Universal Closest Refinement on Measurable Bipartite Relations
T-H. Hubert Chan

TL;DR
This paper introduces a measure-theoretic framework for the universal closest refinement problem on measurable bipartite relations, establishing a unique structure via convex refinement and equilibrium characterization.
Contribution
It develops a measure-theoretic approach with new tools to identify the universally closest refinement pairs in bipartite relations under divergence criteria.
Findings
Level-optimal maximin criterion characterizes the refinement structure.
Convex refinement minimizes relevant divergence functionals.
Universally closest pairs are characterized by divergence satisfying data-processing inequality.
Abstract
We study the universal closest refinement problem on measurable bipartite relations over standard Borel spaces. Given prescribed side measures, the feasible class consists of finite refinement plans concentrated on the relation and carrying one fixed marginal. The main question is whether this highly nonunique class nevertheless contains a mathematically distinguished class of refinements. We show that the correct one-sided extremal criterion is level-optimal maximin, a levelwise maximin principle formulated through truncation and overflow profiles. We then prove that this structure is exactly the one selected by convex refinement: every minimizer of a strictly convex refinement criterion is level-optimal maximin, while every level-optimal maximin refinement minimizes the full class of relevant proper lower semicontinuous convex divergence functionals. Proportional response then…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
