Y-Shaped Generative Flows

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Asadulaev, Arip, Semenov, Semyon, Shtanchaev, Abduragim, Moulines, Eric, Karray, Fakhri, Takac, Martin
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918322363170816
author Asadulaev, Arip
Semenov, Semyon
Shtanchaev, Abduragim
Moulines, Eric
Karray, Fakhri
Takac, Martin
author_facet Asadulaev, Arip
Semenov, Semyon
Shtanchaev, Abduragim
Moulines, Eric
Karray, Fakhri
Takac, Martin
contents Modern continuous-time generative models typically induce \emph{V-shaped} flows: each sample travels independently along a nearly straight trajectory from the prior to the data. Although effective, this independent movement overlooks the hierarchical structures that exist in real-world data. To address this, we introduce \emph{Y-shaped generative flows}, a framework in which samples travel together along shared pathways before branching off to target-specific endpoints. Our formulation is theoretically justified, yet remains practical, requiring only minimal modifications to standard velocity-driven models. We implement this through a scalable, neural network-based training objective. Experiments on synthetic, image, and biological datasets demonstrate that our method recovers hierarchy-aware structures, improves distributional metrics over strong flow-based baselines, and reaches targets in fewer steps.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11955
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Y-Shaped Generative Flows
Asadulaev, Arip
Semenov, Semyon
Shtanchaev, Abduragim
Moulines, Eric
Karray, Fakhri
Takac, Martin
Machine Learning
Artificial Intelligence
Modern continuous-time generative models typically induce \emph{V-shaped} flows: each sample travels independently along a nearly straight trajectory from the prior to the data. Although effective, this independent movement overlooks the hierarchical structures that exist in real-world data. To address this, we introduce \emph{Y-shaped generative flows}, a framework in which samples travel together along shared pathways before branching off to target-specific endpoints. Our formulation is theoretically justified, yet remains practical, requiring only minimal modifications to standard velocity-driven models. We implement this through a scalable, neural network-based training objective. Experiments on synthetic, image, and biological datasets demonstrate that our method recovers hierarchy-aware structures, improves distributional metrics over strong flow-based baselines, and reaches targets in fewer steps.
title Y-Shaped Generative Flows
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2510.11955