arxiv: 2604.10274 · v1 · submitted 2026-04-11 · 🧮 math.FA

Recognition: 2 theorem links

· Lean Theorem

Universal Closest Refinement on Measurable Bipartite Relations

T-H. Hubert Chan

Authors on Pith no claims yet

Pith reviewed 2026-05-10 15:44 UTC · model grok-4.3

classification 🧮 math.FA

keywords universal closest refinementmeasurable bipartite relationslevel-optimal maximinconvex divergence functionalsdata-processing inequalitystandard Borel spacesrefinement plansproportional response

0 comments

The pith

Level-optimal maximin refinements paired with proportional response are exactly the universally closest pairs for every divergence obeying the data-processing inequality on measurable bipartite relations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper asks whether the nonunique set of finite refinement plans concentrated on a measurable bipartite relation with one fixed marginal contains a mathematically distinguished subclass. It defines level-optimal maximin via truncation and overflow profiles as the one-sided extremal criterion that selects canonical refinements. The central theorem shows that every such level-optimal maximin plan, once paired with its proportional-response partner on the opposite side, is closest under every proper lower semicontinuous convex divergence that satisfies the data-processing inequality under measurable post-processing. The converse holds as well: any universally closest pair must arise in this way. The result therefore supplies a canonical selection rule that works uniformly across an entire family of divergences.

Core claim

Every universally closest refinement pair consists of a level-optimal maximin refinement on one side and its proportional-response partner on the other, and such pairs are exactly the minimizers of every proper lower semicontinuous convex functional satisfying the data-processing inequality under measurable post-processing. Conversely, every pair that is universally closest among all feasible pairs has this structure. The same pairs also admit an equilibrium characterization in an associated continuum economy whose commodities are measures.

What carries the argument

Level-optimal maximin, the levelwise maximin principle defined through truncation and overflow profiles that selects the distinguished one-sided refinements.

If this is right

Every minimizer of a strictly convex refinement criterion must be level-optimal maximin.
Every level-optimal maximin refinement minimizes the entire class of relevant convex divergence functionals.
The structure admits a converse paired characterization when the closest-pair criterion is strictly convex.
Level-optimal maximin pairs receive an equilibrium-theoretic description as solutions to a continuum economy with measure-valued commodities.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The measurable max-flow min-cut duality and symmetric density decomposition developed here may simplify numerical computation of such pairs on concrete relations.
The uniform selection rule could be used to define canonical couplings or channels in settings where multiple refinements are possible.
Extensions to non-Borel spaces would require checking whether the same measurable selection and disintegration arguments continue to hold.

Load-bearing premise

The underlying spaces are standard Borel spaces, the bipartite relation is measurable, and the divergences are proper lower semicontinuous convex functionals that satisfy the data-processing inequality under measurable post-processing.

What would settle it

A concrete measurable bipartite relation together with a divergence obeying the data-processing inequality for which either a level-optimal maximin pair fails to be closest or a closest pair fails to be level-optimal maximin with proportional response.

read the original 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 identifies the opposite-side partner and yields a universally closest refinement pair. Our main theorem shows that every such pair is universally closest among all feasible pairs for every divergence satisfying the data-processing inequality under measurable post-processing, and conversely that every universally closest pair has this structure. We also prove a converse paired characterization for strictly convex closest pairs. Finally, we give an equilibrium-theoretic characterization of level-optimal maximin pairs through a naturally associated continuum economy with measure-valued commodities. The measure-theoretic theory requires new tools beyond the finite case, including disintegration, measurable selection, measurable max-flow/min-cut duality, augmentation arguments, and a symmetric density decomposition separating absolutely continuous and singular components.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives a bidirectional characterization of level-optimal maximin pairs as the universally closest refinements for DPI-satisfying divergences in the measurable bipartite setting.

read the letter

The central result is that level-optimal maximin refinements are exactly the ones that minimize every proper lsc convex divergence obeying the data-processing inequality under measurable post-processing, with a converse that any such universally closest pair must have this structure. This is paired with an equilibrium characterization via a continuum economy with measure-valued commodities. The work extends the finite case by building the necessary tools: disintegration, measurable selection, a measurable max-flow/min-cut duality, augmentation arguments, and symmetric density decomposition to separate absolutely continuous and singular parts. These let the authors prove the equivalences without circularity and handle the non-uniqueness of feasible plans on standard Borel spaces. The levelwise truncation and overflow profiles give a concrete way to identify the distinguished pairs, and the proportional response step pairs them across sides. The assumptions are standard—measurable relations, finite plans with one fixed marginal, proper convex divergences—but they are stated clearly and the proofs appear to rest on independent measure-theoretic facts rather than fitted parameters. One minor limitation is that the setting stays within standard Borel spaces, so extensions to more general measurable spaces would need extra work. The abstract and stress-test outline show no internal gaps in the chain of equivalences. This is for researchers in optimal transport or functional analysis who need explicit structure inside non-unique classes of plans. A reader already working with measurable couplings or divergence minimization would find the constructions and the equilibrium view useful. It deserves a serious referee because the technical development is substantial and the claims are falsifiable through the listed tools.

