跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.19587 · v1

Explicit Regularization for Flow Matching

首次提交 2026-11-15 03:32
版本更新 2026-11-15 03:32
雷达观测 2026-11-17 12:06
arXiv ↗PDF ↗

摘要

Learning the underlying probability distribution from samples is a core problem in statistics and applied mathematics. Generative models based on diffusion and flow matching learn a velocity field whose flow transports a known distribution to the data distribution. They produce novel, realistic samples, although the empirical learning problem is ill posed near the data distribution. Its minimizer transports the samples onto the training data, and small perturbations of the data change the velocity field strongly near the end of the flow. The observed generalization is often attributed to implicit regularization by the network parameterization, early stopping of the training, and the time integration during sampling. Explicit regularization has advanced the theory and computation of many inverse problems and has recently been proposed to prevent memorization in generative models. With a regularizer on the magnitude of the velocity, however, the field approaches the memorizing velocity unless the regularization parameter grows to infinity fast enough, and it then stays bounded by a constant that is small only for large parameters. We show that penalizing the derivatives of order kk of the velocity renders the DD-dimensional empirical problem well posed uniformly in time if 2k>D2k > D and the training points are in general position. Our experiments in one and two dimensions discretize the velocity with splines and use direct solvers for the regularized least-squares problems to exclude optimization error. Refining the discretization in space and time changes the error of the Hessian-regularized flows only marginally, whereas without the regularizer the error tends to that of memorization. These flows generalize, with errors comparable to or lower than those of well-tuned kernel density estimates.

作者备注:40 pages (main text pp. 1-24, references pp. 24-26, appendices A-C pp. 27-40), 4 figures, 11 tables

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-15 03:32 · 最新观测版本

Explicit Regularization for Flow Matching

Learning the underlying probability distribution from samples is a core problem in statistics and applied mathematics. Generative models based on diffusion and flow matching learn a velocity field whose flow transports a known distribution to the data distribution. They produce novel, realistic samples, although the empirical learning problem is ill posed near the data distribution. Its minimizer transports the samples onto the training data, and small perturbations of the data change the velocity field strongly near the end of the flow. The observed generalization is often attributed to implicit regularization by the network parameterization, early stopping of the training, and the time integration during sampling. Explicit regularization has advanced the theory and computation of many inverse problems and has recently been proposed to prevent memorization in generative models. With a regularizer on the magnitude of the velocity, however, the field approaches the memorizing velocity unless the regularization parameter grows to infinity fast enough, and it then stays bounded by a constant that is small only for large parameters. We show that penalizing the derivatives of order kk of the velocity renders the DD-dimensional empirical problem well posed uniformly in time if 2k>D2k > D and the training points are in general position. Our experiments in one and two dimensions discretize the velocity with splines and use direct solvers for the regularized least-squares problems to exclude optimization error. Refining the discretization in space and time changes the error of the Hessian-regularized flows only marginally, whereas without the regularizer the error tends to that of memorization. These flows generalize, with errors comparable to or lower than those of well-tuned kernel density estimates.

查看此版本解读 →查看此版本原文 ↗