Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-14
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.

Killed and absorbed kernels are probability kernels

Statement

For every probability kernel K and measurable D, the absorbed kernel KDabs and killed kernel KDkill of the preceding definition are probability kernels on (E,E) and (EΔ,EΔ) respectively.

Facts & Assumptions

Given: A probability kernel K and D∈E.

[F1]

The absorbed and killed candidates, including the cemetery sigma-algebra, are the formulas in Killed and absorbed transition kernels.

[F2]

A probability kernel has probability-measure sections and measurable evaluation functions. (Measure kernel and probability kernel)

Proof

1.1F1F2

Fix x∈E. If x∈D, the absorbed section in [F1] is δx; if [F1, F2] x∈Dc, it is K(x,⋅). Hence every section is countably additive, has empty-set mass zero and total mass one. For fixed A∈E, its evaluation is 1D(x)1A(x)+1Dc(x)K(x,A), a measurable function by [F2]. Thus KDabs is a probability kernel.

1.2F1F2

Fix x∈D. The killed section is the restriction [F1, F2] A↦K(x,A∩D) on E, plus an atom of mass K(x,Dc) at Δ. It is countably additive and its total mass is K(x,D)+K(x,Dc)=1. If x∈Dc∪{Δ}, the section is δΔ. Thus all killed sections are probability measures, including D=∅ and D=E.

2.1F1F2step 1.2∎

Fix B∈EΔ, put A=B∩E, and let [F1, F2, step 1.2] ε=1B(Δ). On E the killed evaluation is 1D(x){K(x,A∩D)+εK(x,Dc)}+1Dc(x)ε, which is E-measurable by [F2]. Its value at the measurable singleton {Δ} is ε, so the full evaluation is EΔ-measurable. Therefore KDkill is a probability kernel. The empty and full target sets give respectively zero and one in every case. No choice principle is used.

Depends on

Used by

Dependency tree · two levels

4 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