Saved in:
Bibliographic Details
Main Authors: Wang, Zimo, Mehta, Ishit, Lu, Haolin, Sun, Chung-En, Yan, Ge, Weng, Tsui-Wei, Li, Tzu-Mao
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.02928
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908807938965504
author Wang, Zimo
Mehta, Ishit
Lu, Haolin
Sun, Chung-En
Yan, Ge
Weng, Tsui-Wei
Li, Tzu-Mao
author_facet Wang, Zimo
Mehta, Ishit
Lu, Haolin
Sun, Chung-En
Yan, Ge
Weng, Tsui-Wei
Li, Tzu-Mao
contents Time-unconditional generative models learn time-independent denoising vector fields. But without time conditioning, the same noisy input may correspond to multiple noise levels and different denoising directions, which interferes with the supervision signal. Inspired by distance field modeling, we propose Distance Marching, a new time-unconditional approach with two principled inference methods. Crucially, we design losses that focus on closer targets. This yields denoising directions better directed toward the data manifold. Across architectures, Distance Marching consistently improves FID by 13.5% on CIFAR-10 and ImageNet over recent time-unconditional baselines. For class-conditional ImageNet generation, despite removing time input, Distance Marching surpasses flow matching using our losses and inference methods. It achieves lower FID than flow matching's final performance using 60% of the sampling steps and 13.6% lower FID on average across backbone sizes. Moreover, our distance prediction is also helpful for early stopping during sampling and for OOD detection. We hope distance field modeling can serve as a principled lens for generative modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02928
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Distance Marching for Generative Modeling
Wang, Zimo
Mehta, Ishit
Lu, Haolin
Sun, Chung-En
Yan, Ge
Weng, Tsui-Wei
Li, Tzu-Mao
Machine Learning
Time-unconditional generative models learn time-independent denoising vector fields. But without time conditioning, the same noisy input may correspond to multiple noise levels and different denoising directions, which interferes with the supervision signal. Inspired by distance field modeling, we propose Distance Marching, a new time-unconditional approach with two principled inference methods. Crucially, we design losses that focus on closer targets. This yields denoising directions better directed toward the data manifold. Across architectures, Distance Marching consistently improves FID by 13.5% on CIFAR-10 and ImageNet over recent time-unconditional baselines. For class-conditional ImageNet generation, despite removing time input, Distance Marching surpasses flow matching using our losses and inference methods. It achieves lower FID than flow matching's final performance using 60% of the sampling steps and 13.6% lower FID on average across backbone sizes. Moreover, our distance prediction is also helpful for early stopping during sampling and for OOD detection. We hope distance field modeling can serve as a principled lens for generative modeling.
title Distance Marching for Generative Modeling
topic Machine Learning
url https://arxiv.org/abs/2602.02928