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.
Finite relative-entropy chain rule and event grouping
Statement
Let be finite nonempty sets, let be probability vectors on (a probability vector on a finite nonempty set is a function with ), and assume for every . For probability vectors on a finite set with strictly positive, define with zero-mass terms omitted. For , write and . Define the first-coordinate marginals by and . For with , define ; define for every , since strict positivity of gives . These are probability vectors on ; terms weighted by are omitted. Then For every subset , grouping the atoms into and its complement cannot increase relative entropy: with zero-weight terms omitted. For the empty or full event, the binary divergence is , and the grouping bound reduces to , with equality exactly when . For a nonempty proper event , equality holds exactly when, on each of and having positive -mass, the conditional probability vectors of and agree.
Statement references
Facts & Assumptions
Given: Finite nonempty sets , probability vectors on , strict positivity of , and the natural logarithm convention.
Nonnegativity and equality for finite classical relative entropy: for finite probability vectors with positive reference vector, relative entropy as defined above is nonnegative and vanishes exactly when the probability vectors coincide; its proof derives the scalar inequality from the logarithm derivative.
Proof
Write and . Strict positivity of gives and for each . If , then and by [F2]. Multiply by and sum over positive joint atoms. For each with , summing the first term over gives , and summing the second gives . Atoms with contribute no terms. This is exactly the claimed chain rule.
If is empty or all of , then is respectively or , so the binary divergence is under the stated zero-weight convention. The grouping bound is then , with equality exactly when by [F1]. For a nonempty proper , strict positivity gives . Put and . When , the conditional laws on are and ; the latter is strictly positive on . The logarithmic identity for each with yields If , both the sum and the weighted conditional term are zero, so the same identity holds with that term omitted. Applying the identical calculation to and adding gives , with zero-weight conditional terms omitted. Both weighted conditional divergences are nonnegative by [F1], which proves the grouping inequality, including or . Equality holds precisely when every conditional divergence with positive weight vanishes; by the equality clause of [F1], this is equivalent to agreement of the conditional probability vectors on each positive--mass block. Zero--mass blocks impose no conditional-law condition.