OctoInformBeta

OpenAI preprint, September 2026 · Appendix A. The fixed-point ingredients

Appendix A (setting): finite probability simplices: Nonnegative real coordinates with total mass one

Probability weights on finite index sets.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Let II be a finite index set, and write ℝI\mathbb R^I for the space of real coordinate families a=(ai)i∈Ia=(a_i)_{i\in I}.

The Definition

Definition 1 (Finite probability simplex). The probability weight vectors on II form the set W(I)={a∈ℝI:ai≥0 for every i∈I,∑i∈Iai=1}.W(I)=\left\{a\in\mathbb R^I : a_i\geq 0\text{ for every }i\in I, \quad \sum_{i\in I}a_i=1\right\}.(1)✦ The probability simplex Δ(I)\Delta(I) on the finite index set II is the set of probability weight vectors given by Δ(I):=W(I).\Delta(I):=W(I).(2)✦

In the paper, the result reads: for every integer m≥0m\geq 0, the finite-dimensional probability simplex is Δm={a∈ℝm+1:ai≥0(0≤i≤m),∑i=0mai=1}\Delta_m=\{a\in\mathbb R^{m+1}:a_i\geq 0\ (0\leq i\leq m), \sum_{i=0}^{m}a_i=1\}.

What Is Checked

The definitions are stated over arbitrary finite index sets and real-valued coordinate families.