Alphabeta Math
AlgorithmAlgorithm: Literature-sourcedProof: Not applicablePipeline-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.

Exhaustive finite-horizon POMDP alpha-vector backup

Definition

Input a finite POMDP, a horizon H∈N, and its discount γ∈[0,1]. A depth-zero policy tree is the empty continuation and has the zero vector α0:S→R. Recursively, a depth-(h+1) policy tree consists of a root action a∈A and one depth-h continuation tree for each observation o∈Ω. If the continuation vectors are (αo)o∈Ω, assign the tree the vector

αa,(αo)(s)=r(s,a)+γ∑s′∈ST(s′∣s,a)∑o∈ΩO(o∣s′)αo(s′)(s∈S).

Set Γ0={α0}. For each h=0,…,H−1, form Γh+1 from every action a∈A and every tuple (αo)o∈Ω∈ΓhΩ using the displayed formula, treating Γh+1 as a set so equal vectors are stored once. Return (Γ0,…,ΓH). This procedure exhaustively enumerates finite policy-tree vectors; it does not perform linear-program witness pruning and makes no bit-complexity or polynomial-time claim.

Remarks

The diagram shows a depth-two policy tree: the root chooses a0; after the observation x the continuation chooses ax, and after y it chooses ay. The outgoing edges represent observations, not hidden states or rewards. Every observation has a branch, including an observation that has probability zero at a particular belief. The algebraic definition above, rather than the diagram, determines the procedure and its outputs.

diagram failed to render:
\begin{tikzpicture}[level distance=1.25cm,sibling distance=3.0cm]
  \node[draw,rounded corners] {$a_0$}
    child { node[draw,rounded corners] {$a_x$} edge from parent node[left] {$x$} }
    child { node[draw,rounded corners] {$a_y$} edge from parent node[right] {$y$} };
\end{tikzpicture}

Depends on

Used by

Dependency tree · two levels

13 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