No-Regret Generative Modeling via Parabolic Monge-Ampère PDE

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Deb, Nabarun, Liang, Tengyuan
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910091844780032
author Deb, Nabarun
Liang, Tengyuan
author_facet Deb, Nabarun
Liang, Tengyuan
contents We introduce a novel generative modeling framework based on a discretized parabolic Monge-Ampère PDE, which emerges as a continuous limit of the Sinkhorn algorithm commonly used in optimal transport. Our method performs iterative refinement in the space of Brenier maps using a mirror gradient descent step. We establish theoretical guarantees for generative modeling through the lens of no-regret analysis, demonstrating that the iterates converge to the optimal Brenier map under a variety of step-size schedules. As a technical contribution, we derive a new Evolution Variational Inequality tailored to the parabolic Monge-Ampère PDE, connecting geometry, transportation cost, and regret. Our framework accommodates non-log-concave target distributions, constructs an optimal sampling process via the Brenier map, and integrates favorable learning techniques from generative adversarial networks and score-based diffusion models. As direct applications, we illustrate how our theory paves new pathways for generative modeling and variational inference.
format Preprint
id arxiv_https___arxiv_org_abs_2504_09279
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle No-Regret Generative Modeling via Parabolic Monge-Ampère PDE
Deb, Nabarun
Liang, Tengyuan
Machine Learning
Optimization and Control
Statistics Theory
49Q22, 49N99, 65K10, 68Q32
We introduce a novel generative modeling framework based on a discretized parabolic Monge-Ampère PDE, which emerges as a continuous limit of the Sinkhorn algorithm commonly used in optimal transport. Our method performs iterative refinement in the space of Brenier maps using a mirror gradient descent step. We establish theoretical guarantees for generative modeling through the lens of no-regret analysis, demonstrating that the iterates converge to the optimal Brenier map under a variety of step-size schedules. As a technical contribution, we derive a new Evolution Variational Inequality tailored to the parabolic Monge-Ampère PDE, connecting geometry, transportation cost, and regret. Our framework accommodates non-log-concave target distributions, constructs an optimal sampling process via the Brenier map, and integrates favorable learning techniques from generative adversarial networks and score-based diffusion models. As direct applications, we illustrate how our theory paves new pathways for generative modeling and variational inference.
title No-Regret Generative Modeling via Parabolic Monge-Ampère PDE
topic Machine Learning
Optimization and Control
Statistics Theory
49Q22, 49N99, 65K10, 68Q32
url https://arxiv.org/abs/2504.09279