面向多模态分布的流匹配
Flow Matching for Multimodal Distributions
把源分布从单一高斯换成数据自适应的混合高斯，并配合模式相关的数据耦合，可让流匹配模型的传输路径更短更直，从而在更少训练轮次和采样步数下取得更优 FID；这一「源-耦合协同设计」思路可直接迁移到图像/视频流匹配生成器的训练加速上。
原文依据 we propose an efficient source and coupling co-design method termed Mixture-Modeling Flow Matching (MM-FM)
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期，视觉基础模型被证明能够通过揭示内在的流形并集结构、降低潜空间/目标分布的复杂度，来提升基于流的生成模型的效率。本文利用流形并集结构的多模态特性，旨在进一步提升流匹配模型的学习与推理效率。为此，我们提出一种高效的源分布与耦合协同设计方法，称为混合建模流匹配（MM-FM），它将数据自适应的多模态源分布（以高斯混合模型实现）与依赖模式的数据耦合相结合。前者缩短了源分布与目标分布之间的距离，后者促进了局部且更直的流。我们还推导了理论结果，从定量意义上证实了我们的直觉。在 ImageNet256x256 上使用多模态 DINOv2-B 潜变量的实验中，MM-FM 展现出更优的学习效率与最优的无条件生成质量：仅用 80 个 epoch 并配合自引导即达到 FID=2.74。
英文原文摘要 1084 字符
Recently, vision foundation models have been shown to boost the efficiency of flow-based generative models by revealing the intrinsic union-of-manifold structures and lowering the complexity of the latent/target distribution. In this paper, we exploit the multimodality aspect of the union-of-manifold structures, and aim to further improve the learning and inference efficiency for flow-matching models. To this end, we propose an efficient source and coupling co-design method termed Mixture-Modeling Flow Matching (MM-FM), by integrating a data-adaptive multimodal source distribution (implemented as Gaussian mixture models) and mode-dependent data coupling. The former shortens the distance between the source and the target, and the latter promotes local and straighter flows. We also derive theoretical results to confirm our intuition in a quantitative sense. In our experiments on ImageNet256x256 with multimodal DINOv2-B latents, MM-FM exhibits superior learning efficiency and state-of-the-art unconditional generation quality: FID=2.74 with autoguidance in only 80 epochs.