arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.11547cs.LGcs.AI
收藏

提升流:用于流匹配的锐度感知采样

Sharpen Your Flow: Sharpness-Aware Sampling for Flow Matching

Lawrence Berkeley National Laboratory(伯克利国家实验室);International Computer Science Institute(国际计算机科学研究所);Department of Mathematics, KTH Royal Institute of Technology(皇家理工学院数学系);Nordita, KTH Royal Institute of Technology and Stockholm University(KTH皇家理工学院与斯德哥尔摩大学联合研究所);Center for Applied Mathematics, Cornell University(康奈尔大学应用数学中心)

Aditi Gupta, Soon Hoe Lim, Annan Yu, N. Benjamin Erichson

AI总结:

本文提出SharpEuler,一种无需训练的采样器,通过预训练模型离线分析校准轨迹上学习速度场变化最快的位置,生成时间步网格以优化生成效率,提升样本质量。

摘要:

流匹配模型通过数值积分学习的速度场生成样本,每一步积分都需要神经网络评估。快速生成需要有效利用小固定评估预算:关键问题不仅在于如何积分,还在于采样器应在哪里分配步骤。我们提出SharpEuler,一种免训练的采样器,通过离线分析预训练模型,估计校准轨迹上学习速度场变化最快的位置,生成时间步网格以适应任意的推理预算。在测试时,采样仍保持普通的欧拉积分,与均匀调度相同数量的模型评估。我们通过三个原则证明SharpEuler:数值原则识别轨迹加速为欧拉离散化误差的主要来源,变分原则推导基于锐度的幂律时间步密度,统计保证显示有限样本校准采样器在终端分布水平上是稳定的。实验表明,SharpEuler在固定预算下提升样本质量,减少跨模式泄漏并增加模式覆盖。

英文摘要:

Flow matching models generate samples by numerically integrating a learned velocity field, with each integration step requiring a neural network evaluation. Fast generation therefore requires using a small fixed evaluation budget effectively: the key question is not only how to integrate the flow, but where the sampler should spend its steps. We propose SharpEuler, a training-free sampler that profiles a pretrained model offline by estimating where the learned velocity field changes most rapidly along calibration trajectories. This finite-difference estimate defines a solver-aware sharpness profile, which is smoothed and converted by a quantile transform into a timestep grid for any desired inference budget. At test time, sampling remains ordinary Euler integration with the same number of model evaluations as a uniform schedule. We justify SharpEuler using three principles: a numerical principle identifying trajectory acceleration as the leading source of Euler discretization error, a variational principle deriving sharpness-based power-law timestep densities, and a statistical guarantee showing that the finite-sample calibrated sampler is stable at the terminal distribution level. Our experiments show that SharpEuler improves sample quality at fixed budgets, reducing inter-mode leakage and increasing mode coverage.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