Referee Report

0 major / 2 minor

Summary. The paper studies 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 with one fixed marginal. The central claim is a bidirectional characterization: level-optimal maximin refinements (defined via truncation and overflow profiles) are exactly the universally closest pairs for every proper lower semicontinuous convex divergence satisfying the data-processing inequality under measurable post-processing, with a converse for strictly convex cases. The paper develops supporting tools including disintegration, measurable selection, measurable max-flow/min-cut duality, augmentation arguments, and symmetric density decomposition, and provides an equilibrium characterization via an associated continuum economy with measure-valued commodities.

Significance. If the main theorem holds, the work supplies a divergence-independent structural criterion for selecting canonical refinements in non-unique measurable settings, unifying optimization and extremal properties in a way that extends finite-case results. The new measure-theoretic machinery (measurable max-flow/min-cut duality and symmetric density decomposition) is a standalone contribution with potential uses in optimal transport, information theory, and continuum economies. The bidirectional nature of the characterization and the equilibrium interpretation add robustness and falsifiability.

minor comments (2)

[Abstract] The abstract packs the main theorem and its converses into a single dense sentence; splitting the statement of the bidirectional characterization would improve immediate readability.
[Introduction] Notation for truncation and overflow profiles is introduced without an early summary table or diagram; a small illustrative example in the introduction would clarify the levelwise maximin principle before the technical development.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the detailed summary, positive assessment of significance, and recommendation of minor revision. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper establishes bidirectional characterizations between level-optimal maximin refinements and universally closest pairs for convex divergences satisfying the DPI, using newly developed independent tools including disintegration, measurable selection, measurable max-flow/min-cut duality, augmentation arguments, and symmetric density decomposition. These tools enable proofs of equivalences and converses without reducing to self-definitions or fitted parameters. The equilibrium characterization via continuum economy supplies additional independent support. No load-bearing steps equate to inputs by construction or rely on unverified self-citation chains.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The development rests on standard properties of Borel spaces and convex divergences; no free parameters are introduced, and no new entities are postulated.

axioms (2)

standard math Standard Borel spaces admit disintegrations, measurable selections, and measurable max-flow/min-cut duality
Invoked repeatedly to construct and compare measurable refinement plans.
domain assumption Divergences are proper lower semicontinuous convex functionals satisfying the data-processing inequality under measurable post-processing
Required for the definition of universal closeness and the main equivalence.

pith-pipeline@v0.9.0 · 5538 in / 1318 out tokens · 34572 ms · 2026-05-10T15:44:15.010376+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

every level-optimal maximin refinement minimizes the full class of relevant proper lower semicontinuous convex divergence functionals... universally closest among all feasible pairs for every divergence satisfying the data-processing inequality
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean costAlphaLog_high_calibrated_iff unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

level-optimal maximin... truncation and overflow profiles... hockey-stick divergences HS_γ

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

2 extracted references · 1 canonical work pages

[1]

[Aum64] Robert J. Aumann. Markets with a continuum of traders.Econometrica, 32(1/2):39–50, 1964. [Aum66] Robert J. Aumann. Existence of competitive equilibria in markets with a continuum of traders.Econometrica, 34(1):1–17, 1966. [Bla51] David Blackwell. Comparison of experiments. In Jerzy Neyman, editor,Proceedings of the Second Berkeley Symposium on Mat...

1964
[2]

[Kal21] Olav Kallenberg.Foundations of Modern Probability, volume 99 ofProbability Theory and Stochastic Modelling

Press, Princeton, NJ, 1974. [Kal21] Olav Kallenberg.Foundations of Modern Probability, volume 99 ofProbability Theory and Stochastic Modelling. Springer, Cham, 3 edition, 2021. [Kel84] Hans G. Kellerer. Duality theorems for marginal problems.Zeitschrift f¨ ur Wahrschein- lichkeitstheorie und Verwandte Gebiete, 67(4):399–432, 1984. [KRN65] Kazimierz Kurato...

work page doi:10.1007/bf01213846 1974