a<0, {ρ>a}=[0,∞]; for a=0 it is (0,∞); and for a>0 it is (0,1/a). These sets are Borel by [F8], so [F9] makes ρ Borel. By [F10], ρ∘m is measurable on Θ×X. Pulling it back along the measurable coordinate projection from the finite product space, which is measurable by [F12], gives a product-measurable function; multiplying by p preserves measurability by [F11]. This product restricts to q=p/m on V×Z, so q is measurable on the valid-fibre set.

4.1step 2.2step 2.1step 3.2step 3.1∎

Steps 2.2, 2.1, 3.2, and 3.1 prove the marginal measurability, normalization, valid-fibre density, and integrable-exchange clauses.

Depends on

Used by

Cited to discharge well-definedness by Latent-variable joint and observed-data factorization.

Dependency tree · two levels

52 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