arXiv：cs.LG· F\'elix Dedek, Makoto Yamada·· 3 小时前AI 评分33
面向 Flow-Map 蒸馏的几何感知时间重参数化
Geometry-Aware Time Reparameterization for Flow-Map Distillation
AI 导读
研究者提出一种几何感知的时间重参数化方法，通过给法向加速度较大的轨迹段分配更多学生时间，提升 Flow-Map 蒸馏效果。该方法推导出可均衡法向加速度统计量的共享时钟，并在蒸馏前一次性估计，无需重训教师模型或增加学生参数与推理开销。在合成数据、CIFAR-10 和 CelebA-64 上，相同推理预算下样本质量优于恒等时间蒸馏，一步与两步图像生成均有提升。
正文
Abstract:Flow-map distillation enables one- and few-step generation by learning finite-time transitions of a pretrained generative ODE. We investigate whether changing the teacher's time parameterization can make these transitions easier to learn. Motivated by the hypothesis that trajectory segments with large normal acceleration are harder to distill, we propose a geometry-aware time reparameterization that allocates more student time to these regions while preserving the teacher's geometric paths and terminal distribution. We derive a shared clock that equalizes a population normal-acceleration statistic under suitable assumptions, and construct a practical approximation from robust, regularized estimates across teacher trajectories. We incorporate this clock into Lagrangian flow-map distillation, using the transformed time coordinate to condition the student. The clock is estimated once before distillation and requires neither teacher retraining nor additional student parameters or inference-time network evaluations. Experiments on synthetic data, CIFAR-10, and CelebA-64 show improved sample quality over identity-time distillation at matched inference budgets, including improvements in one- and two-step image generation. The gains in one-step generation, where no intermediate sampling times can be adjusted, highlight the benefits of time reparameterization during distillation.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.02427 [cs.LG] |
| (or arXiv:2610.02427v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.02427 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Félix Dedek [view email]
[v1]
Thu, 1 Oct 2026 19:49:43 UTC (939 KB)
来源：arXiv：cs.LG · arxiv.org