Research

Paper

TESTING February 24, 2026

A Long-Short Flow-Map Perspective for Drifting Models

Authors

Zhiqi Li, Bo Zhu

Abstract

This paper provides a reinterpretation of the Drifting Model~\cite{deng2026generative} through a semigroup-consistent long-short flow-map factorization. We show that a global transport process can be decomposed into a long-horizon flow map followed by a short-time terminal flow map admitting a closed-form optimal velocity representation, and that taking the terminal interval length to zero recovers exactly the drifting field together with a conservative impulse term required for flow-map consistency. Based on this perspective, we propose a new likelihood learning formulation that aligns the long-short flow-map decomposition with density evolution under transport. We validate the framework through both theoretical analysis and empirical evaluations on benchmark tests, and further provide a theoretical interpretation of the feature-space optimization while highlighting several open problems for future study.

Metadata

arXiv ID: 2602.20463
Provider: ARXIV
Primary Category: cs.LG
Published: 2026-02-24
Fetched: 2026-02-25 06:05

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.20463v1</id>\n    <title>A Long-Short Flow-Map Perspective for Drifting Models</title>\n    <updated>2026-02-24T01:48:52Z</updated>\n    <link href='https://arxiv.org/abs/2602.20463v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.20463v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>This paper provides a reinterpretation of the Drifting Model~\\cite{deng2026generative} through a semigroup-consistent long-short flow-map factorization. We show that a global transport process can be decomposed into a long-horizon flow map followed by a short-time terminal flow map admitting a closed-form optimal velocity representation, and that taking the terminal interval length to zero recovers exactly the drifting field together with a conservative impulse term required for flow-map consistency. Based on this perspective, we propose a new likelihood learning formulation that aligns the long-short flow-map decomposition with density evolution under transport. We validate the framework through both theoretical analysis and empirical evaluations on benchmark tests, and further provide a theoretical interpretation of the feature-space optimization while highlighting several open problems for future study.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <published>2026-02-24T01:48:52Z</published>\n    <arxiv:comment>25 pages, 7 figures</arxiv:comment>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Zhiqi Li</name>\n    </author>\n    <author>\n      <name>Bo Zhu</name>\n    </author>\n  </entry>"
}